Original Research

Nonalcoholic fatty liver disease (NAFLD) is a silent epidemic affecting nearly 1 in 3 Americans and is increasing within the Veterans Health Administration (VHA).^1,2 NAFLD independently increases the risk of type 2 diabetes mellitus (T2DM), cardiovascular disease, chronic kidney disease, cirrhosis, liver cancer, and death and impairs health-related quality of life (QOL).³ NAFLD primarily affects those with metabolic risk factors (prediabetes, T2DM, and metabolic syndrome) or obesity (Figure 1).^4,5 

Among those with NAFLD, most have nonalcoholic fatty liver (NAFL), or simple steatosis, affecting > 5% of liver cells (Figure 2).¹² 

In most patients (80%), NAFLD progresses slowly over decades. The progression is related to continuing insulin resistance.^15,16 Greater disease progression is seen in patients with T2DM or concomitant chronic liver disease (such as hepatitis C).^10,11,16 Patients with NAFLD who develop advanced fibrosis or cirrhosis experience increased rates of overall mortality, liver-related events, and liver transplantation.^1,9,17,18 Within the VHA, NAFLD is the third most common cause of cirrhosis and HCC, occurring at an average age of 66 and 70 years, respectively.¹⁹Less commonly, HCC also can occur in NAFLD without cirrhosis.²⁰

Although no pharmaceuticals are yet approved to treat NAFLD, even modest weight loss is beneficial. For example, weight loss > 4% improves fatty liver, ≥ 7% improves liver inflammation, and ≥ 10% decreases liver fibrosis (or scarring).^21-23 In patients with a prior lack of success with weight loss, weight loss medications may be beneficial for short-term use.²⁴ When comparing the effects of diet, exercise, obesity pharmacotherapy, and combinations for these approaches, intensive lifestyle modification with exercise had the greatest, most enduring benefit.²⁵ Additionally, bariatric (weight loss) surgery has significantly improved health and liver-related outcomes for patients with NASH.²⁶

In at-risk veterans, NAFLD has myriad negative effects on health and QOL. To improve its early identification and management in the VHA, we summarize strategies that all providers can use to screen and treat patients for this condition.

Screening for Advanced Fibrosis

Advanced fibrosis in NAFLD is diagnosed by analyzing adequately sized liver biopsies.^27,28 However, noninvasive approaches to quantify advanced fibrosis by imaging or use of a simple fibrosis prediction score also are available. Imaging modalities include measuring liver stiffness, using transient elastography (FibroScan, Waltham, MA) or magnetic resonance elastography.^1,29-31 Fibrosis prediction scores use common clinical and laboratory data to predict the presence or absence of advanced fibrosis (Table 1).²⁹ 

Does This Patient Have NAFLD?

To identify NAFLD, patients with metabolic syndrome and modest or no alcohol use are first assessed for liver injury with ALT, AST, and complete blood count (Figure 3; Case 1).¹⁶ 

Next, common underlying liver diseases that cause liver injury should be excluded by hepatitis B and C virus serology.^11,16 Other underlying liver diseases are uncommon and should be assessed only if clinically indicated. 

Evaluation of fasting glucose or hemoglobin A_1c (HbA_1c)can identify undiagnosed T2DM. NAFL, or simple steatosis, is independently associated with an increased risk of T2DM, cardiovascular and kidney disease, yet not overall mortality.¹⁶ Over 10 to 20 years, few patients (4%) with simple steatosis progress to cirrhosis.³⁹In contrast, NAFLD advanced fibrosis significantly increases overall and liver-related mortality and can be assessed with high probability by calculating the patient’s FIB-4, even in those with normal liver enzymes.^11,16 Patients with highly probable advanced fibrosis merit evaluation by hepatology or gastroenterology (Figure 3).

In NAFLD, simple steatosis can resolve, and NASH can significantly improve with 7% to 10% weight loss.^16,23,40 Patients with simple steatosis on imaging and normal liver enzymes should be monitored with periodic liver enzymes and fibrosis prediction scores (eg, FIB-4) and encouraged to pursue intensive lifestyle intervention.^16,33 Without weight loss and exercise interventions metabolic syndrome, T2DM, and NAFLD may progress.

Patients with combined liver steatosis and liver enzyme elevations may exhibit NASH and warrant an evaluation by a hepatologist or gastroenterologist for consideration of additional testing or liver biopsy.¹⁶While ALT elevations often have been used as a marker of NASH, ALT can be normal in NASH and in advanced fibrosis.^41,42 A liver biopsy is required to establish the diagnosis of NASH, which progresses to cirrhosis in 15% to 20% over a 10- to 20-year follow-up period (Case 2).³⁹ Fibrosis prediction scores also can evaluate the probability of advanced fibrosis in these patients.

Encouraging Patients to Pursue Intensive Lifestyle InterventionS

Most veterans wish to collaborate in their care (Table 3, Figure 4) yet experience many barriers, such as low health literacy, high rates of comorbidities, and ongoing drug/alcohol misuse.^43,44 

In addition to patient education, motivational interviewing significantly improves weight loss, resulting in a 3.3 lb (1.5 kg) increased weight loss in the intervention group vs the control group in weight loss studies.⁴⁶By being supportive, empathic, and clearly sharing the rationale for change, motivational interviewing is a collaborative conversation to guide patients to strengthen their motivation and commitment to change.⁴⁷ It helps patients examine and address their ambivalence—most recognize they should exercise and lose weight, but it can be difficult.

To start the conversation, the health care provider can explain that NAFLD increases the risk of T2DM, heart disease, and liver injury or scarring and can be effectively treated (or improved) with modest weight loss and regular exercise (ie, 14 lb weight loss if 200 lb, or 21 lb weight loss if 300 lb). Exercise can start with a 5-minute walk and build to 30 minutes daily). Then, the provider can ask the following 4 questions:

1. Why would you want to lose weight and exercise?
2. How might you go about it in order to succeed?
3. What are the 3 best reasons for you to do it?
4. How important is it for you to make this change, and why? The provider can also ask the patient to quantify on a scale of 1 to 10: (a) How likely is it that they will make each required change? (b) How hard will each change be for them?
5. The provider then summarizes the patient’s reasons for wanting change, how he/she can effect change, what their best reasons are, and how to successfully change. The provider then asks a final question:
6. So what do you think you will do?

Most patients report feeling engaged, empowered, open, and understood with motivational interviewing. People are “persuaded by what they hear themselves say,” increasing motivation to change.⁴⁷

This personalized action plan facilitates successful health behavior change.⁴⁸ Action plans should integrate daily routines. For example, by placing the scale near the toothbrush, daily weighing is encouraged. Daily weighing is associated with significantly greater weight loss and less weight regain.⁴⁹ In a 6-month, randomized controlled weight loss trial in men and women, daily weighing (using a scale that automatically transmitted weight data), with weekly e-mails and tailored feedback yielded an overall 9% weight loss and increased use of exercise and diet behaviors associated with weight loss in comparison with those who weighed themselves less than weekly.⁵⁰ This simple daily measure seems to reinforce a patient’s action plan.

Adherence to an action plan significantly improves with patient education, peer or social support, and addressing barriers to adherence.⁵¹ For example, by providing support with weekly text messaging of “How are you?” and addressing the issues that patients reported in a large randomized treatment trial, adherence was significantly improved.⁵² In VHA patients with low health literacy, peer support or telephone coaching also has proven effective in increasing weight loss and glycemic control in patients with T2DM.^53,54 Providing multidisciplinary team support during intensive lifestyle intervention, providers can partner with patients to address questions or issues and applaud progress.

Effective VHA interventions

In an ethnically diverse population of patients with prediabetes, up to 7% weight loss was observed in the Diabetes Prevention Program (DPP).⁵⁵ In this study patients were randomized to placebo; metformin 850 mg twice daily; or a lifestyle-modification program in which they received one-on-one culturally sensitive, individualized lessons in diet, moderate exercise (≥ 150 minutes weekly), and behavior modification from case managers over 16 sessions. Lessons were reinforced in both group and individual sessions. This intervention was associated with an average of 6% weight loss at 6 months (half of participants attained 7% weight loss) and a 58% decrease in the rate of progression to T2DM over a nearly 3-year follow-up of this population with prediabetes compared with that of the placebo group.⁵⁵ Over a 15-year follow-up, the intensive lifestyle intervention group sustained a 27% decrease in the incidence of T2DM compared with that of the placebo group.⁵⁶ To emulate the success of the DPP in the VHA, a web-based DPP-like study in female veterans was performed with online coaching and daily weighing. This study achieved a 5.2% weight loss from baseline at 4 months.⁵⁷

To improve outcomes, the VHA MOVE! Weight Management Program has been revised to include more sustained intervention (16 sessions) and multiple modes for participating—in person, by telephone, via video, via MOVE! Coach phone app, or any combination.⁵⁸ Using shared decision making between patients with NAFLD and their providers, a customized MOVE! weight loss program can be developed to enable sustained intensive lifestyle intervention: hypocaloric diet, ≥ 150 minutes of moderate exercise weekly, and behavioral change.

In addition to intensive lifestyle intervention, a prospective study found that bariatric surgery significantly improved outcomes in patients with NASH, with most patients experiencing resolution of their NASH and nearly half exhibiting significantly improved fibrosis.²⁶ In the VHA, bariatric surgery has yielded excellent long-term outcomes, with 21% sustained weight loss from baseline (vs matched nonsurgical population) at 10 years postoperatively in patients undergoing Roux-en-Y gastric bypass.⁵⁹ Bariatric surgery also results in long-term remission of T2DM in most patients and significant improvement in hypertension and dyslipidemia.⁶⁰ The risks of bariatric surgery include 3% serious complications, 1% reoperation rates, and 0.4% 30-day mortality.^61,62 Bariatric surgery can be considered in patients with BMI > 40 or in patients with BMI > 35 who have comorbidities and do not have decompensated cirrhosis.^63,64

Beyond weight loss, more favorable liver-related outcomes and lower rates of advanced liver fibrosis are observed in those consuming filtered coffee; a reduction in liver steatosis also is observed with adherence to a Mediterranean diet.^65,66 In NAFLD, statins may improve liver chemistries and fibrosis; this class of medications can be used safely even in the presence of an elevated ALT.^11,67As a risk factor for chronic liver disease, alcohol consumption of ≥ 4 drinks per day or > 14 drinks per week for men or > 7 drinks per week for women should be avoided in patients with NAFLD.¹¹

Conclusion

Nonalcoholic fatty liver disease independently increases the risk of T2DM, cardiovascular disease and kidney disease. With its rates increasing in the VHA, earlier identification and intervention is warranted in patients at high risk (ie, those with metabolic syndrome, obesity, and T2DM).² 

NASH is more frequent in those with liver enzyme elevations or with an elevated FIB-4 and is associated with a long-term risk of cirrhosis. These patients merit referral to hepatology or gastroenterology for further evaluation and consideration of a liver biopsy to identify NASH. Patients with likely NAFLD without liver enzyme elevations can be further evaluated with FIB-4 scores to assess their probability of advanced liver fibrosis and potential need for referral to hepatology or gastroenterology.

Early NAFLD detection and intervention with intensive lifestyle modifications has the potential to avert progression to advanced fibrosis—and its associated increased overall and liver-related mortality, and impaired QOL.^3,16,18  Although FIB-4 is a validated predictor of advanced fibrosis, this score is not yet used nationally to identify and risk stratify NAFLD in the VHA. Additionally, the very low use of VHA diet/exercise programs in eligible patients contributes to NAFLD progression.⁶⁸ The cost-effective DPP has successfully yielded weight loss in patients with prediabetes and decreases in the incidence of T2DM through motivational interviewing and intensive lifestyle intervention.⁵⁵ 

To improve NAFLD management, providers can successfully engage patients through motivational interviewing for intensive lifestyle intervention. Their resulting weight loss is enhanced with a personalized action plan, daily weighing, and peer support. When NAFLD is identified in patients with metabolic risk factors, the probability of advanced fibrosis is easily assessed in those with elevated FIB-4 scores who merit gastrointestinal referral.^33,37

In all those identified with NAFLD, disease information should be provided to patients and their families. Intensive lifestyle modification targeting a ≥ 7% weight loss is recommended; motivational interviewing can increase commitment to change and yield a customized action plan for sustained weight loss. Working with the support and encouragement of their team of primary care providers, dieticians, and MOVE! coaches, patients can actively engage to improve their NAFLD and overall health.

Nonalcoholic fatty liver disease (NAFLD) is a silent epidemic affecting nearly 1 in 3 Americans and is increasing within the Veterans Health Administration (VHA).^1,2 NAFLD independently increases the risk of type 2 diabetes mellitus (T2DM), cardiovascular disease, chronic kidney disease, cirrhosis, liver cancer, and death and impairs health-related quality of life (QOL).³ NAFLD primarily affects those with metabolic risk factors (prediabetes, T2DM, and metabolic syndrome) or obesity (Figure 1).^4,5 

Among those with NAFLD, most have nonalcoholic fatty liver (NAFL), or simple steatosis, affecting > 5% of liver cells (Figure 2).¹² 

In most patients (80%), NAFLD progresses slowly over decades. The progression is related to continuing insulin resistance.^15,16 Greater disease progression is seen in patients with T2DM or concomitant chronic liver disease (such as hepatitis C).^10,11,16 Patients with NAFLD who develop advanced fibrosis or cirrhosis experience increased rates of overall mortality, liver-related events, and liver transplantation.^1,9,17,18 Within the VHA, NAFLD is the third most common cause of cirrhosis and HCC, occurring at an average age of 66 and 70 years, respectively.¹⁹Less commonly, HCC also can occur in NAFLD without cirrhosis.²⁰

Although no pharmaceuticals are yet approved to treat NAFLD, even modest weight loss is beneficial. For example, weight loss > 4% improves fatty liver, ≥ 7% improves liver inflammation, and ≥ 10% decreases liver fibrosis (or scarring).^21-23 In patients with a prior lack of success with weight loss, weight loss medications may be beneficial for short-term use.²⁴ When comparing the effects of diet, exercise, obesity pharmacotherapy, and combinations for these approaches, intensive lifestyle modification with exercise had the greatest, most enduring benefit.²⁵ Additionally, bariatric (weight loss) surgery has significantly improved health and liver-related outcomes for patients with NASH.²⁶

In at-risk veterans, NAFLD has myriad negative effects on health and QOL. To improve its early identification and management in the VHA, we summarize strategies that all providers can use to screen and treat patients for this condition.

Screening for Advanced Fibrosis

Advanced fibrosis in NAFLD is diagnosed by analyzing adequately sized liver biopsies.^27,28 However, noninvasive approaches to quantify advanced fibrosis by imaging or use of a simple fibrosis prediction score also are available. Imaging modalities include measuring liver stiffness, using transient elastography (FibroScan, Waltham, MA) or magnetic resonance elastography.^1,29-31 Fibrosis prediction scores use common clinical and laboratory data to predict the presence or absence of advanced fibrosis (Table 1).²⁹ 

Does This Patient Have NAFLD?

To identify NAFLD, patients with metabolic syndrome and modest or no alcohol use are first assessed for liver injury with ALT, AST, and complete blood count (Figure 3; Case 1).¹⁶ 

Next, common underlying liver diseases that cause liver injury should be excluded by hepatitis B and C virus serology.^11,16 Other underlying liver diseases are uncommon and should be assessed only if clinically indicated. 

Evaluation of fasting glucose or hemoglobin A_1c (HbA_1c)can identify undiagnosed T2DM. NAFL, or simple steatosis, is independently associated with an increased risk of T2DM, cardiovascular and kidney disease, yet not overall mortality.¹⁶ Over 10 to 20 years, few patients (4%) with simple steatosis progress to cirrhosis.³⁹In contrast, NAFLD advanced fibrosis significantly increases overall and liver-related mortality and can be assessed with high probability by calculating the patient’s FIB-4, even in those with normal liver enzymes.^11,16 Patients with highly probable advanced fibrosis merit evaluation by hepatology or gastroenterology (Figure 3).

In NAFLD, simple steatosis can resolve, and NASH can significantly improve with 7% to 10% weight loss.^16,23,40 Patients with simple steatosis on imaging and normal liver enzymes should be monitored with periodic liver enzymes and fibrosis prediction scores (eg, FIB-4) and encouraged to pursue intensive lifestyle intervention.^16,33 Without weight loss and exercise interventions metabolic syndrome, T2DM, and NAFLD may progress.

Patients with combined liver steatosis and liver enzyme elevations may exhibit NASH and warrant an evaluation by a hepatologist or gastroenterologist for consideration of additional testing or liver biopsy.¹⁶While ALT elevations often have been used as a marker of NASH, ALT can be normal in NASH and in advanced fibrosis.^41,42 A liver biopsy is required to establish the diagnosis of NASH, which progresses to cirrhosis in 15% to 20% over a 10- to 20-year follow-up period (Case 2).³⁹ Fibrosis prediction scores also can evaluate the probability of advanced fibrosis in these patients.

Encouraging Patients to Pursue Intensive Lifestyle InterventionS

Most veterans wish to collaborate in their care (Table 3, Figure 4) yet experience many barriers, such as low health literacy, high rates of comorbidities, and ongoing drug/alcohol misuse.^43,44 

In addition to patient education, motivational interviewing significantly improves weight loss, resulting in a 3.3 lb (1.5 kg) increased weight loss in the intervention group vs the control group in weight loss studies.⁴⁶By being supportive, empathic, and clearly sharing the rationale for change, motivational interviewing is a collaborative conversation to guide patients to strengthen their motivation and commitment to change.⁴⁷ It helps patients examine and address their ambivalence—most recognize they should exercise and lose weight, but it can be difficult.

To start the conversation, the health care provider can explain that NAFLD increases the risk of T2DM, heart disease, and liver injury or scarring and can be effectively treated (or improved) with modest weight loss and regular exercise (ie, 14 lb weight loss if 200 lb, or 21 lb weight loss if 300 lb). Exercise can start with a 5-minute walk and build to 30 minutes daily). Then, the provider can ask the following 4 questions:

1. Why would you want to lose weight and exercise?
2. How might you go about it in order to succeed?
3. What are the 3 best reasons for you to do it?
4. How important is it for you to make this change, and why? The provider can also ask the patient to quantify on a scale of 1 to 10: (a) How likely is it that they will make each required change? (b) How hard will each change be for them?
5. The provider then summarizes the patient’s reasons for wanting change, how he/she can effect change, what their best reasons are, and how to successfully change. The provider then asks a final question:
6. So what do you think you will do?

Most patients report feeling engaged, empowered, open, and understood with motivational interviewing. People are “persuaded by what they hear themselves say,” increasing motivation to change.⁴⁷

This personalized action plan facilitates successful health behavior change.⁴⁸ Action plans should integrate daily routines. For example, by placing the scale near the toothbrush, daily weighing is encouraged. Daily weighing is associated with significantly greater weight loss and less weight regain.⁴⁹ In a 6-month, randomized controlled weight loss trial in men and women, daily weighing (using a scale that automatically transmitted weight data), with weekly e-mails and tailored feedback yielded an overall 9% weight loss and increased use of exercise and diet behaviors associated with weight loss in comparison with those who weighed themselves less than weekly.⁵⁰ This simple daily measure seems to reinforce a patient’s action plan.

Adherence to an action plan significantly improves with patient education, peer or social support, and addressing barriers to adherence.⁵¹ For example, by providing support with weekly text messaging of “How are you?” and addressing the issues that patients reported in a large randomized treatment trial, adherence was significantly improved.⁵² In VHA patients with low health literacy, peer support or telephone coaching also has proven effective in increasing weight loss and glycemic control in patients with T2DM.^53,54 Providing multidisciplinary team support during intensive lifestyle intervention, providers can partner with patients to address questions or issues and applaud progress.

Effective VHA interventions

In an ethnically diverse population of patients with prediabetes, up to 7% weight loss was observed in the Diabetes Prevention Program (DPP).⁵⁵ In this study patients were randomized to placebo; metformin 850 mg twice daily; or a lifestyle-modification program in which they received one-on-one culturally sensitive, individualized lessons in diet, moderate exercise (≥ 150 minutes weekly), and behavior modification from case managers over 16 sessions. Lessons were reinforced in both group and individual sessions. This intervention was associated with an average of 6% weight loss at 6 months (half of participants attained 7% weight loss) and a 58% decrease in the rate of progression to T2DM over a nearly 3-year follow-up of this population with prediabetes compared with that of the placebo group.⁵⁵ Over a 15-year follow-up, the intensive lifestyle intervention group sustained a 27% decrease in the incidence of T2DM compared with that of the placebo group.⁵⁶ To emulate the success of the DPP in the VHA, a web-based DPP-like study in female veterans was performed with online coaching and daily weighing. This study achieved a 5.2% weight loss from baseline at 4 months.⁵⁷

To improve outcomes, the VHA MOVE! Weight Management Program has been revised to include more sustained intervention (16 sessions) and multiple modes for participating—in person, by telephone, via video, via MOVE! Coach phone app, or any combination.⁵⁸ Using shared decision making between patients with NAFLD and their providers, a customized MOVE! weight loss program can be developed to enable sustained intensive lifestyle intervention: hypocaloric diet, ≥ 150 minutes of moderate exercise weekly, and behavioral change.

In addition to intensive lifestyle intervention, a prospective study found that bariatric surgery significantly improved outcomes in patients with NASH, with most patients experiencing resolution of their NASH and nearly half exhibiting significantly improved fibrosis.²⁶ In the VHA, bariatric surgery has yielded excellent long-term outcomes, with 21% sustained weight loss from baseline (vs matched nonsurgical population) at 10 years postoperatively in patients undergoing Roux-en-Y gastric bypass.⁵⁹ Bariatric surgery also results in long-term remission of T2DM in most patients and significant improvement in hypertension and dyslipidemia.⁶⁰ The risks of bariatric surgery include 3% serious complications, 1% reoperation rates, and 0.4% 30-day mortality.^61,62 Bariatric surgery can be considered in patients with BMI > 40 or in patients with BMI > 35 who have comorbidities and do not have decompensated cirrhosis.^63,64

Beyond weight loss, more favorable liver-related outcomes and lower rates of advanced liver fibrosis are observed in those consuming filtered coffee; a reduction in liver steatosis also is observed with adherence to a Mediterranean diet.^65,66 In NAFLD, statins may improve liver chemistries and fibrosis; this class of medications can be used safely even in the presence of an elevated ALT.^11,67As a risk factor for chronic liver disease, alcohol consumption of ≥ 4 drinks per day or > 14 drinks per week for men or > 7 drinks per week for women should be avoided in patients with NAFLD.¹¹

Conclusion

Nonalcoholic fatty liver disease independently increases the risk of T2DM, cardiovascular disease and kidney disease. With its rates increasing in the VHA, earlier identification and intervention is warranted in patients at high risk (ie, those with metabolic syndrome, obesity, and T2DM).² 

NASH is more frequent in those with liver enzyme elevations or with an elevated FIB-4 and is associated with a long-term risk of cirrhosis. These patients merit referral to hepatology or gastroenterology for further evaluation and consideration of a liver biopsy to identify NASH. Patients with likely NAFLD without liver enzyme elevations can be further evaluated with FIB-4 scores to assess their probability of advanced liver fibrosis and potential need for referral to hepatology or gastroenterology.

Early NAFLD detection and intervention with intensive lifestyle modifications has the potential to avert progression to advanced fibrosis—and its associated increased overall and liver-related mortality, and impaired QOL.^3,16,18  Although FIB-4 is a validated predictor of advanced fibrosis, this score is not yet used nationally to identify and risk stratify NAFLD in the VHA. Additionally, the very low use of VHA diet/exercise programs in eligible patients contributes to NAFLD progression.⁶⁸ The cost-effective DPP has successfully yielded weight loss in patients with prediabetes and decreases in the incidence of T2DM through motivational interviewing and intensive lifestyle intervention.⁵⁵ 

To improve NAFLD management, providers can successfully engage patients through motivational interviewing for intensive lifestyle intervention. Their resulting weight loss is enhanced with a personalized action plan, daily weighing, and peer support. When NAFLD is identified in patients with metabolic risk factors, the probability of advanced fibrosis is easily assessed in those with elevated FIB-4 scores who merit gastrointestinal referral.^33,37

In all those identified with NAFLD, disease information should be provided to patients and their families. Intensive lifestyle modification targeting a ≥ 7% weight loss is recommended; motivational interviewing can increase commitment to change and yield a customized action plan for sustained weight loss. Working with the support and encouragement of their team of primary care providers, dieticians, and MOVE! coaches, patients can actively engage to improve their NAFLD and overall health.

Nonalcoholic fatty liver disease (NAFLD) is a silent epidemic affecting nearly 1 in 3 Americans and is increasing within the Veterans Health Administration (VHA).^1,2 NAFLD independently increases the risk of type 2 diabetes mellitus (T2DM), cardiovascular disease, chronic kidney disease, cirrhosis, liver cancer, and death and impairs health-related quality of life (QOL).³ NAFLD primarily affects those with metabolic risk factors (prediabetes, T2DM, and metabolic syndrome) or obesity (Figure 1).^4,5 

Among those with NAFLD, most have nonalcoholic fatty liver (NAFL), or simple steatosis, affecting > 5% of liver cells (Figure 2).¹² 

In most patients (80%), NAFLD progresses slowly over decades. The progression is related to continuing insulin resistance.^15,16 Greater disease progression is seen in patients with T2DM or concomitant chronic liver disease (such as hepatitis C).^10,11,16 Patients with NAFLD who develop advanced fibrosis or cirrhosis experience increased rates of overall mortality, liver-related events, and liver transplantation.^1,9,17,18 Within the VHA, NAFLD is the third most common cause of cirrhosis and HCC, occurring at an average age of 66 and 70 years, respectively.¹⁹Less commonly, HCC also can occur in NAFLD without cirrhosis.²⁰

Although no pharmaceuticals are yet approved to treat NAFLD, even modest weight loss is beneficial. For example, weight loss > 4% improves fatty liver, ≥ 7% improves liver inflammation, and ≥ 10% decreases liver fibrosis (or scarring).^21-23 In patients with a prior lack of success with weight loss, weight loss medications may be beneficial for short-term use.²⁴ When comparing the effects of diet, exercise, obesity pharmacotherapy, and combinations for these approaches, intensive lifestyle modification with exercise had the greatest, most enduring benefit.²⁵ Additionally, bariatric (weight loss) surgery has significantly improved health and liver-related outcomes for patients with NASH.²⁶

In at-risk veterans, NAFLD has myriad negative effects on health and QOL. To improve its early identification and management in the VHA, we summarize strategies that all providers can use to screen and treat patients for this condition.

Screening for Advanced Fibrosis

Advanced fibrosis in NAFLD is diagnosed by analyzing adequately sized liver biopsies.^27,28 However, noninvasive approaches to quantify advanced fibrosis by imaging or use of a simple fibrosis prediction score also are available. Imaging modalities include measuring liver stiffness, using transient elastography (FibroScan, Waltham, MA) or magnetic resonance elastography.^1,29-31 Fibrosis prediction scores use common clinical and laboratory data to predict the presence or absence of advanced fibrosis (Table 1).²⁹ 

Does This Patient Have NAFLD?

To identify NAFLD, patients with metabolic syndrome and modest or no alcohol use are first assessed for liver injury with ALT, AST, and complete blood count (Figure 3; Case 1).¹⁶ 

Next, common underlying liver diseases that cause liver injury should be excluded by hepatitis B and C virus serology.^11,16 Other underlying liver diseases are uncommon and should be assessed only if clinically indicated. 

Evaluation of fasting glucose or hemoglobin A_1c (HbA_1c)can identify undiagnosed T2DM. NAFL, or simple steatosis, is independently associated with an increased risk of T2DM, cardiovascular and kidney disease, yet not overall mortality.¹⁶ Over 10 to 20 years, few patients (4%) with simple steatosis progress to cirrhosis.³⁹In contrast, NAFLD advanced fibrosis significantly increases overall and liver-related mortality and can be assessed with high probability by calculating the patient’s FIB-4, even in those with normal liver enzymes.^11,16 Patients with highly probable advanced fibrosis merit evaluation by hepatology or gastroenterology (Figure 3).

In NAFLD, simple steatosis can resolve, and NASH can significantly improve with 7% to 10% weight loss.^16,23,40 Patients with simple steatosis on imaging and normal liver enzymes should be monitored with periodic liver enzymes and fibrosis prediction scores (eg, FIB-4) and encouraged to pursue intensive lifestyle intervention.^16,33 Without weight loss and exercise interventions metabolic syndrome, T2DM, and NAFLD may progress.

Patients with combined liver steatosis and liver enzyme elevations may exhibit NASH and warrant an evaluation by a hepatologist or gastroenterologist for consideration of additional testing or liver biopsy.¹⁶While ALT elevations often have been used as a marker of NASH, ALT can be normal in NASH and in advanced fibrosis.^41,42 A liver biopsy is required to establish the diagnosis of NASH, which progresses to cirrhosis in 15% to 20% over a 10- to 20-year follow-up period (Case 2).³⁹ Fibrosis prediction scores also can evaluate the probability of advanced fibrosis in these patients.

Encouraging Patients to Pursue Intensive Lifestyle InterventionS

Most veterans wish to collaborate in their care (Table 3, Figure 4) yet experience many barriers, such as low health literacy, high rates of comorbidities, and ongoing drug/alcohol misuse.^43,44 

In addition to patient education, motivational interviewing significantly improves weight loss, resulting in a 3.3 lb (1.5 kg) increased weight loss in the intervention group vs the control group in weight loss studies.⁴⁶By being supportive, empathic, and clearly sharing the rationale for change, motivational interviewing is a collaborative conversation to guide patients to strengthen their motivation and commitment to change.⁴⁷ It helps patients examine and address their ambivalence—most recognize they should exercise and lose weight, but it can be difficult.

To start the conversation, the health care provider can explain that NAFLD increases the risk of T2DM, heart disease, and liver injury or scarring and can be effectively treated (or improved) with modest weight loss and regular exercise (ie, 14 lb weight loss if 200 lb, or 21 lb weight loss if 300 lb). Exercise can start with a 5-minute walk and build to 30 minutes daily). Then, the provider can ask the following 4 questions:

1. Why would you want to lose weight and exercise?
2. How might you go about it in order to succeed?
3. What are the 3 best reasons for you to do it?
4. How important is it for you to make this change, and why? The provider can also ask the patient to quantify on a scale of 1 to 10: (a) How likely is it that they will make each required change? (b) How hard will each change be for them?
5. The provider then summarizes the patient’s reasons for wanting change, how he/she can effect change, what their best reasons are, and how to successfully change. The provider then asks a final question:
6. So what do you think you will do?

Most patients report feeling engaged, empowered, open, and understood with motivational interviewing. People are “persuaded by what they hear themselves say,” increasing motivation to change.⁴⁷

This personalized action plan facilitates successful health behavior change.⁴⁸ Action plans should integrate daily routines. For example, by placing the scale near the toothbrush, daily weighing is encouraged. Daily weighing is associated with significantly greater weight loss and less weight regain.⁴⁹ In a 6-month, randomized controlled weight loss trial in men and women, daily weighing (using a scale that automatically transmitted weight data), with weekly e-mails and tailored feedback yielded an overall 9% weight loss and increased use of exercise and diet behaviors associated with weight loss in comparison with those who weighed themselves less than weekly.⁵⁰ This simple daily measure seems to reinforce a patient’s action plan.

Adherence to an action plan significantly improves with patient education, peer or social support, and addressing barriers to adherence.⁵¹ For example, by providing support with weekly text messaging of “How are you?” and addressing the issues that patients reported in a large randomized treatment trial, adherence was significantly improved.⁵² In VHA patients with low health literacy, peer support or telephone coaching also has proven effective in increasing weight loss and glycemic control in patients with T2DM.^53,54 Providing multidisciplinary team support during intensive lifestyle intervention, providers can partner with patients to address questions or issues and applaud progress.

Effective VHA interventions

In an ethnically diverse population of patients with prediabetes, up to 7% weight loss was observed in the Diabetes Prevention Program (DPP).⁵⁵ In this study patients were randomized to placebo; metformin 850 mg twice daily; or a lifestyle-modification program in which they received one-on-one culturally sensitive, individualized lessons in diet, moderate exercise (≥ 150 minutes weekly), and behavior modification from case managers over 16 sessions. Lessons were reinforced in both group and individual sessions. This intervention was associated with an average of 6% weight loss at 6 months (half of participants attained 7% weight loss) and a 58% decrease in the rate of progression to T2DM over a nearly 3-year follow-up of this population with prediabetes compared with that of the placebo group.⁵⁵ Over a 15-year follow-up, the intensive lifestyle intervention group sustained a 27% decrease in the incidence of T2DM compared with that of the placebo group.⁵⁶ To emulate the success of the DPP in the VHA, a web-based DPP-like study in female veterans was performed with online coaching and daily weighing. This study achieved a 5.2% weight loss from baseline at 4 months.⁵⁷

To improve outcomes, the VHA MOVE! Weight Management Program has been revised to include more sustained intervention (16 sessions) and multiple modes for participating—in person, by telephone, via video, via MOVE! Coach phone app, or any combination.⁵⁸ Using shared decision making between patients with NAFLD and their providers, a customized MOVE! weight loss program can be developed to enable sustained intensive lifestyle intervention: hypocaloric diet, ≥ 150 minutes of moderate exercise weekly, and behavioral change.

In addition to intensive lifestyle intervention, a prospective study found that bariatric surgery significantly improved outcomes in patients with NASH, with most patients experiencing resolution of their NASH and nearly half exhibiting significantly improved fibrosis.²⁶ In the VHA, bariatric surgery has yielded excellent long-term outcomes, with 21% sustained weight loss from baseline (vs matched nonsurgical population) at 10 years postoperatively in patients undergoing Roux-en-Y gastric bypass.⁵⁹ Bariatric surgery also results in long-term remission of T2DM in most patients and significant improvement in hypertension and dyslipidemia.⁶⁰ The risks of bariatric surgery include 3% serious complications, 1% reoperation rates, and 0.4% 30-day mortality.^61,62 Bariatric surgery can be considered in patients with BMI > 40 or in patients with BMI > 35 who have comorbidities and do not have decompensated cirrhosis.^63,64

Beyond weight loss, more favorable liver-related outcomes and lower rates of advanced liver fibrosis are observed in those consuming filtered coffee; a reduction in liver steatosis also is observed with adherence to a Mediterranean diet.^65,66 In NAFLD, statins may improve liver chemistries and fibrosis; this class of medications can be used safely even in the presence of an elevated ALT.^11,67As a risk factor for chronic liver disease, alcohol consumption of ≥ 4 drinks per day or > 14 drinks per week for men or > 7 drinks per week for women should be avoided in patients with NAFLD.¹¹

Conclusion

Nonalcoholic fatty liver disease independently increases the risk of T2DM, cardiovascular disease and kidney disease. With its rates increasing in the VHA, earlier identification and intervention is warranted in patients at high risk (ie, those with metabolic syndrome, obesity, and T2DM).² 

NASH is more frequent in those with liver enzyme elevations or with an elevated FIB-4 and is associated with a long-term risk of cirrhosis. These patients merit referral to hepatology or gastroenterology for further evaluation and consideration of a liver biopsy to identify NASH. Patients with likely NAFLD without liver enzyme elevations can be further evaluated with FIB-4 scores to assess their probability of advanced liver fibrosis and potential need for referral to hepatology or gastroenterology.

Early NAFLD detection and intervention with intensive lifestyle modifications has the potential to avert progression to advanced fibrosis—and its associated increased overall and liver-related mortality, and impaired QOL.^3,16,18  Although FIB-4 is a validated predictor of advanced fibrosis, this score is not yet used nationally to identify and risk stratify NAFLD in the VHA. Additionally, the very low use of VHA diet/exercise programs in eligible patients contributes to NAFLD progression.⁶⁸ The cost-effective DPP has successfully yielded weight loss in patients with prediabetes and decreases in the incidence of T2DM through motivational interviewing and intensive lifestyle intervention.⁵⁵ 

To improve NAFLD management, providers can successfully engage patients through motivational interviewing for intensive lifestyle intervention. Their resulting weight loss is enhanced with a personalized action plan, daily weighing, and peer support. When NAFLD is identified in patients with metabolic risk factors, the probability of advanced fibrosis is easily assessed in those with elevated FIB-4 scores who merit gastrointestinal referral.^33,37

In all those identified with NAFLD, disease information should be provided to patients and their families. Intensive lifestyle modification targeting a ≥ 7% weight loss is recommended; motivational interviewing can increase commitment to change and yield a customized action plan for sustained weight loss. Working with the support and encouragement of their team of primary care providers, dieticians, and MOVE! coaches, patients can actively engage to improve their NAFLD and overall health.

Females are the fastest growing population to seek care at the Veterans Health Administration (VHA).¹ Based on a 2014 study examining prevalence of military sexual trauma (MST), it is estimated that about one-third of females in the military screen positive for MST, and the rates are higher for younger veterans.² Military sexual trauma includes both rape and any sexual activity that occurred without consent; offensive sexual remarks or advances can also represent MST. The issue of MST, therefore, is an important one to address adequately, especially for female veterans who are screened through the VHA system.

Since 1992, the VHA has been required to provide services for MST, defined as “sexual harassment that is threatening in character or physical assault of a sexual nature that occurred while the victim was in the military.”³ Despite this mandate, it has taken many years for all VHA hospitals to adopt recommended screening tools to identify survivors of MST and give them proper resources. Only half of VHA hospitals adopted screening 6 years after the policy change.⁴ In addition, the environment in which the survivors receive MST care may trigger posttraumatic stress symptoms as many of the other patients seeking care at the VHA hospital resemble the perpetrators.⁵ Thus, up to half of females who report a history of MST do not receive care for their MST through the VHA.⁶

Having a history of MST significantly increases the risks of developing mental health disorders, including posttraumatic stress disorder (PTSD), major depressive disorder, generalized anxiety disorder, and suicidal ideation.² This group also has overall decreased quality of life (QOL). Female veterans have increased sexual dysfunction and dissatisfaction, which is heightened with a history of MST.⁷ Addressing MST requires treatment of all aspects of life affected by MST, such as mental health, sexual function, and QOL. The quality of treatment for MST through VHA hospitals deserves attention and likely still requires improvement with better incorporation of the patient’s perspective.

Qualitative research allows for incorporation of the patient’s perspective and is useful for exploring new ideas and themes.⁸ Current qualitative research using individual interviews of MST survivors focuses more on mental health treatment modalities through the VHA system and how resources are used within the system.^9,10 While it is important to understand the quantity of these resources, their quality also should be explored. Research has identified unique gender-specific concerns such as female-only mental health groups.¹⁰ However, there has been less focus on how to improve current therapies and the treatment modalities (regardless of whether it is a community service or at the VHA system) females find most helpful. There is a gap in understanding the patient’s perspective and assessment of current MST treatments as well as the unmet needs both within and outside of the VHA system. Therefore, the purpose of this study is 2-fold: (1) examine the utilization of VHA services for MST, as well as outside services, through focusgroup sessions; and (2) to offer specific recommendations for improving MST treatment for female veterans from the patient’s perspective.

Methods

After obtaining institutional review board approval (16-H192), females who screened positive for a history of MST, using the validated MST screening questionnaire, were recruited from the Women’s Continuity Clinic, Urology clinic, and via a research flyer placed within key locations at the New Mexico Veterans Affairs (VA) Health Care System (NMVAHCS).¹¹ Inclusion criteria were veterans aged > 18 years who could speak and understand English. Those who agreed to participate attended any 1 of 5 focus groups. Prior to initiation of the focus groups, the investigators generated a focus-group script, including specific questions or probes to explore treatment, unmet needs (such as other health conditions the veteran associated with MST that were not being addressed), and recommendations for care improvement.

Subjects granted consent privately prior to conduction of the focus group. Each participant completed a basic demographic (age, race, ethnicity) and clinical history (including pain conditions and therapy received for MST). These characteristics were evaluated with descriptive statistics, including means and frequencies.

The focus groups took place on the NM VAHCS Raymond G. Murphy VA Medical Center campus in a private conference room and were moderated by nonmedical research personnel experienced in focus-group moderation. Focus groups were recorded and transcribed. An iterative process was used with revisions to the script and probe questions as needed. Focus groups were planned for 2 hours but were allowed to continue at the participants’ discretion.

The de-identified transcripts were uploaded to the web-based qualitative engine Dedoose 6.2.21 software (Los Angeles, CA) and coded. Using grounded theory, the codes were grouped into themes and subsequently organized into emergent concepts.^8,12 Following constant comparative methodology, ideas were compared and combined between each focus group.^8,13 After completion of the focus groups, the generated ideas were organized and refined to create a conceptual framework that represented the collective ideas from the focus groups.

Results

Between January and June 2017, 5 focus groups with 17 participants were conducted; each session lasted about 3 hours. The average age was 52 ± 8.3 years, and were from a diverse racial and ethnic background. Most reported that > 20 years had passed since the first MST, and care-seeking for the first time was > 11 years after the trauma, although symptoms related to the MST most frequently began within 1 year of the trauma (Table 1). 

Preliminary Themes

The Trauma

Focus-group participants noted improved therapies offered by the VA but challenges obtaining health care:

“…because I’m really trying to deal with it and just be happy and get my joy back and deal with the isolation.”

“Another way that the memories affected me was barricading myself in my own house, starting from the front door.”

Male-Dominated VA

Participants also noted that, along with screening improving the system, dedicated female staff and service connection are important:

“The Womens Clinic is nice, and it’s nice to know that I can go there and I’m not having to discuss everything with men all over the place.”

“The other thing... that would be really good for survivors of MST, is help with disability.”

While the focus-group participants found dedicated women’s clinics helpful and providing improved care, the overall VA environment remains male-dominated:

“Because it’s really hard to relax and be vulnerable and be in your body and in your emotions if there‘s a bunch of penises around. When I saw these guys on the floor I’m like, I ain’t going in there.”

This male-dominated sense also incorporated a feeling of being misunderstood by a system that has traditionally cared for male veterans:

“People don‘t understand. They think, oh, you‘re overreacting, but they don’t know what it feels like to be inside.”

“I wouldn’t say they treat you like a second citizen, but it’s like almost every appointment I go to that’s not in the Women’s Clinic, the secretaries or whatever will be like ‘Oh, are you looking for somebody, or...’

Assumption Females Are Not Veterans

“There was an older gentleman behind me, they were like ‘Are you checking him in?’ I said, ‘I’m sure he’ll check himself in, but I’m checking myself in.’”

Participants also reported that there is an assumption that you’re not a veteran when you’re female:

“All of the care should be geared to be the same. And we know we need to recognize that men have their issues, and women will have their issues. But we don’t need to just say ‘all women have this issue, throw them over there.’”

Self-Doubt

“The world doesn’t validate rape, you asked for it, it was what you were wearing, it was what you said.”

Ongoing efforts to have female-only spaces, therapy groups, and support networks were encouraged by all 5 focus groups. These themes, provided the foundation for emergent concepts regarding patients’ perceptions of their treatment for MST: (1) Improvement has been slow but measurable; (2) VA cares more about male veterans; (3) The isolation from MST is pervasive; (4) It’s hard to navigate the VA system or any health care when you’re traumatized; and (5) Sexual assault leaves lasting self-doubt that providers need to address.

Isolation

Because there are barriers to seeking care the overarching method for coping with the effects of MST was isolation.

Overcoming the isolation was essential to seeking any care. Participants reported years of living alone, avoiding social situations and contexts, and difficulty with basic tasks because of the isolation.

“That the coping skills, that the isolation is a coping skill and all these things, and that I had to do that to survive.”

Lack of family and provider support and the VHA’s perceived focus on male veterans perpetuated this sense of isolation. Additionally, feeding the isolation were other maladaptive behaviors, such as alcoholism, weight gain, and anger.

“I was always an athlete until my MST, and I still find myself drinking whisky and wanting to smoke pot. It’s not that I want to, I guess it gives me a sense of relief, because my MST made me an alcoholic.”

Participants reported that successful treatment of MST must include treatment of other maladaptive behaviors and specific provider-behavior changes.

At times, providers contribute to female MST survivors’ feeling undervalued:

“I had an hour session and she kept looking at her watch and blowing me off, and I finally said, okay, I’m done, good-bye, after 45 minutes.”

Validation

Participants’ suggestions to improve MST treatment, including goal sharing, validation, knowledge, and support:

“They should have staff awareness groups, or focus groups to teach them the same thing that the patients are receiving as far as how to handle yourself, how to interact with others. Don’t bring your sh** from home into your job. You’re an employee, don’t take it personal.” (

Therapies for MST

A variety of modalities was recommended as helpful in coping with symptoms associated with MST. One female noted her therapy dog allowed her to get her first Papanicolaou (Pap) smear in years:

“Pelvic exams are like the seventh circle of hell. Like, God, you’d think I was being abducted by aliens or something. Last time, up here, they let me bring my little dog, which was extraordinarily helpful for me.”

For others, more traditional therapy such as prolonged exposure therapy or cognitive behavioral therapy, was helpful.

“After my prolonged exposure therapy; it saved my life. I’m not suicidal, and the only thing that’s really, really affected is sometimes I still have to sleep with a night light. Over 80% of the symptoms that I had and the problems that I had were alleviated with the therapy.”

Other veterans noted alternative therapies as beneficial for overcoming trauma:

“Yoga has really helped me with dealing with chronic pain and letting go of things that no longer serve me, and remembering about the inhale, the exhale, there’s a pause between the exhale and an inhale, where that’s where I make my choices, my thoughts, catch it, check it, change it, challenge my thoughts, that’s really, really helped me.”

From these concepts, and the specific suggestions female veterans provided for improvement in care, we developed a pictorial conceptual framework of the results. 

Discussion

This qualitative study of the quality of MST treatment with specific suggestions for improvement shows that the underlying force impacting health care in female survivors of MST is isolation. In turn, that isolation is perpetuated by personal beliefs, mental health, lack of support, and the VHA culture. While there was improvement in VHA care noted, female veterans offered many specific suggestions—simple ones that could be rapidly implemented—to enhance care. Many of these suggestions were targeted at provider-level behaviors such as validation, goal setting, knowledge (both about the military and about MST), and support.

Previous work showed that tangible (ie, words, being present) support rather than broad social support only generally helps reduces posttraumatic stress symptoms.¹⁵ These researchers found that tangible support moderated the relationship between number of lifetime traumas and PTSD. Schumm and colleagues also found that high social support predicted lower PTSD severity for inner-city women who experienced both child abuse and adult rape.¹⁶ A prior meta-analysis found social support was the strongest correlate of PTSD (effect size = 0.4).¹⁷

Our finding that female MST survivors desire verbal support from physicians may point to the inherent sense that validation helps healing, demonstrated by this meta-analysis. Importantly, the focus group participants did not specify the type of physician (psychiatrist, primary care provider, gynecologist, surgeon, etc) who needed to provide this support. Thus, we believe this suggestion is applicable to all physician interactions when the history of MST comes up. Physicians may be unaware of their profound impact in helping women recover from MST. This validation may also apply to survivors of other types of sexual trauma.

A second simple suggestion that arose from the focus groups was the need for broader options for MST therapy. Current data on the locations female veterans are treated for MST include specialty MST clinics, specialty PTSD clinics, psychosocial rehabilitation, and substance use disorder clinics, showing a wide range of settings.¹⁸ But female veterans are also asking for more services, including animal therapy, art therapy, yoga, and tai chi. While it may not be possible to offer every resource at every VHA facility, partnering with community services may help fulfill this veteran need. The advent of telehealth may also help address female veterans’ concerns about being surrounded by male patients and should be further explored.

The focus groups’ third suggestion for improvement in MST was better treatment for the health problems associated with sexual trauma, such as chronic pelvic pain, sexual dysfunction, and weight gain. It is important to note that the female veterans provided this list of associated health conditions from the broader facilitator question “What health problems do you think you have because of MST?” Females correctly identified common sequelae of sexual abuse, including pelvic pain and sexual dysfunction.^14,19 Weight gain and obesity have been associated with childhood sexual trauma and abuse, but they are not well studied in MST and may be worth further exploration.^20,21

Limitations

There are several inherent weaknesses in this study. The female veterans who agreed to participate in the focus group may not be representative of the entire population, particularly as survivors may be reluctant to talk about their MST experience. The participants in our focus groups were most commonly 2 decades past the MST and their experience with therapy may differ from that of women more recently traumatized and engaged in therapy. However, the fact that many of these females were still receiving some form of therapy 20 years after the traumatic event deserves attention.

Recall bias may have affected how female veterans described their experiences with MST treatment. We did not inquire about the timing of therapy and whether they sought VA care first, followed by community care, or vice versa. Finally, although the data were analyzed separately by 3 investigators, biases in data analysis may arise with qualitative methods.

Strengths of the study included the inherent patient-centered approach and ability to analyze data not readily extracted from patient records or validated questionnaires. Additionally, this qualitative approach allows for the discovery of patient-driven ideas and concerns. Our focus groups also contained a majority of minority females (including Hispanic and American Indian) populations that are frequently underrepresented in research.

Conclusion

Our data show there is still substantial room for improvement in the therapies and in the physician-level care for MST. While each treatment experience was unique, the collective agreement was that multimodal therapy was beneficial. However, the isolation that often comes from MST makes accessing care and treatment challenging. A crucial component to combating this isolation is provider validation and support for the female’s experience with MST. The simple act of hearing “I believe you” from the provider can make a huge impact on continuing to seek care and overcoming the consequences of MST.

Females are the fastest growing population to seek care at the Veterans Health Administration (VHA).¹ Based on a 2014 study examining prevalence of military sexual trauma (MST), it is estimated that about one-third of females in the military screen positive for MST, and the rates are higher for younger veterans.² Military sexual trauma includes both rape and any sexual activity that occurred without consent; offensive sexual remarks or advances can also represent MST. The issue of MST, therefore, is an important one to address adequately, especially for female veterans who are screened through the VHA system.

Since 1992, the VHA has been required to provide services for MST, defined as “sexual harassment that is threatening in character or physical assault of a sexual nature that occurred while the victim was in the military.”³ Despite this mandate, it has taken many years for all VHA hospitals to adopt recommended screening tools to identify survivors of MST and give them proper resources. Only half of VHA hospitals adopted screening 6 years after the policy change.⁴ In addition, the environment in which the survivors receive MST care may trigger posttraumatic stress symptoms as many of the other patients seeking care at the VHA hospital resemble the perpetrators.⁵ Thus, up to half of females who report a history of MST do not receive care for their MST through the VHA.⁶

Having a history of MST significantly increases the risks of developing mental health disorders, including posttraumatic stress disorder (PTSD), major depressive disorder, generalized anxiety disorder, and suicidal ideation.² This group also has overall decreased quality of life (QOL). Female veterans have increased sexual dysfunction and dissatisfaction, which is heightened with a history of MST.⁷ Addressing MST requires treatment of all aspects of life affected by MST, such as mental health, sexual function, and QOL. The quality of treatment for MST through VHA hospitals deserves attention and likely still requires improvement with better incorporation of the patient’s perspective.

Qualitative research allows for incorporation of the patient’s perspective and is useful for exploring new ideas and themes.⁸ Current qualitative research using individual interviews of MST survivors focuses more on mental health treatment modalities through the VHA system and how resources are used within the system.^9,10 While it is important to understand the quantity of these resources, their quality also should be explored. Research has identified unique gender-specific concerns such as female-only mental health groups.¹⁰ However, there has been less focus on how to improve current therapies and the treatment modalities (regardless of whether it is a community service or at the VHA system) females find most helpful. There is a gap in understanding the patient’s perspective and assessment of current MST treatments as well as the unmet needs both within and outside of the VHA system. Therefore, the purpose of this study is 2-fold: (1) examine the utilization of VHA services for MST, as well as outside services, through focusgroup sessions; and (2) to offer specific recommendations for improving MST treatment for female veterans from the patient’s perspective.

Methods

After obtaining institutional review board approval (16-H192), females who screened positive for a history of MST, using the validated MST screening questionnaire, were recruited from the Women’s Continuity Clinic, Urology clinic, and via a research flyer placed within key locations at the New Mexico Veterans Affairs (VA) Health Care System (NMVAHCS).¹¹ Inclusion criteria were veterans aged > 18 years who could speak and understand English. Those who agreed to participate attended any 1 of 5 focus groups. Prior to initiation of the focus groups, the investigators generated a focus-group script, including specific questions or probes to explore treatment, unmet needs (such as other health conditions the veteran associated with MST that were not being addressed), and recommendations for care improvement.

Subjects granted consent privately prior to conduction of the focus group. Each participant completed a basic demographic (age, race, ethnicity) and clinical history (including pain conditions and therapy received for MST). These characteristics were evaluated with descriptive statistics, including means and frequencies.

The focus groups took place on the NM VAHCS Raymond G. Murphy VA Medical Center campus in a private conference room and were moderated by nonmedical research personnel experienced in focus-group moderation. Focus groups were recorded and transcribed. An iterative process was used with revisions to the script and probe questions as needed. Focus groups were planned for 2 hours but were allowed to continue at the participants’ discretion.

The de-identified transcripts were uploaded to the web-based qualitative engine Dedoose 6.2.21 software (Los Angeles, CA) and coded. Using grounded theory, the codes were grouped into themes and subsequently organized into emergent concepts.^8,12 Following constant comparative methodology, ideas were compared and combined between each focus group.^8,13 After completion of the focus groups, the generated ideas were organized and refined to create a conceptual framework that represented the collective ideas from the focus groups.

Results

Between January and June 2017, 5 focus groups with 17 participants were conducted; each session lasted about 3 hours. The average age was 52 ± 8.3 years, and were from a diverse racial and ethnic background. Most reported that > 20 years had passed since the first MST, and care-seeking for the first time was > 11 years after the trauma, although symptoms related to the MST most frequently began within 1 year of the trauma (Table 1). 

Preliminary Themes

The Trauma

Focus-group participants noted improved therapies offered by the VA but challenges obtaining health care:

“…because I’m really trying to deal with it and just be happy and get my joy back and deal with the isolation.”

“Another way that the memories affected me was barricading myself in my own house, starting from the front door.”

Male-Dominated VA

Participants also noted that, along with screening improving the system, dedicated female staff and service connection are important:

“The Womens Clinic is nice, and it’s nice to know that I can go there and I’m not having to discuss everything with men all over the place.”

“The other thing... that would be really good for survivors of MST, is help with disability.”

While the focus-group participants found dedicated women’s clinics helpful and providing improved care, the overall VA environment remains male-dominated:

“Because it’s really hard to relax and be vulnerable and be in your body and in your emotions if there‘s a bunch of penises around. When I saw these guys on the floor I’m like, I ain’t going in there.”

This male-dominated sense also incorporated a feeling of being misunderstood by a system that has traditionally cared for male veterans:

“People don‘t understand. They think, oh, you‘re overreacting, but they don’t know what it feels like to be inside.”

“I wouldn’t say they treat you like a second citizen, but it’s like almost every appointment I go to that’s not in the Women’s Clinic, the secretaries or whatever will be like ‘Oh, are you looking for somebody, or...’

Assumption Females Are Not Veterans

“There was an older gentleman behind me, they were like ‘Are you checking him in?’ I said, ‘I’m sure he’ll check himself in, but I’m checking myself in.’”

Participants also reported that there is an assumption that you’re not a veteran when you’re female:

“All of the care should be geared to be the same. And we know we need to recognize that men have their issues, and women will have their issues. But we don’t need to just say ‘all women have this issue, throw them over there.’”

Self-Doubt

“The world doesn’t validate rape, you asked for it, it was what you were wearing, it was what you said.”

Ongoing efforts to have female-only spaces, therapy groups, and support networks were encouraged by all 5 focus groups. These themes, provided the foundation for emergent concepts regarding patients’ perceptions of their treatment for MST: (1) Improvement has been slow but measurable; (2) VA cares more about male veterans; (3) The isolation from MST is pervasive; (4) It’s hard to navigate the VA system or any health care when you’re traumatized; and (5) Sexual assault leaves lasting self-doubt that providers need to address.

Isolation

Because there are barriers to seeking care the overarching method for coping with the effects of MST was isolation.

Overcoming the isolation was essential to seeking any care. Participants reported years of living alone, avoiding social situations and contexts, and difficulty with basic tasks because of the isolation.

“That the coping skills, that the isolation is a coping skill and all these things, and that I had to do that to survive.”

Lack of family and provider support and the VHA’s perceived focus on male veterans perpetuated this sense of isolation. Additionally, feeding the isolation were other maladaptive behaviors, such as alcoholism, weight gain, and anger.

“I was always an athlete until my MST, and I still find myself drinking whisky and wanting to smoke pot. It’s not that I want to, I guess it gives me a sense of relief, because my MST made me an alcoholic.”

Participants reported that successful treatment of MST must include treatment of other maladaptive behaviors and specific provider-behavior changes.

At times, providers contribute to female MST survivors’ feeling undervalued:

“I had an hour session and she kept looking at her watch and blowing me off, and I finally said, okay, I’m done, good-bye, after 45 minutes.”

Validation

Participants’ suggestions to improve MST treatment, including goal sharing, validation, knowledge, and support:

“They should have staff awareness groups, or focus groups to teach them the same thing that the patients are receiving as far as how to handle yourself, how to interact with others. Don’t bring your sh** from home into your job. You’re an employee, don’t take it personal.” (

Therapies for MST

A variety of modalities was recommended as helpful in coping with symptoms associated with MST. One female noted her therapy dog allowed her to get her first Papanicolaou (Pap) smear in years:

“Pelvic exams are like the seventh circle of hell. Like, God, you’d think I was being abducted by aliens or something. Last time, up here, they let me bring my little dog, which was extraordinarily helpful for me.”

For others, more traditional therapy such as prolonged exposure therapy or cognitive behavioral therapy, was helpful.

“After my prolonged exposure therapy; it saved my life. I’m not suicidal, and the only thing that’s really, really affected is sometimes I still have to sleep with a night light. Over 80% of the symptoms that I had and the problems that I had were alleviated with the therapy.”

Other veterans noted alternative therapies as beneficial for overcoming trauma:

“Yoga has really helped me with dealing with chronic pain and letting go of things that no longer serve me, and remembering about the inhale, the exhale, there’s a pause between the exhale and an inhale, where that’s where I make my choices, my thoughts, catch it, check it, change it, challenge my thoughts, that’s really, really helped me.”

From these concepts, and the specific suggestions female veterans provided for improvement in care, we developed a pictorial conceptual framework of the results. 

Discussion

This qualitative study of the quality of MST treatment with specific suggestions for improvement shows that the underlying force impacting health care in female survivors of MST is isolation. In turn, that isolation is perpetuated by personal beliefs, mental health, lack of support, and the VHA culture. While there was improvement in VHA care noted, female veterans offered many specific suggestions—simple ones that could be rapidly implemented—to enhance care. Many of these suggestions were targeted at provider-level behaviors such as validation, goal setting, knowledge (both about the military and about MST), and support.

Previous work showed that tangible (ie, words, being present) support rather than broad social support only generally helps reduces posttraumatic stress symptoms.¹⁵ These researchers found that tangible support moderated the relationship between number of lifetime traumas and PTSD. Schumm and colleagues also found that high social support predicted lower PTSD severity for inner-city women who experienced both child abuse and adult rape.¹⁶ A prior meta-analysis found social support was the strongest correlate of PTSD (effect size = 0.4).¹⁷

Our finding that female MST survivors desire verbal support from physicians may point to the inherent sense that validation helps healing, demonstrated by this meta-analysis. Importantly, the focus group participants did not specify the type of physician (psychiatrist, primary care provider, gynecologist, surgeon, etc) who needed to provide this support. Thus, we believe this suggestion is applicable to all physician interactions when the history of MST comes up. Physicians may be unaware of their profound impact in helping women recover from MST. This validation may also apply to survivors of other types of sexual trauma.

A second simple suggestion that arose from the focus groups was the need for broader options for MST therapy. Current data on the locations female veterans are treated for MST include specialty MST clinics, specialty PTSD clinics, psychosocial rehabilitation, and substance use disorder clinics, showing a wide range of settings.¹⁸ But female veterans are also asking for more services, including animal therapy, art therapy, yoga, and tai chi. While it may not be possible to offer every resource at every VHA facility, partnering with community services may help fulfill this veteran need. The advent of telehealth may also help address female veterans’ concerns about being surrounded by male patients and should be further explored.

The focus groups’ third suggestion for improvement in MST was better treatment for the health problems associated with sexual trauma, such as chronic pelvic pain, sexual dysfunction, and weight gain. It is important to note that the female veterans provided this list of associated health conditions from the broader facilitator question “What health problems do you think you have because of MST?” Females correctly identified common sequelae of sexual abuse, including pelvic pain and sexual dysfunction.^14,19 Weight gain and obesity have been associated with childhood sexual trauma and abuse, but they are not well studied in MST and may be worth further exploration.^20,21

Limitations

There are several inherent weaknesses in this study. The female veterans who agreed to participate in the focus group may not be representative of the entire population, particularly as survivors may be reluctant to talk about their MST experience. The participants in our focus groups were most commonly 2 decades past the MST and their experience with therapy may differ from that of women more recently traumatized and engaged in therapy. However, the fact that many of these females were still receiving some form of therapy 20 years after the traumatic event deserves attention.

Recall bias may have affected how female veterans described their experiences with MST treatment. We did not inquire about the timing of therapy and whether they sought VA care first, followed by community care, or vice versa. Finally, although the data were analyzed separately by 3 investigators, biases in data analysis may arise with qualitative methods.

Strengths of the study included the inherent patient-centered approach and ability to analyze data not readily extracted from patient records or validated questionnaires. Additionally, this qualitative approach allows for the discovery of patient-driven ideas and concerns. Our focus groups also contained a majority of minority females (including Hispanic and American Indian) populations that are frequently underrepresented in research.

Conclusion

Our data show there is still substantial room for improvement in the therapies and in the physician-level care for MST. While each treatment experience was unique, the collective agreement was that multimodal therapy was beneficial. However, the isolation that often comes from MST makes accessing care and treatment challenging. A crucial component to combating this isolation is provider validation and support for the female’s experience with MST. The simple act of hearing “I believe you” from the provider can make a huge impact on continuing to seek care and overcoming the consequences of MST.

Females are the fastest growing population to seek care at the Veterans Health Administration (VHA).¹ Based on a 2014 study examining prevalence of military sexual trauma (MST), it is estimated that about one-third of females in the military screen positive for MST, and the rates are higher for younger veterans.² Military sexual trauma includes both rape and any sexual activity that occurred without consent; offensive sexual remarks or advances can also represent MST. The issue of MST, therefore, is an important one to address adequately, especially for female veterans who are screened through the VHA system.

Since 1992, the VHA has been required to provide services for MST, defined as “sexual harassment that is threatening in character or physical assault of a sexual nature that occurred while the victim was in the military.”³ Despite this mandate, it has taken many years for all VHA hospitals to adopt recommended screening tools to identify survivors of MST and give them proper resources. Only half of VHA hospitals adopted screening 6 years after the policy change.⁴ In addition, the environment in which the survivors receive MST care may trigger posttraumatic stress symptoms as many of the other patients seeking care at the VHA hospital resemble the perpetrators.⁵ Thus, up to half of females who report a history of MST do not receive care for their MST through the VHA.⁶

Having a history of MST significantly increases the risks of developing mental health disorders, including posttraumatic stress disorder (PTSD), major depressive disorder, generalized anxiety disorder, and suicidal ideation.² This group also has overall decreased quality of life (QOL). Female veterans have increased sexual dysfunction and dissatisfaction, which is heightened with a history of MST.⁷ Addressing MST requires treatment of all aspects of life affected by MST, such as mental health, sexual function, and QOL. The quality of treatment for MST through VHA hospitals deserves attention and likely still requires improvement with better incorporation of the patient’s perspective.

Qualitative research allows for incorporation of the patient’s perspective and is useful for exploring new ideas and themes.⁸ Current qualitative research using individual interviews of MST survivors focuses more on mental health treatment modalities through the VHA system and how resources are used within the system.^9,10 While it is important to understand the quantity of these resources, their quality also should be explored. Research has identified unique gender-specific concerns such as female-only mental health groups.¹⁰ However, there has been less focus on how to improve current therapies and the treatment modalities (regardless of whether it is a community service or at the VHA system) females find most helpful. There is a gap in understanding the patient’s perspective and assessment of current MST treatments as well as the unmet needs both within and outside of the VHA system. Therefore, the purpose of this study is 2-fold: (1) examine the utilization of VHA services for MST, as well as outside services, through focusgroup sessions; and (2) to offer specific recommendations for improving MST treatment for female veterans from the patient’s perspective.

Methods

After obtaining institutional review board approval (16-H192), females who screened positive for a history of MST, using the validated MST screening questionnaire, were recruited from the Women’s Continuity Clinic, Urology clinic, and via a research flyer placed within key locations at the New Mexico Veterans Affairs (VA) Health Care System (NMVAHCS).¹¹ Inclusion criteria were veterans aged > 18 years who could speak and understand English. Those who agreed to participate attended any 1 of 5 focus groups. Prior to initiation of the focus groups, the investigators generated a focus-group script, including specific questions or probes to explore treatment, unmet needs (such as other health conditions the veteran associated with MST that were not being addressed), and recommendations for care improvement.

Subjects granted consent privately prior to conduction of the focus group. Each participant completed a basic demographic (age, race, ethnicity) and clinical history (including pain conditions and therapy received for MST). These characteristics were evaluated with descriptive statistics, including means and frequencies.

The focus groups took place on the NM VAHCS Raymond G. Murphy VA Medical Center campus in a private conference room and were moderated by nonmedical research personnel experienced in focus-group moderation. Focus groups were recorded and transcribed. An iterative process was used with revisions to the script and probe questions as needed. Focus groups were planned for 2 hours but were allowed to continue at the participants’ discretion.

The de-identified transcripts were uploaded to the web-based qualitative engine Dedoose 6.2.21 software (Los Angeles, CA) and coded. Using grounded theory, the codes were grouped into themes and subsequently organized into emergent concepts.^8,12 Following constant comparative methodology, ideas were compared and combined between each focus group.^8,13 After completion of the focus groups, the generated ideas were organized and refined to create a conceptual framework that represented the collective ideas from the focus groups.

Results

Between January and June 2017, 5 focus groups with 17 participants were conducted; each session lasted about 3 hours. The average age was 52 ± 8.3 years, and were from a diverse racial and ethnic background. Most reported that > 20 years had passed since the first MST, and care-seeking for the first time was > 11 years after the trauma, although symptoms related to the MST most frequently began within 1 year of the trauma (Table 1). 

Preliminary Themes

The Trauma

Focus-group participants noted improved therapies offered by the VA but challenges obtaining health care:

“…because I’m really trying to deal with it and just be happy and get my joy back and deal with the isolation.”

“Another way that the memories affected me was barricading myself in my own house, starting from the front door.”

Male-Dominated VA

Participants also noted that, along with screening improving the system, dedicated female staff and service connection are important:

“The Womens Clinic is nice, and it’s nice to know that I can go there and I’m not having to discuss everything with men all over the place.”

“The other thing... that would be really good for survivors of MST, is help with disability.”

While the focus-group participants found dedicated women’s clinics helpful and providing improved care, the overall VA environment remains male-dominated:

“Because it’s really hard to relax and be vulnerable and be in your body and in your emotions if there‘s a bunch of penises around. When I saw these guys on the floor I’m like, I ain’t going in there.”

This male-dominated sense also incorporated a feeling of being misunderstood by a system that has traditionally cared for male veterans:

“People don‘t understand. They think, oh, you‘re overreacting, but they don’t know what it feels like to be inside.”

“I wouldn’t say they treat you like a second citizen, but it’s like almost every appointment I go to that’s not in the Women’s Clinic, the secretaries or whatever will be like ‘Oh, are you looking for somebody, or...’

Assumption Females Are Not Veterans

“There was an older gentleman behind me, they were like ‘Are you checking him in?’ I said, ‘I’m sure he’ll check himself in, but I’m checking myself in.’”

Participants also reported that there is an assumption that you’re not a veteran when you’re female:

“All of the care should be geared to be the same. And we know we need to recognize that men have their issues, and women will have their issues. But we don’t need to just say ‘all women have this issue, throw them over there.’”

Self-Doubt

“The world doesn’t validate rape, you asked for it, it was what you were wearing, it was what you said.”

Ongoing efforts to have female-only spaces, therapy groups, and support networks were encouraged by all 5 focus groups. These themes, provided the foundation for emergent concepts regarding patients’ perceptions of their treatment for MST: (1) Improvement has been slow but measurable; (2) VA cares more about male veterans; (3) The isolation from MST is pervasive; (4) It’s hard to navigate the VA system or any health care when you’re traumatized; and (5) Sexual assault leaves lasting self-doubt that providers need to address.

Isolation

Because there are barriers to seeking care the overarching method for coping with the effects of MST was isolation.

Overcoming the isolation was essential to seeking any care. Participants reported years of living alone, avoiding social situations and contexts, and difficulty with basic tasks because of the isolation.

“That the coping skills, that the isolation is a coping skill and all these things, and that I had to do that to survive.”

Lack of family and provider support and the VHA’s perceived focus on male veterans perpetuated this sense of isolation. Additionally, feeding the isolation were other maladaptive behaviors, such as alcoholism, weight gain, and anger.

“I was always an athlete until my MST, and I still find myself drinking whisky and wanting to smoke pot. It’s not that I want to, I guess it gives me a sense of relief, because my MST made me an alcoholic.”

Participants reported that successful treatment of MST must include treatment of other maladaptive behaviors and specific provider-behavior changes.

At times, providers contribute to female MST survivors’ feeling undervalued:

“I had an hour session and she kept looking at her watch and blowing me off, and I finally said, okay, I’m done, good-bye, after 45 minutes.”

Validation

Participants’ suggestions to improve MST treatment, including goal sharing, validation, knowledge, and support:

“They should have staff awareness groups, or focus groups to teach them the same thing that the patients are receiving as far as how to handle yourself, how to interact with others. Don’t bring your sh** from home into your job. You’re an employee, don’t take it personal.” (

Therapies for MST

A variety of modalities was recommended as helpful in coping with symptoms associated with MST. One female noted her therapy dog allowed her to get her first Papanicolaou (Pap) smear in years:

“Pelvic exams are like the seventh circle of hell. Like, God, you’d think I was being abducted by aliens or something. Last time, up here, they let me bring my little dog, which was extraordinarily helpful for me.”

For others, more traditional therapy such as prolonged exposure therapy or cognitive behavioral therapy, was helpful.

“After my prolonged exposure therapy; it saved my life. I’m not suicidal, and the only thing that’s really, really affected is sometimes I still have to sleep with a night light. Over 80% of the symptoms that I had and the problems that I had were alleviated with the therapy.”

Other veterans noted alternative therapies as beneficial for overcoming trauma:

“Yoga has really helped me with dealing with chronic pain and letting go of things that no longer serve me, and remembering about the inhale, the exhale, there’s a pause between the exhale and an inhale, where that’s where I make my choices, my thoughts, catch it, check it, change it, challenge my thoughts, that’s really, really helped me.”

From these concepts, and the specific suggestions female veterans provided for improvement in care, we developed a pictorial conceptual framework of the results. 

Discussion

This qualitative study of the quality of MST treatment with specific suggestions for improvement shows that the underlying force impacting health care in female survivors of MST is isolation. In turn, that isolation is perpetuated by personal beliefs, mental health, lack of support, and the VHA culture. While there was improvement in VHA care noted, female veterans offered many specific suggestions—simple ones that could be rapidly implemented—to enhance care. Many of these suggestions were targeted at provider-level behaviors such as validation, goal setting, knowledge (both about the military and about MST), and support.

Previous work showed that tangible (ie, words, being present) support rather than broad social support only generally helps reduces posttraumatic stress symptoms.¹⁵ These researchers found that tangible support moderated the relationship between number of lifetime traumas and PTSD. Schumm and colleagues also found that high social support predicted lower PTSD severity for inner-city women who experienced both child abuse and adult rape.¹⁶ A prior meta-analysis found social support was the strongest correlate of PTSD (effect size = 0.4).¹⁷

Our finding that female MST survivors desire verbal support from physicians may point to the inherent sense that validation helps healing, demonstrated by this meta-analysis. Importantly, the focus group participants did not specify the type of physician (psychiatrist, primary care provider, gynecologist, surgeon, etc) who needed to provide this support. Thus, we believe this suggestion is applicable to all physician interactions when the history of MST comes up. Physicians may be unaware of their profound impact in helping women recover from MST. This validation may also apply to survivors of other types of sexual trauma.

A second simple suggestion that arose from the focus groups was the need for broader options for MST therapy. Current data on the locations female veterans are treated for MST include specialty MST clinics, specialty PTSD clinics, psychosocial rehabilitation, and substance use disorder clinics, showing a wide range of settings.¹⁸ But female veterans are also asking for more services, including animal therapy, art therapy, yoga, and tai chi. While it may not be possible to offer every resource at every VHA facility, partnering with community services may help fulfill this veteran need. The advent of telehealth may also help address female veterans’ concerns about being surrounded by male patients and should be further explored.

The focus groups’ third suggestion for improvement in MST was better treatment for the health problems associated with sexual trauma, such as chronic pelvic pain, sexual dysfunction, and weight gain. It is important to note that the female veterans provided this list of associated health conditions from the broader facilitator question “What health problems do you think you have because of MST?” Females correctly identified common sequelae of sexual abuse, including pelvic pain and sexual dysfunction.^14,19 Weight gain and obesity have been associated with childhood sexual trauma and abuse, but they are not well studied in MST and may be worth further exploration.^20,21

Limitations

There are several inherent weaknesses in this study. The female veterans who agreed to participate in the focus group may not be representative of the entire population, particularly as survivors may be reluctant to talk about their MST experience. The participants in our focus groups were most commonly 2 decades past the MST and their experience with therapy may differ from that of women more recently traumatized and engaged in therapy. However, the fact that many of these females were still receiving some form of therapy 20 years after the traumatic event deserves attention.

Recall bias may have affected how female veterans described their experiences with MST treatment. We did not inquire about the timing of therapy and whether they sought VA care first, followed by community care, or vice versa. Finally, although the data were analyzed separately by 3 investigators, biases in data analysis may arise with qualitative methods.

Strengths of the study included the inherent patient-centered approach and ability to analyze data not readily extracted from patient records or validated questionnaires. Additionally, this qualitative approach allows for the discovery of patient-driven ideas and concerns. Our focus groups also contained a majority of minority females (including Hispanic and American Indian) populations that are frequently underrepresented in research.

Conclusion

Our data show there is still substantial room for improvement in the therapies and in the physician-level care for MST. While each treatment experience was unique, the collective agreement was that multimodal therapy was beneficial. However, the isolation that often comes from MST makes accessing care and treatment challenging. A crucial component to combating this isolation is provider validation and support for the female’s experience with MST. The simple act of hearing “I believe you” from the provider can make a huge impact on continuing to seek care and overcoming the consequences of MST.

The Centers of Medicare and Medicaid Services (CMS) has introduced new payment models that tie quality and value incentives to 90% of fee-for-service payments and provide 50% of Medicare payments through alternative payment models.¹ The push toward value comes after productivity-based physician reimbursement (ie, fee for service) has been associated with poor quality care, including delayed diagnoses, complications, readmissions, increased length of stay, and high costs of care.^2-5 The method of physician payment is widely believed to affect clinical behavior by incentivizing doing more, coding for more, and billing for more.^6-7 Although payment systems may be used to achieve policy objectives,⁸ little is known about the association of different payment systems with the culture of delivering value-based care among frontline clinicians.

Culture is defined as a system of shared assumptions, values, beliefs, and norms within an environment and has a powerful role in shaping clinician practice patterns.^9-12 The culture within medicine currently contributes to the overuse of resources,^11,13 and a culture for improvement is correlated with clinical outcomes. A systematic review found a consistent association between positive organization culture and improved outcomes including mortality.¹⁴ Across health systems, institutions with high scores on patient safety culture surveys have shown improvements in clinical behaviors and patient outcomes.^15-18

In this study, we aim to describe high-value care culture among internal medicine hospitalists across diverse hospitals and evaluate the relationship between physician reimbursement and high-value care culture.

Study Design

This study is an observational, cross-sectional survey-based study of hospitalists from 12 hospitals in California between January and June 2016.

Study Population

A total of 12 hospitals with hospitalist programs in California were chosen to represent three types of hospitals (ie, four university, four community, and four safety net). Safety-net hospitals, which traditionally serve low-income and medically and socially vulnerable patients were defined as those in the top quartile (ie, greater than 0.5) of their Disproportionate Share Index (DSH), which measures Medicaid patient load.^19-20

To select hospitals with varying value-based care performance, we stratified using CMS value-based purchasing (VBP) scores from fiscal year 2015; these scores have been used to adjust reimbursement for just over 3,000 hospitals in the VBP program of CMS.^22,23 CMS calculates the VBP total performance score as a composite of four domains: (1) clinical processes of care (20% of total performance); (2) patient satisfaction (30%); (3) patient outcomes, including mortality and complications (30%); and (4) cost defined by Medicare payment per beneficiary (20%).²¹ Established quality measures are based on data reported by participating hospitals and chart abstraction during 2011-2014.²² Although other clinical measures of care intensity have been used as proxies of value-based care,^23,24 we used the measure of value that has been publically reported by the CMS VBP given its wide use and effects on reimbursements for 80% of hospitals in the CMS VBP program in 2015.²⁵ We obtained institution-level data from the CMS VBP Program and Hospital Compare files. Each of the three types of hospitals represented institutions with low, middle, and high VBP performance (split in tertiles) as reported by the CMS VBP program. To increase the number of participants in tertiles with fewer hospitalists, a fourth hospital was selected for each hospital type.

We excluded individual hospitalists who primarily identified as working in subspecialty divisions and those who spent less than eight weeks during the last year providing direct patient care on inpatient internal medicine services at the studied institution.

Measurement

Hospitalists were asked to complete the High-Value Care Culture Survey (HVCCS^TM), which measures the culture of value-based decision making among frontline clinicians.²⁶ Similar to other validated surveys for the assessment of patient safety culture,^27,28 the HVCCS can be used to identify target areas for improvement. The survey includes four domains: (1) leadership and health system messaging, (2) data transparency and access, (3) comfort with cost conversations, and (4) blame-free environment. This tool was developed by using a two-phase national modified Delphi process. It was evaluated at two academic centers to complete factor analysis and assess internal consistency, reliability, and validity among internal medicine hospitalists and residents. Validation included estimating product-moment correlation of overall HVCCS scores and domain scores with the CMS institutional VBP scores. HVCCS scores are standardized to a 0-100 point scale for each of the four domains and are then averaged to obtain an overall score.²⁶

In the survey, value was defined as the quality of care provided to patients in relation to the costs required to deliver that care, and high-value care was defined as care that tried to maximize quality while minimizing costs. Quality was defined as the degree to which health services increased the likelihood of desired health outcomes that are safe, effective, patient centered, timely, equitable, and consistent with current professional knowledge. Cost was defined as the negative financial, physical, and emotional effects on patients and the health system.²⁶

Data Analysis

We described the overall institutional mean high-value care culture and domain scores measured by the HVCCS, hospitalist demographics and training experiences, and hospital characteristics. We also described individual survey items. Descriptive statistics were stratified and compared on the basis of hospital type (ie, safety net, community, or university). We assessed the relationship between the clinician perception of reimbursement structure within their divisions and individually reported high-value care culture scores using bivariate and multilevel linear regression. We hypothesized that compared with hospitalists who were paid with salaries or wages, those who reported reimbursement with productivity adjustments may report lower HVCCS scores and those who reported reimbursement with quality or value adjustments may report higher HVCCS scores. We adjusted for physician- and hospital-level characteristics, including age, gender, and training track, and considered hospital type and size as random effects.

This study was approved by the Institutional Review Board at all 12 sites. All analyses were conducted using STATA^® 13.0 (College Station, Texas).

Hospitalist Characteristics

A total of 255 (68.9%, 255/370) hospitalists across all sites completed the survey. Of these respondents, 135 were female (50.6%). On average, hospitalists were 39 years of age (SD 6.8), trained in categorical tracks (221; 86.7%), and had previously trained for 14.3 months at a safety-net hospital (SD 14.2). In total, 166 hospitalists (65.1%) reported being paid with salary or wages, 77 (30.2%) with salary plus productivity adjustments, and 12 (4.7%) with salary plus quality or value adjustments. Moreover, 123 (48.6%) hospitalists agreed that funding for their group depended on the volume of services they delivered. Community-based hospitalists reported higher rates of reimbursement with salary plus productivity (47; 32.0%) compared with their counterparts from university-based (24; 28.2%) and safety-net based programs (6; 26.1%). Among the three different hospital types, significant differences exist in hospitalist mean age (P < .001), gender (P = .01), and the number of months training in a safety-net hospital (P = .02; Table 1).

Hospital Characteristics

Of the 12 study sites, four from each type of hospital (ie, safety-net based, community based, and university based) and four representing each value-based purchasing performance tertile (ie, high, middle, and low) were included. Eleven (91.7%) sites were located in urban areas with an average DSH index of 0.40 (SD 0.23), case mix index of 1.97 (SD 0.28), and bed size of 435.5 (SD 146.0; Table 1).

In multilevel regression modeling across all 12 sites, hospitalists from community-based hospitalist programs reported lower mean HVCCS scores (β = −4.4, 95% CI −8.1 to −0.7; Table 2) than those from other hospital types.

High-Value Care Culture Survey Scores

The mean HVCCS score was 50.2 (SD 13.6), and mean domain scores across all sites were 65.4 (SD 15.6) for leadership and health system messaging, 32.4 (SD 22.8) for data transparency and access, 52.1 (SD 19.7) for comfort with cost conversations, and 50.7 (SD 21.4) for blame-free environment (Table 1). For the majority (two-thirds) of individual HVCCS items, more than 30% of hospitalists across all sites agreed or strongly agreed that components of a low-value care culture exist within their institutions. For example, over 80% of hospitalists reported low transparency and limited access to data (see Appendix I for complete survey responses).

Hospitalists reported different HVCCS domains as strengths or weaknesses within their institutions in accordance with hospital type. Compared with university-based and safety-net-based hospitalists, community-based hospitalists reported lower scores in having a blame-free environment (466, SD 21.8). Nearly 50% reported that the clinicians’ fear of legal repercussions affects their frequency of ordering unneeded tests or procedures, and 30% reported that individual clinicians are blamed for complications. Nearly 40% reported that clinicians are uncomfortable discussing the costs of tests or treatments with patients and reported that clinicians do not feel that physicians should discuss costs with patients. Notably, community-based hospitalists uniquely differed in how they reported components of leadership and health system messaging. Over 60% reported a work climate or role modeling supportive of delivering quality care at lower costs. Only 48%, however, reported success seen from implemented efforts, and 45% reported weighing costs in clinical decision making (Table 1, Appendix I).

University-based hospitalists had significantly higher scores in leadership and health system messaging (67.4, SD 16.9) than community-based and safety-net-based hospitalists. They reported that their institutions consider their suggestions to improve quality care at low cost (75%), openly discuss ways to deliver this care (64%), and are actively implementing projects (73%). However, only 54% reported seeing success from implemented high-value care efforts (Table 1, Appendix I).

Safety-net hospitalists reported lower scores in leadership and health system messaging (56.8, SD 10.5) than university-based and community-based hospitalists. Few hospitalists reported a work climate (26%) or role modeling (30%) that is supportive of delivering quality care at low costs, openly discusses ways to deliver this care (35%), encourages frontline clinicians to pursue improvement projects (57%), or actively implements projects (26%). They also reported higher scores in the blame-free environment domain (59.8, SD 22.3; Table 1; Appendix 1).

Productivity Adjustments and High-Value Care Culture

In multilevel regression modeling, hospitalists who reported reimbursement with salary plus productivity adjustments had a lower mean HVCCS score (β = −6.2, 95% CI −9.9 to –2.5) than those who reported payment with salary or wages alone. Further multilevel regression modeling for each HVCCS domain revealed that hospitalists who reported reimbursement with salary plus productivity adjustments had lower scores in the leadership and health system messaging domain (β = −4.9, 95% CI −9.3 to −0.6) and data transparency and access domain (β = −10.7, 95% CI −16.7 to −4.6). No statistically significant difference was found between hospitalists who reported reimbursement with quality or value adjustments.

Understanding the drivers that are associated with a high-value care culture is necessary as payment models for hospitals transition from volume-based to value-based care. In this study, we found a meaningful association (β = −6.2) between clinician reimbursement schemes and measures of high-value care culture. A six-point change in the HVCCS score would correspond with a hospital moving from the top quartile to the median, which represents a significant change in performance. The relationship between clinician reimbursement schemes and high-value care culture may be a bidirectional relationship. Fee for service, the predominant payment scheme, places pressure on clinicians to maximize volume, focus on billing, and provide reactive care.^7,29 Conversely, payment schemes that avoid these incentives (ie, salary, wages, and adjustments for quality or value), especially if incentives are felt by frontline clinicians, may better align with goals for long-term health outcomes for patient populations and reduce excess visits and services.^2-6,8,30-34 At the same time, hospitals with a strong high-value care culture may be more likely to introduce shared savings programs and alternative payment models than those without. Through these decisions, the leadership can play an important role in creating an environment for change.³⁴ Similar to the study sites, hospitals in California have a higher percentage of risk-based payments than hospitals in other states (>22%)³⁵ and may also provide incentives to promote a high-value care culture or affect local physician compensation models.

Hospitals have options in how they choose to pay their clinicians, and these decisions may have downstream effects, such as building or eroding high-value care culture among clinicians or staff. A dose-response relationship between physician compensation models and value culture is plausible (salary with productivity < salary only < salary with value incentive). However, we did not find a statistically significant difference for salary with value incentive. This result may be attributed to the relatively small sample size in this study.

Hospitals can also improve their internal processes, organizational structure, and align their institutional payment contracts with those that emphasize value over fee-for-service-based incentives to increase value in care delivery.³⁶ The operation of hospitals is challenging when competing payment incentives are used at the same time,⁷ and leadership will likely achieve more success in improving a high-value care culture and value performance when all efforts, including clinician and institutional payment, are aligned.^37-38

Enduring large systems redesign will require directing attention to local organizational culture. For the majority of individual HVCCS items, 30% or more hospitalists across all sites agreed or strongly agreed that components of low-value care culture exist within their institutions. This response demonstrates a lack of focus on culture to address high-value care improvement among the study sites. Division and program leaders can begin measuring culture within their groups to develop new interventions that target culture change and improve value.³⁴ No single panacea exists for the value improvement of hospitalist programs in California across all hospital types and sites.

Unique trends, however, emerge among each hospital type that could direct future improvements. In addition to all sites requiring increased transparency and access to data, community-based hospitalists identified the need for improvement in the creation of a blame-free environment, comfort with cost conversations, and aspects of leadership and health system messaging. While a high proportion of these hospitalists reported a work culture and role modeling that support the delivery of quality care at low costs, opportunities to create open discussion and frontline involvement in improvement efforts, weigh costs into clinical decision making, and cost conversations with patients exist. We hypothesize that these opportunities exist because community-based hospitals create infrastructure and technology to drive improvement that is often unseen by frontline providers. University-based hospitalists performed higher on three of the four domains compared with their counterparts but may have opportunities to promote a blame-free environment. A great proportion of these hospitalists reported the occurrence of open discussion and active projects within their institutions but also identified opportunities for the improvement of project implementation. Safety-net hospitalists reported the need to improve leadership and health system messaging across most domain items. Further study is required to evaluate reasons for safety-net hospitalists’ responses. We hypothesize that these responses may be related to having limited institutional resources to provide data and coordinated care and different institutional payment models. Each of these sites could identify trends in specific questions identified by the HVCCS for improvement in the high-value care culture.²⁵

Our study evaluated 12 hospitalist programs in California that represent hospitals of different sizes and institutional VBP performance. A large multisite study that evaluates HVCCS across other specialties and disciplines in medicine, all regions of the country, and ambulatory care settings may be conducted in the future. Community-based hospitalist programs also reported low mean HVCCS scores, and further studies could better understand this relationship.

The limitations of the study include its small subgroup sample size and the lack of a gold standard for the measurement of high-value care. As expected, hospitalist groups among safety-net hospitals in California are small, and we may have been underpowered to determine some correlations presented by safety-net sites when stratifying by hospital type. Other correlations also may have been limited by sample size, including differences in HVCCS scores based on reimbursement and hospital type and the correlation between a blame-free environment and reimbursement type. Additionally, the field lacks a gold standard for the measurement of high-value care to help stratify institutional value performance for site selection. The VBP measure presents policy implications and is currently the best available measure with recent value data for over 3,000 hospitals nationally and representing various types of hospitals. This study is also cross-sectional and may benefit from the further evaluation of organizational culture over time and across other settings.

The HVCCS can identify clear targets for improvement and has been evaluated among internal medicine hospitalists. Hospitalists who are paid partly based on productivity reported low measures of high-value care culture at their institutions. As the nation moves toward increasingly value-based payment models, hospitals can strive to improve their understanding of their individual culture for value and begin addressing gaps.

Acknowledgments

The authors wish to thank Michael Lazarus, MD from the University of California Los Angeles; Robert Wachter, MD, James Harrison, PhD; Victoria Valencia, MPH from Dell Medical School at the University of Texas at Austin; Mithu Molla, MD from University of California Davis; Gregory Seymann, MD from the University of California San Diego; Bindu Swaroop, MD and Alpesh Amin, MD from University of California Irvine; Jessica Murphy, DO and Danny Sam, MD from Kaiser Permanente Santa Clara; Thomas Baudendistel, MD and Rajeeva Ranga, MD from Kaiser Permanente Oakland; Yile Ding, MD from California Pacific Medical Center; Soma Wali, MD from Los Angeles County/ OliveView UCLA Medical Center; Anshu Abhat, MD, MPH from the LA BioMed Institute at Los Angeles County/ Harbor-UCLA Medical Center; Steve Tringali, MD from Community Regional Medical Center Fresno; and Dan Dworsky, MD from Scripps Green Hospital for their site leadership and participation with the study.

Disclosures

Dr. Gupta is the Director of the Teaching Value in Healthcare Learning Network at Costs of Care. Dr. Moriates receives royalties from McGraw Hill for the textbook “Understanding Value-based Healthcare” outside of the submitted work and is the Director of Implementation at Costs of Care.

The Centers of Medicare and Medicaid Services (CMS) has introduced new payment models that tie quality and value incentives to 90% of fee-for-service payments and provide 50% of Medicare payments through alternative payment models.¹ The push toward value comes after productivity-based physician reimbursement (ie, fee for service) has been associated with poor quality care, including delayed diagnoses, complications, readmissions, increased length of stay, and high costs of care.^2-5 The method of physician payment is widely believed to affect clinical behavior by incentivizing doing more, coding for more, and billing for more.^6-7 Although payment systems may be used to achieve policy objectives,⁸ little is known about the association of different payment systems with the culture of delivering value-based care among frontline clinicians.

Culture is defined as a system of shared assumptions, values, beliefs, and norms within an environment and has a powerful role in shaping clinician practice patterns.^9-12 The culture within medicine currently contributes to the overuse of resources,^11,13 and a culture for improvement is correlated with clinical outcomes. A systematic review found a consistent association between positive organization culture and improved outcomes including mortality.¹⁴ Across health systems, institutions with high scores on patient safety culture surveys have shown improvements in clinical behaviors and patient outcomes.^15-18

In this study, we aim to describe high-value care culture among internal medicine hospitalists across diverse hospitals and evaluate the relationship between physician reimbursement and high-value care culture.

Study Design

This study is an observational, cross-sectional survey-based study of hospitalists from 12 hospitals in California between January and June 2016.

Study Population

A total of 12 hospitals with hospitalist programs in California were chosen to represent three types of hospitals (ie, four university, four community, and four safety net). Safety-net hospitals, which traditionally serve low-income and medically and socially vulnerable patients were defined as those in the top quartile (ie, greater than 0.5) of their Disproportionate Share Index (DSH), which measures Medicaid patient load.^19-20

To select hospitals with varying value-based care performance, we stratified using CMS value-based purchasing (VBP) scores from fiscal year 2015; these scores have been used to adjust reimbursement for just over 3,000 hospitals in the VBP program of CMS.^22,23 CMS calculates the VBP total performance score as a composite of four domains: (1) clinical processes of care (20% of total performance); (2) patient satisfaction (30%); (3) patient outcomes, including mortality and complications (30%); and (4) cost defined by Medicare payment per beneficiary (20%).²¹ Established quality measures are based on data reported by participating hospitals and chart abstraction during 2011-2014.²² Although other clinical measures of care intensity have been used as proxies of value-based care,^23,24 we used the measure of value that has been publically reported by the CMS VBP given its wide use and effects on reimbursements for 80% of hospitals in the CMS VBP program in 2015.²⁵ We obtained institution-level data from the CMS VBP Program and Hospital Compare files. Each of the three types of hospitals represented institutions with low, middle, and high VBP performance (split in tertiles) as reported by the CMS VBP program. To increase the number of participants in tertiles with fewer hospitalists, a fourth hospital was selected for each hospital type.

We excluded individual hospitalists who primarily identified as working in subspecialty divisions and those who spent less than eight weeks during the last year providing direct patient care on inpatient internal medicine services at the studied institution.

Measurement

Hospitalists were asked to complete the High-Value Care Culture Survey (HVCCS^TM), which measures the culture of value-based decision making among frontline clinicians.²⁶ Similar to other validated surveys for the assessment of patient safety culture,^27,28 the HVCCS can be used to identify target areas for improvement. The survey includes four domains: (1) leadership and health system messaging, (2) data transparency and access, (3) comfort with cost conversations, and (4) blame-free environment. This tool was developed by using a two-phase national modified Delphi process. It was evaluated at two academic centers to complete factor analysis and assess internal consistency, reliability, and validity among internal medicine hospitalists and residents. Validation included estimating product-moment correlation of overall HVCCS scores and domain scores with the CMS institutional VBP scores. HVCCS scores are standardized to a 0-100 point scale for each of the four domains and are then averaged to obtain an overall score.²⁶

In the survey, value was defined as the quality of care provided to patients in relation to the costs required to deliver that care, and high-value care was defined as care that tried to maximize quality while minimizing costs. Quality was defined as the degree to which health services increased the likelihood of desired health outcomes that are safe, effective, patient centered, timely, equitable, and consistent with current professional knowledge. Cost was defined as the negative financial, physical, and emotional effects on patients and the health system.²⁶

Data Analysis

We described the overall institutional mean high-value care culture and domain scores measured by the HVCCS, hospitalist demographics and training experiences, and hospital characteristics. We also described individual survey items. Descriptive statistics were stratified and compared on the basis of hospital type (ie, safety net, community, or university). We assessed the relationship between the clinician perception of reimbursement structure within their divisions and individually reported high-value care culture scores using bivariate and multilevel linear regression. We hypothesized that compared with hospitalists who were paid with salaries or wages, those who reported reimbursement with productivity adjustments may report lower HVCCS scores and those who reported reimbursement with quality or value adjustments may report higher HVCCS scores. We adjusted for physician- and hospital-level characteristics, including age, gender, and training track, and considered hospital type and size as random effects.

This study was approved by the Institutional Review Board at all 12 sites. All analyses were conducted using STATA^® 13.0 (College Station, Texas).

Hospitalist Characteristics

A total of 255 (68.9%, 255/370) hospitalists across all sites completed the survey. Of these respondents, 135 were female (50.6%). On average, hospitalists were 39 years of age (SD 6.8), trained in categorical tracks (221; 86.7%), and had previously trained for 14.3 months at a safety-net hospital (SD 14.2). In total, 166 hospitalists (65.1%) reported being paid with salary or wages, 77 (30.2%) with salary plus productivity adjustments, and 12 (4.7%) with salary plus quality or value adjustments. Moreover, 123 (48.6%) hospitalists agreed that funding for their group depended on the volume of services they delivered. Community-based hospitalists reported higher rates of reimbursement with salary plus productivity (47; 32.0%) compared with their counterparts from university-based (24; 28.2%) and safety-net based programs (6; 26.1%). Among the three different hospital types, significant differences exist in hospitalist mean age (P < .001), gender (P = .01), and the number of months training in a safety-net hospital (P = .02; Table 1).

Hospital Characteristics

Of the 12 study sites, four from each type of hospital (ie, safety-net based, community based, and university based) and four representing each value-based purchasing performance tertile (ie, high, middle, and low) were included. Eleven (91.7%) sites were located in urban areas with an average DSH index of 0.40 (SD 0.23), case mix index of 1.97 (SD 0.28), and bed size of 435.5 (SD 146.0; Table 1).

In multilevel regression modeling across all 12 sites, hospitalists from community-based hospitalist programs reported lower mean HVCCS scores (β = −4.4, 95% CI −8.1 to −0.7; Table 2) than those from other hospital types.

High-Value Care Culture Survey Scores

The mean HVCCS score was 50.2 (SD 13.6), and mean domain scores across all sites were 65.4 (SD 15.6) for leadership and health system messaging, 32.4 (SD 22.8) for data transparency and access, 52.1 (SD 19.7) for comfort with cost conversations, and 50.7 (SD 21.4) for blame-free environment (Table 1). For the majority (two-thirds) of individual HVCCS items, more than 30% of hospitalists across all sites agreed or strongly agreed that components of a low-value care culture exist within their institutions. For example, over 80% of hospitalists reported low transparency and limited access to data (see Appendix I for complete survey responses).

Hospitalists reported different HVCCS domains as strengths or weaknesses within their institutions in accordance with hospital type. Compared with university-based and safety-net-based hospitalists, community-based hospitalists reported lower scores in having a blame-free environment (466, SD 21.8). Nearly 50% reported that the clinicians’ fear of legal repercussions affects their frequency of ordering unneeded tests or procedures, and 30% reported that individual clinicians are blamed for complications. Nearly 40% reported that clinicians are uncomfortable discussing the costs of tests or treatments with patients and reported that clinicians do not feel that physicians should discuss costs with patients. Notably, community-based hospitalists uniquely differed in how they reported components of leadership and health system messaging. Over 60% reported a work climate or role modeling supportive of delivering quality care at lower costs. Only 48%, however, reported success seen from implemented efforts, and 45% reported weighing costs in clinical decision making (Table 1, Appendix I).

University-based hospitalists had significantly higher scores in leadership and health system messaging (67.4, SD 16.9) than community-based and safety-net-based hospitalists. They reported that their institutions consider their suggestions to improve quality care at low cost (75%), openly discuss ways to deliver this care (64%), and are actively implementing projects (73%). However, only 54% reported seeing success from implemented high-value care efforts (Table 1, Appendix I).

Safety-net hospitalists reported lower scores in leadership and health system messaging (56.8, SD 10.5) than university-based and community-based hospitalists. Few hospitalists reported a work climate (26%) or role modeling (30%) that is supportive of delivering quality care at low costs, openly discusses ways to deliver this care (35%), encourages frontline clinicians to pursue improvement projects (57%), or actively implements projects (26%). They also reported higher scores in the blame-free environment domain (59.8, SD 22.3; Table 1; Appendix 1).

Productivity Adjustments and High-Value Care Culture

In multilevel regression modeling, hospitalists who reported reimbursement with salary plus productivity adjustments had a lower mean HVCCS score (β = −6.2, 95% CI −9.9 to –2.5) than those who reported payment with salary or wages alone. Further multilevel regression modeling for each HVCCS domain revealed that hospitalists who reported reimbursement with salary plus productivity adjustments had lower scores in the leadership and health system messaging domain (β = −4.9, 95% CI −9.3 to −0.6) and data transparency and access domain (β = −10.7, 95% CI −16.7 to −4.6). No statistically significant difference was found between hospitalists who reported reimbursement with quality or value adjustments.

Understanding the drivers that are associated with a high-value care culture is necessary as payment models for hospitals transition from volume-based to value-based care. In this study, we found a meaningful association (β = −6.2) between clinician reimbursement schemes and measures of high-value care culture. A six-point change in the HVCCS score would correspond with a hospital moving from the top quartile to the median, which represents a significant change in performance. The relationship between clinician reimbursement schemes and high-value care culture may be a bidirectional relationship. Fee for service, the predominant payment scheme, places pressure on clinicians to maximize volume, focus on billing, and provide reactive care.^7,29 Conversely, payment schemes that avoid these incentives (ie, salary, wages, and adjustments for quality or value), especially if incentives are felt by frontline clinicians, may better align with goals for long-term health outcomes for patient populations and reduce excess visits and services.^2-6,8,30-34 At the same time, hospitals with a strong high-value care culture may be more likely to introduce shared savings programs and alternative payment models than those without. Through these decisions, the leadership can play an important role in creating an environment for change.³⁴ Similar to the study sites, hospitals in California have a higher percentage of risk-based payments than hospitals in other states (>22%)³⁵ and may also provide incentives to promote a high-value care culture or affect local physician compensation models.

Hospitals have options in how they choose to pay their clinicians, and these decisions may have downstream effects, such as building or eroding high-value care culture among clinicians or staff. A dose-response relationship between physician compensation models and value culture is plausible (salary with productivity < salary only < salary with value incentive). However, we did not find a statistically significant difference for salary with value incentive. This result may be attributed to the relatively small sample size in this study.

Hospitals can also improve their internal processes, organizational structure, and align their institutional payment contracts with those that emphasize value over fee-for-service-based incentives to increase value in care delivery.³⁶ The operation of hospitals is challenging when competing payment incentives are used at the same time,⁷ and leadership will likely achieve more success in improving a high-value care culture and value performance when all efforts, including clinician and institutional payment, are aligned.^37-38

Enduring large systems redesign will require directing attention to local organizational culture. For the majority of individual HVCCS items, 30% or more hospitalists across all sites agreed or strongly agreed that components of low-value care culture exist within their institutions. This response demonstrates a lack of focus on culture to address high-value care improvement among the study sites. Division and program leaders can begin measuring culture within their groups to develop new interventions that target culture change and improve value.³⁴ No single panacea exists for the value improvement of hospitalist programs in California across all hospital types and sites.

Unique trends, however, emerge among each hospital type that could direct future improvements. In addition to all sites requiring increased transparency and access to data, community-based hospitalists identified the need for improvement in the creation of a blame-free environment, comfort with cost conversations, and aspects of leadership and health system messaging. While a high proportion of these hospitalists reported a work culture and role modeling that support the delivery of quality care at low costs, opportunities to create open discussion and frontline involvement in improvement efforts, weigh costs into clinical decision making, and cost conversations with patients exist. We hypothesize that these opportunities exist because community-based hospitals create infrastructure and technology to drive improvement that is often unseen by frontline providers. University-based hospitalists performed higher on three of the four domains compared with their counterparts but may have opportunities to promote a blame-free environment. A great proportion of these hospitalists reported the occurrence of open discussion and active projects within their institutions but also identified opportunities for the improvement of project implementation. Safety-net hospitalists reported the need to improve leadership and health system messaging across most domain items. Further study is required to evaluate reasons for safety-net hospitalists’ responses. We hypothesize that these responses may be related to having limited institutional resources to provide data and coordinated care and different institutional payment models. Each of these sites could identify trends in specific questions identified by the HVCCS for improvement in the high-value care culture.²⁵

Our study evaluated 12 hospitalist programs in California that represent hospitals of different sizes and institutional VBP performance. A large multisite study that evaluates HVCCS across other specialties and disciplines in medicine, all regions of the country, and ambulatory care settings may be conducted in the future. Community-based hospitalist programs also reported low mean HVCCS scores, and further studies could better understand this relationship.

The limitations of the study include its small subgroup sample size and the lack of a gold standard for the measurement of high-value care. As expected, hospitalist groups among safety-net hospitals in California are small, and we may have been underpowered to determine some correlations presented by safety-net sites when stratifying by hospital type. Other correlations also may have been limited by sample size, including differences in HVCCS scores based on reimbursement and hospital type and the correlation between a blame-free environment and reimbursement type. Additionally, the field lacks a gold standard for the measurement of high-value care to help stratify institutional value performance for site selection. The VBP measure presents policy implications and is currently the best available measure with recent value data for over 3,000 hospitals nationally and representing various types of hospitals. This study is also cross-sectional and may benefit from the further evaluation of organizational culture over time and across other settings.

The HVCCS can identify clear targets for improvement and has been evaluated among internal medicine hospitalists. Hospitalists who are paid partly based on productivity reported low measures of high-value care culture at their institutions. As the nation moves toward increasingly value-based payment models, hospitals can strive to improve their understanding of their individual culture for value and begin addressing gaps.

Acknowledgments

The authors wish to thank Michael Lazarus, MD from the University of California Los Angeles; Robert Wachter, MD, James Harrison, PhD; Victoria Valencia, MPH from Dell Medical School at the University of Texas at Austin; Mithu Molla, MD from University of California Davis; Gregory Seymann, MD from the University of California San Diego; Bindu Swaroop, MD and Alpesh Amin, MD from University of California Irvine; Jessica Murphy, DO and Danny Sam, MD from Kaiser Permanente Santa Clara; Thomas Baudendistel, MD and Rajeeva Ranga, MD from Kaiser Permanente Oakland; Yile Ding, MD from California Pacific Medical Center; Soma Wali, MD from Los Angeles County/ OliveView UCLA Medical Center; Anshu Abhat, MD, MPH from the LA BioMed Institute at Los Angeles County/ Harbor-UCLA Medical Center; Steve Tringali, MD from Community Regional Medical Center Fresno; and Dan Dworsky, MD from Scripps Green Hospital for their site leadership and participation with the study.

Disclosures

Dr. Gupta is the Director of the Teaching Value in Healthcare Learning Network at Costs of Care. Dr. Moriates receives royalties from McGraw Hill for the textbook “Understanding Value-based Healthcare” outside of the submitted work and is the Director of Implementation at Costs of Care.

The Centers of Medicare and Medicaid Services (CMS) has introduced new payment models that tie quality and value incentives to 90% of fee-for-service payments and provide 50% of Medicare payments through alternative payment models.¹ The push toward value comes after productivity-based physician reimbursement (ie, fee for service) has been associated with poor quality care, including delayed diagnoses, complications, readmissions, increased length of stay, and high costs of care.^2-5 The method of physician payment is widely believed to affect clinical behavior by incentivizing doing more, coding for more, and billing for more.^6-7 Although payment systems may be used to achieve policy objectives,⁸ little is known about the association of different payment systems with the culture of delivering value-based care among frontline clinicians.

Culture is defined as a system of shared assumptions, values, beliefs, and norms within an environment and has a powerful role in shaping clinician practice patterns.^9-12 The culture within medicine currently contributes to the overuse of resources,^11,13 and a culture for improvement is correlated with clinical outcomes. A systematic review found a consistent association between positive organization culture and improved outcomes including mortality.¹⁴ Across health systems, institutions with high scores on patient safety culture surveys have shown improvements in clinical behaviors and patient outcomes.^15-18

In this study, we aim to describe high-value care culture among internal medicine hospitalists across diverse hospitals and evaluate the relationship between physician reimbursement and high-value care culture.

Study Design

This study is an observational, cross-sectional survey-based study of hospitalists from 12 hospitals in California between January and June 2016.

Study Population

A total of 12 hospitals with hospitalist programs in California were chosen to represent three types of hospitals (ie, four university, four community, and four safety net). Safety-net hospitals, which traditionally serve low-income and medically and socially vulnerable patients were defined as those in the top quartile (ie, greater than 0.5) of their Disproportionate Share Index (DSH), which measures Medicaid patient load.^19-20

To select hospitals with varying value-based care performance, we stratified using CMS value-based purchasing (VBP) scores from fiscal year 2015; these scores have been used to adjust reimbursement for just over 3,000 hospitals in the VBP program of CMS.^22,23 CMS calculates the VBP total performance score as a composite of four domains: (1) clinical processes of care (20% of total performance); (2) patient satisfaction (30%); (3) patient outcomes, including mortality and complications (30%); and (4) cost defined by Medicare payment per beneficiary (20%).²¹ Established quality measures are based on data reported by participating hospitals and chart abstraction during 2011-2014.²² Although other clinical measures of care intensity have been used as proxies of value-based care,^23,24 we used the measure of value that has been publically reported by the CMS VBP given its wide use and effects on reimbursements for 80% of hospitals in the CMS VBP program in 2015.²⁵ We obtained institution-level data from the CMS VBP Program and Hospital Compare files. Each of the three types of hospitals represented institutions with low, middle, and high VBP performance (split in tertiles) as reported by the CMS VBP program. To increase the number of participants in tertiles with fewer hospitalists, a fourth hospital was selected for each hospital type.

We excluded individual hospitalists who primarily identified as working in subspecialty divisions and those who spent less than eight weeks during the last year providing direct patient care on inpatient internal medicine services at the studied institution.

Measurement

Hospitalists were asked to complete the High-Value Care Culture Survey (HVCCS^TM), which measures the culture of value-based decision making among frontline clinicians.²⁶ Similar to other validated surveys for the assessment of patient safety culture,^27,28 the HVCCS can be used to identify target areas for improvement. The survey includes four domains: (1) leadership and health system messaging, (2) data transparency and access, (3) comfort with cost conversations, and (4) blame-free environment. This tool was developed by using a two-phase national modified Delphi process. It was evaluated at two academic centers to complete factor analysis and assess internal consistency, reliability, and validity among internal medicine hospitalists and residents. Validation included estimating product-moment correlation of overall HVCCS scores and domain scores with the CMS institutional VBP scores. HVCCS scores are standardized to a 0-100 point scale for each of the four domains and are then averaged to obtain an overall score.²⁶

In the survey, value was defined as the quality of care provided to patients in relation to the costs required to deliver that care, and high-value care was defined as care that tried to maximize quality while minimizing costs. Quality was defined as the degree to which health services increased the likelihood of desired health outcomes that are safe, effective, patient centered, timely, equitable, and consistent with current professional knowledge. Cost was defined as the negative financial, physical, and emotional effects on patients and the health system.²⁶

Data Analysis

We described the overall institutional mean high-value care culture and domain scores measured by the HVCCS, hospitalist demographics and training experiences, and hospital characteristics. We also described individual survey items. Descriptive statistics were stratified and compared on the basis of hospital type (ie, safety net, community, or university). We assessed the relationship between the clinician perception of reimbursement structure within their divisions and individually reported high-value care culture scores using bivariate and multilevel linear regression. We hypothesized that compared with hospitalists who were paid with salaries or wages, those who reported reimbursement with productivity adjustments may report lower HVCCS scores and those who reported reimbursement with quality or value adjustments may report higher HVCCS scores. We adjusted for physician- and hospital-level characteristics, including age, gender, and training track, and considered hospital type and size as random effects.

This study was approved by the Institutional Review Board at all 12 sites. All analyses were conducted using STATA^® 13.0 (College Station, Texas).

Hospitalist Characteristics

A total of 255 (68.9%, 255/370) hospitalists across all sites completed the survey. Of these respondents, 135 were female (50.6%). On average, hospitalists were 39 years of age (SD 6.8), trained in categorical tracks (221; 86.7%), and had previously trained for 14.3 months at a safety-net hospital (SD 14.2). In total, 166 hospitalists (65.1%) reported being paid with salary or wages, 77 (30.2%) with salary plus productivity adjustments, and 12 (4.7%) with salary plus quality or value adjustments. Moreover, 123 (48.6%) hospitalists agreed that funding for their group depended on the volume of services they delivered. Community-based hospitalists reported higher rates of reimbursement with salary plus productivity (47; 32.0%) compared with their counterparts from university-based (24; 28.2%) and safety-net based programs (6; 26.1%). Among the three different hospital types, significant differences exist in hospitalist mean age (P < .001), gender (P = .01), and the number of months training in a safety-net hospital (P = .02; Table 1).

Hospital Characteristics

Of the 12 study sites, four from each type of hospital (ie, safety-net based, community based, and university based) and four representing each value-based purchasing performance tertile (ie, high, middle, and low) were included. Eleven (91.7%) sites were located in urban areas with an average DSH index of 0.40 (SD 0.23), case mix index of 1.97 (SD 0.28), and bed size of 435.5 (SD 146.0; Table 1).

In multilevel regression modeling across all 12 sites, hospitalists from community-based hospitalist programs reported lower mean HVCCS scores (β = −4.4, 95% CI −8.1 to −0.7; Table 2) than those from other hospital types.

High-Value Care Culture Survey Scores

The mean HVCCS score was 50.2 (SD 13.6), and mean domain scores across all sites were 65.4 (SD 15.6) for leadership and health system messaging, 32.4 (SD 22.8) for data transparency and access, 52.1 (SD 19.7) for comfort with cost conversations, and 50.7 (SD 21.4) for blame-free environment (Table 1). For the majority (two-thirds) of individual HVCCS items, more than 30% of hospitalists across all sites agreed or strongly agreed that components of a low-value care culture exist within their institutions. For example, over 80% of hospitalists reported low transparency and limited access to data (see Appendix I for complete survey responses).

Hospitalists reported different HVCCS domains as strengths or weaknesses within their institutions in accordance with hospital type. Compared with university-based and safety-net-based hospitalists, community-based hospitalists reported lower scores in having a blame-free environment (466, SD 21.8). Nearly 50% reported that the clinicians’ fear of legal repercussions affects their frequency of ordering unneeded tests or procedures, and 30% reported that individual clinicians are blamed for complications. Nearly 40% reported that clinicians are uncomfortable discussing the costs of tests or treatments with patients and reported that clinicians do not feel that physicians should discuss costs with patients. Notably, community-based hospitalists uniquely differed in how they reported components of leadership and health system messaging. Over 60% reported a work climate or role modeling supportive of delivering quality care at lower costs. Only 48%, however, reported success seen from implemented efforts, and 45% reported weighing costs in clinical decision making (Table 1, Appendix I).

University-based hospitalists had significantly higher scores in leadership and health system messaging (67.4, SD 16.9) than community-based and safety-net-based hospitalists. They reported that their institutions consider their suggestions to improve quality care at low cost (75%), openly discuss ways to deliver this care (64%), and are actively implementing projects (73%). However, only 54% reported seeing success from implemented high-value care efforts (Table 1, Appendix I).

Safety-net hospitalists reported lower scores in leadership and health system messaging (56.8, SD 10.5) than university-based and community-based hospitalists. Few hospitalists reported a work climate (26%) or role modeling (30%) that is supportive of delivering quality care at low costs, openly discusses ways to deliver this care (35%), encourages frontline clinicians to pursue improvement projects (57%), or actively implements projects (26%). They also reported higher scores in the blame-free environment domain (59.8, SD 22.3; Table 1; Appendix 1).

Productivity Adjustments and High-Value Care Culture

In multilevel regression modeling, hospitalists who reported reimbursement with salary plus productivity adjustments had a lower mean HVCCS score (β = −6.2, 95% CI −9.9 to –2.5) than those who reported payment with salary or wages alone. Further multilevel regression modeling for each HVCCS domain revealed that hospitalists who reported reimbursement with salary plus productivity adjustments had lower scores in the leadership and health system messaging domain (β = −4.9, 95% CI −9.3 to −0.6) and data transparency and access domain (β = −10.7, 95% CI −16.7 to −4.6). No statistically significant difference was found between hospitalists who reported reimbursement with quality or value adjustments.

Understanding the drivers that are associated with a high-value care culture is necessary as payment models for hospitals transition from volume-based to value-based care. In this study, we found a meaningful association (β = −6.2) between clinician reimbursement schemes and measures of high-value care culture. A six-point change in the HVCCS score would correspond with a hospital moving from the top quartile to the median, which represents a significant change in performance. The relationship between clinician reimbursement schemes and high-value care culture may be a bidirectional relationship. Fee for service, the predominant payment scheme, places pressure on clinicians to maximize volume, focus on billing, and provide reactive care.^7,29 Conversely, payment schemes that avoid these incentives (ie, salary, wages, and adjustments for quality or value), especially if incentives are felt by frontline clinicians, may better align with goals for long-term health outcomes for patient populations and reduce excess visits and services.^2-6,8,30-34 At the same time, hospitals with a strong high-value care culture may be more likely to introduce shared savings programs and alternative payment models than those without. Through these decisions, the leadership can play an important role in creating an environment for change.³⁴ Similar to the study sites, hospitals in California have a higher percentage of risk-based payments than hospitals in other states (>22%)³⁵ and may also provide incentives to promote a high-value care culture or affect local physician compensation models.

Hospitals have options in how they choose to pay their clinicians, and these decisions may have downstream effects, such as building or eroding high-value care culture among clinicians or staff. A dose-response relationship between physician compensation models and value culture is plausible (salary with productivity < salary only < salary with value incentive). However, we did not find a statistically significant difference for salary with value incentive. This result may be attributed to the relatively small sample size in this study.

Hospitals can also improve their internal processes, organizational structure, and align their institutional payment contracts with those that emphasize value over fee-for-service-based incentives to increase value in care delivery.³⁶ The operation of hospitals is challenging when competing payment incentives are used at the same time,⁷ and leadership will likely achieve more success in improving a high-value care culture and value performance when all efforts, including clinician and institutional payment, are aligned.^37-38

Enduring large systems redesign will require directing attention to local organizational culture. For the majority of individual HVCCS items, 30% or more hospitalists across all sites agreed or strongly agreed that components of low-value care culture exist within their institutions. This response demonstrates a lack of focus on culture to address high-value care improvement among the study sites. Division and program leaders can begin measuring culture within their groups to develop new interventions that target culture change and improve value.³⁴ No single panacea exists for the value improvement of hospitalist programs in California across all hospital types and sites.

Unique trends, however, emerge among each hospital type that could direct future improvements. In addition to all sites requiring increased transparency and access to data, community-based hospitalists identified the need for improvement in the creation of a blame-free environment, comfort with cost conversations, and aspects of leadership and health system messaging. While a high proportion of these hospitalists reported a work culture and role modeling that support the delivery of quality care at low costs, opportunities to create open discussion and frontline involvement in improvement efforts, weigh costs into clinical decision making, and cost conversations with patients exist. We hypothesize that these opportunities exist because community-based hospitals create infrastructure and technology to drive improvement that is often unseen by frontline providers. University-based hospitalists performed higher on three of the four domains compared with their counterparts but may have opportunities to promote a blame-free environment. A great proportion of these hospitalists reported the occurrence of open discussion and active projects within their institutions but also identified opportunities for the improvement of project implementation. Safety-net hospitalists reported the need to improve leadership and health system messaging across most domain items. Further study is required to evaluate reasons for safety-net hospitalists’ responses. We hypothesize that these responses may be related to having limited institutional resources to provide data and coordinated care and different institutional payment models. Each of these sites could identify trends in specific questions identified by the HVCCS for improvement in the high-value care culture.²⁵

Our study evaluated 12 hospitalist programs in California that represent hospitals of different sizes and institutional VBP performance. A large multisite study that evaluates HVCCS across other specialties and disciplines in medicine, all regions of the country, and ambulatory care settings may be conducted in the future. Community-based hospitalist programs also reported low mean HVCCS scores, and further studies could better understand this relationship.

The limitations of the study include its small subgroup sample size and the lack of a gold standard for the measurement of high-value care. As expected, hospitalist groups among safety-net hospitals in California are small, and we may have been underpowered to determine some correlations presented by safety-net sites when stratifying by hospital type. Other correlations also may have been limited by sample size, including differences in HVCCS scores based on reimbursement and hospital type and the correlation between a blame-free environment and reimbursement type. Additionally, the field lacks a gold standard for the measurement of high-value care to help stratify institutional value performance for site selection. The VBP measure presents policy implications and is currently the best available measure with recent value data for over 3,000 hospitals nationally and representing various types of hospitals. This study is also cross-sectional and may benefit from the further evaluation of organizational culture over time and across other settings.

The HVCCS can identify clear targets for improvement and has been evaluated among internal medicine hospitalists. Hospitalists who are paid partly based on productivity reported low measures of high-value care culture at their institutions. As the nation moves toward increasingly value-based payment models, hospitals can strive to improve their understanding of their individual culture for value and begin addressing gaps.

Acknowledgments

The authors wish to thank Michael Lazarus, MD from the University of California Los Angeles; Robert Wachter, MD, James Harrison, PhD; Victoria Valencia, MPH from Dell Medical School at the University of Texas at Austin; Mithu Molla, MD from University of California Davis; Gregory Seymann, MD from the University of California San Diego; Bindu Swaroop, MD and Alpesh Amin, MD from University of California Irvine; Jessica Murphy, DO and Danny Sam, MD from Kaiser Permanente Santa Clara; Thomas Baudendistel, MD and Rajeeva Ranga, MD from Kaiser Permanente Oakland; Yile Ding, MD from California Pacific Medical Center; Soma Wali, MD from Los Angeles County/ OliveView UCLA Medical Center; Anshu Abhat, MD, MPH from the LA BioMed Institute at Los Angeles County/ Harbor-UCLA Medical Center; Steve Tringali, MD from Community Regional Medical Center Fresno; and Dan Dworsky, MD from Scripps Green Hospital for their site leadership and participation with the study.

Disclosures

Dr. Gupta is the Director of the Teaching Value in Healthcare Learning Network at Costs of Care. Dr. Moriates receives royalties from McGraw Hill for the textbook “Understanding Value-based Healthcare” outside of the submitted work and is the Director of Implementation at Costs of Care.

Patient flow throughout the hospital has been shown to be adversely affected by discharge delays.¹ When hospitals are operating at peak capacity, these delays impact throughput, length of stay (LOS), and cost of care and block patients from the emergency department (ED), postanesthesia recovery unit (PACU), or home awaiting inpatient beds.^2-5 As patients wait in locations not ideal for inpatient care, they may suffer from adverse events and poor satisfaction.^3,6 Several studies have analyzed discharge timing as it relates to ED boarding of admitted patients and demonstrated that early discharges (EDCs) can impact boarding times.^7-9 A number of recent improvement efforts directed at moving discharges earlier in the day have been published.^10-15 However, these improvements are often targeted at specific units or teams within a larger hospital setting and only one is in the pediatric setting.

Lucile Packard Children’s Hospital Stanford (LPCHS) is a 311-bed quaternary care academic women and children’s hospital in Northern California. As our organization expanded, the demand for hospital beds often exceeded capacity. The challenge of overall demand was regularly compounded by a mismatch in bed availability timing – bed demand is early in the day and bed availability is later. This mismatch results in delays for admitted patients waiting in the ED and PACU. Organization leaders identified increasing early discharges (EDCs) as one initiative to contribute to improved patient flow.

Our organization aimed to increase the number of discharges before 11 am across the acute care units from an average of 8% in the 17 months prior to May 2015 to 25% by December 2016. Based on the average number and timing of planned admissions, they hypothesized that 25% of EDCs would decrease ED and PACU wait times.

Setting

We focused our EDC interventions on the 87 acute care beds at LPCHS. All patients discharged from these beds were included in the study. We excluded patients discharged from intensive care, maternity, and nursery. Acute care includes five units, one focused on hematology/oncology (Unit A), one focused on cardiology (Unit B), and the others with a surgical and medical pediatric patient mix (Units C, D, and E). Although physician teams have primary units, due to unit size, patients on teams other than cardiology and hematology/oncology are often spread across multiple units wherever there is a bed (including Units A and B). Most of the frontline care physicians are residents supervised by attendings; however, a minority of patients are cared for by nurse practitioners (NPs) or physician assistants (PAs).

Improvement Team

In early 2015, we formed a multidisciplinary group inclusive of a case manager, frontline nurses, nurse management, pediatric residents, and hospitalist physicians with support from performance improvement. We periodically included physician leaders from other specialties to help initiate changes within their own clinical areas. Our group used Lean A3 thinking¹⁶ to gather information about the current state, formulate the problem statement, analyze the problem, and consider interventions implemented in three Plan–Do-Check-Act (PDCA) cycles. The A3 is a structured tool to analyze problems before jumping to solutions and communicate with stakeholders. We interviewed leaders, nurses, residents, case managers, etc. and observed work processes around discharge. We met weekly to follow data, assess results of interventions, and problem solve.

Barriers and Interventions

The first barrier we identified and addressed was poor identification and shared team mental model of potential EDC patients and lack of preparation when an EDC was identified. In intervention one starting May 2015, charge nurses on Units C, D, and E were each asked to identify one EDC for the following day. The identified patient was discussed at the previously existing afternoon daily unit huddle¹⁷ attended by nurse management, case management, and hospitalist leaders. Following the huddle, the resident, NP, or PA responsible for the patient was paged regarding the EDC plan and tasked with medication reconciliation and discharge paperwork. Others were asked to address their specific area of patient care for discharge (eg, case manager–supplies, nursing–education). The patient was identified on the unit white board with a yellow magnet (use of a visual control¹⁸), so that all would be aware of the EDC. An e-mail was sent to case management, nurse leaders, and patient placement coordinators regarding the planned EDCs. Finally, the EDCs were discussed during regularly scheduled huddles throughout the evening and into the next day.¹⁷

Despite this first intervention, we noted that progress toward increased EDCs was slow. Thus, we spent approximately seven days (spread over one month) further observing the work processes.¹⁹ Over five days, we asked each unit’s charge nurse every hour which patients were waiting to be discharged and the primary reason for waiting. From this information, we created a pareto chart demonstrating that rounds were the highest contributor to waiting (Appendix A). Thus, our second intervention was a daily physician morning huddle that the four nonsurgical physician teams (excluding cardiology, hematology/oncology) implemented one team at a time between November 2015 and February 2016. At the huddle, previously identified EDCs (located on any of the five units) were confirmed and preparatory work was completed (inclusive of the discharge order) before rounds. Further, the attending and resident physicians were to see the patient before or at the start of rounds.

Our working group still observed slow EDC improvement and sought feedback from all providers. EDC was described as “extra” work, apart from routine practices and culture. In addition, our interventions had not addressed most discharges on Units A and B. Consequently, our third intervention in February 2016 aimed to recognize and incentivize teams, units, and individuals for EDC successes. Units and/or physician teams that met 25% of EDCs the previous week were acknowledged through hospital-wide screensavers and certificates of appreciation signed by the Chief Nursing Officer. Units and/or physician teams that met 25% of EDC the previous month were acknowledged with a trophy. Residents received coffee cards for each EDC (though not without controversy among the improvement group as we acknowledged that all providers contributed to EDCs). Finally, weekly, we shared an EDC dashboard displaying unit, team, and organization performance at the hospital-wide leader huddle. We also e-mailed the dashboard regularly to division chiefs, medical directors, and nursing leaders.

Measures

Our primary outcome was percentage of EDCs (based on the time the patient left the room) across acute care. Secondary outcome measures were median wait times for an inpatient bed from the ED (time bed requested to the time patient left the ED) and the average PACU wait time (time the patient is ready to leave the PACU to time the patient left the PACU) per admitted patient. We also assessed balancing measures, including discharge satisfaction, seven-day readmission rates, and LOS. We obtained the mean discharge satisfaction score from the organization’s Press Ganey survey results across acute care (the three discharge questions’ mean – “degree … you felt ready to have your child discharged,” “speed of discharge process …,” and “instructions… to care for your child…”). We obtained seven-day readmission rates from acute care discharges using the hospital’s regularly reported data. We assessed patient characteristics, including sex, age, case mix index (CMI; >2 vs <2), insurance type (nongovernment vs government), day of discharge (weekend vs weekday), and LOS from those patients categorized as inpatients. Complete patient characteristics were not available for observation (InterQual^® criteria) status patients.

Analysis

We used descriptive statistics to describe the inpatient population characteristics by analyzing differences when EDC did and did not occur using chi-square and the Mann–Whitney U tests. Patients with missing data were removed from analyses that incorporated patient factors.

To assess our primary outcome, we used an interrupted time series analysis assessing the percentage of EDC in the total population before any intervention (May 2015) and after the last intervention (March 2016). We used the Durbin–Watson statistic to assess autocorrelation of errors in our regression models. As we had only patient characteristics for the inpatient population, we repeated the analysis including only inpatients and accounting for patient factors significantly associated with EDC.

As units and physician teams had differential exposure to the interventions, we performed a subanalysis (using interrupted time series) creating groups based on the combination of interventions to which a patient’s discharge was exposed (based on unit and physician team at discharge). Patient discharges from group 1 (medical patients on Units C, D, and E) were exposed to all three interventions, group 2 patient discharges (medical patients on Units A and B) were exposed to interventions 2 and 3, group 3 (cardiology, hematology/oncology, surgical patients on Units A and B) were exposed to intervention 3, and group 4 (surgical, cardiology, hematology/oncology patients on Units C, D, and E) were exposed to interventions 1 and 3 (Figure 1). Interrupted time series models were fit using the R Statistical Software Package.²⁰

There were 16,175 discharges on acute care from January 2014 through December 2016. Across all acute care units, EDCs increased from an average of 8.8% before the start of interventions (May 2015) to 15.8% after all interventions (February 2016). From the estimated trend in the preintervention period, there was a jump of 3.9% to the start of the postintervention trend (P = .02; Figure 2). Furthermore, there was an increase of 0.48% (95% CI 0.15-0.82%; P < .01) per month in the trend of the slope between the pre- and postintervention. The autocorrelation function and the Durbin–Watson test did not show evidence of autocorrelation (P = .85). Lack of evidence for autocorrelation in this and each of our subsequent fitted models led to excluding an autocorrelation parameter from our models.

From 16,175 discharges, 1,764 (11%) were assigned to observation status. Among inpatients (14,411), patients with missing values (CMI, insurance status) were also excluded (n = 66, 0.5%). Among the remaining 14,345 inpatients, 54% were males, 50% were government-insured, and 1,645 (11.5%) were discharged early. The average age was 8.5 years, the average LOS was seven days, and the median CMI was 2.2. Children who were younger, had shorter LOS, CMI <2, and nongovernment insurance were more likely to be discharged early (P < .01 for all). For each of these variables, F-tests were performed to determine whether there was a statistically significant reduction in variation by adding the variable to our initial model. None of the variables alone or in combination led to a statistically significant reduction in variation. Including these factors in the interrupted time series did not change the significance of the results (jump at postintervention start 3.6%, 95% CI 0.7%-7.2%; P = .02, slope increased by 0.59% per month, 95% CI 0.29-0.89%; P < .01).

In the subgroup analysis, we did not account for patient factors as they did not change the results in the analysis of total population. Though each group had a greater percentage of EDCs in the postintervention period, the changes in slopes and jumps were primarily nonsignificant (Figure 3). Only the change in slope in group 4 was significant (1.1%, 95% CI 0.3-1.9%; P = .01).

The percentage of patients who left the hospital prior to 11 am significantly improved after a number of interventions aimed at emphasizing EDC and discharge task completion earlier within the hospital stay. Our EDC improvement was associated with improved ED and PACU wait times without negatively impacting discharge satisfaction, seven-day readmissions, or LOS.

It is difficult to compare our EDC improvements to those of previous studies, as we are unaware of published data on pediatric EDC efforts across an entire hospital. In addition, studies have reported discharges prior to different times in the day (noon, 1 pm, etc).^{12, 13} Our interventions were similar to those of Wertheimer et al.,¹¹ including the use of interdisciplinary rounds, identification of potential EDCs the afternoon before discharge, and “reward and recognition.” Wertheimer also sent an e-mail about EDCs to a multidisciplinary group, which was then updated as conditions changed. Unlike Wertheimer, we did not include physicians in our e-mail due to the large number and frequently changing physician teams. Our EDC rate prior to 11 am was lower than their achieved rate of 35% prior to 12 pm. When we assessed our discharges using 12 pm, our rate was still lower (22%-28%), but a direct comparison was complicated by different patient populations. Still, our study adds to the evidence that interdisciplinary rounds and reward and recognition lead to earlier discharge. In addition, this study builds upon Wertheimer’s results as although they later assessed the timing of ED admissions as a result of their EDC improvements, they did not directly assess inpatient bed wait times as we did in our study.¹⁴

As providers of all types were aware of the constant push for beds due to canceled surgeries, delayed admissions and intensive care transfers, and the inability to accept admission, it is difficult to compare the subgroups directly. Furthermore, although physician teams and units are distinct, individuals (nurses, case managers, trainees) may rotate through different units and teams and we cannot account for individual influences on EDCs depending on exposure to interventions over time. Although all groups improved, the improvement in slope in group 4 (exposed to interventions 1 and 3) was the only significant change. As group 4 contained a large number of surgical patients who often have more predictable hospital stays, perhaps this group was more responsive to the interventions.

Our EDC improvements were associated with a decrease in ED and PACU bed wait times. Importantly, we did not address potential confounding factors impacting these times such as total hospital admission volumes, ED and PACU patient complexity, and distribution of ED and PACU admission requests throughout the day. Modeling has suggested that EDCs could also improve ED flow,⁷ but studies implementing EDC have not necessarily assessed this outcome.^10-15 One study retrospectively evaluated ED boarding times in the context of an EDC improvement effort and found a decrease in boarding times.²¹ This decrease is important as ED boarders may be at a higher risk for adverse events, a longer LOS, and more readmissions.^3,7 Less is known about prolonged PACU wait times; however, studies have reported delays in receiving patients from the operating room (OR), which could presumably impact timeliness of other scheduled procedures and patient satisfaction.^22-24 It is worth noting that OR holds as a result of PACU backups happened more frequently at our institution before our EDC work.

Our limitations include that individual providers in the various groups were not completely blind to the interventions and groups often comprised distinct patient populations. Second, LPCHS has a high CMI and LOS relative to most other children’s hospitals, complicating comparison with patient populations at other children’s hospitals. In addition, our work was done at this single institution. However, since a higher CMI was associated with a lower probability of EDC, hospitals with a lower CMI may have a greater opportunity for EDC improvements. Third, hospital systems are more impacted by low EDCs when operating at high occupancy (as we were at LPCHS); thus, improvements in ED and PACU wait times for inpatient beds might not be noted for hospitals operating with a >10% inventory of beds.²⁵ Importantly, our hospital had multiple daily management structures in place, which we harnessed for our interventions, and better patient flow was a key hospital initiative garnering improvement of resources. Hospitals without these resources may have more difficulty implementing similar interventions. Finally, other work to improve patient flow was concurrently implemented, including matching numbers of scheduled OR admissions with anticipated capacity, which probably also contributed to the decrease in ED and PACU wait times.

We found that a multimodal intervention was associated with more EDCs and improved ED and PACU bed wait times. We observed no impact on discharge satisfaction or readmissions. Our EDC improvement efforts may guide institutions operating at high capacity and aiming to improve EDCs to improve patient flow.

Acknowledgments

The authors would like to acknowledge all those engaged in the early discharge work at LPCHS. They would like to particularly acknowledge Ava Rezvani for her engagement and work in helping to implement the interventions.

Disclosures

The authors have no conflicts of interest relevant to this article to disclose. The authors have no financial relationships relevant to this article to disclose.

Funding

This project was accomplished without specific funding. Funding for incentives was provided by the Lucile Packard Children’s Hospital Stanford.

Patient flow throughout the hospital has been shown to be adversely affected by discharge delays.¹ When hospitals are operating at peak capacity, these delays impact throughput, length of stay (LOS), and cost of care and block patients from the emergency department (ED), postanesthesia recovery unit (PACU), or home awaiting inpatient beds.^2-5 As patients wait in locations not ideal for inpatient care, they may suffer from adverse events and poor satisfaction.^3,6 Several studies have analyzed discharge timing as it relates to ED boarding of admitted patients and demonstrated that early discharges (EDCs) can impact boarding times.^7-9 A number of recent improvement efforts directed at moving discharges earlier in the day have been published.^10-15 However, these improvements are often targeted at specific units or teams within a larger hospital setting and only one is in the pediatric setting.

Lucile Packard Children’s Hospital Stanford (LPCHS) is a 311-bed quaternary care academic women and children’s hospital in Northern California. As our organization expanded, the demand for hospital beds often exceeded capacity. The challenge of overall demand was regularly compounded by a mismatch in bed availability timing – bed demand is early in the day and bed availability is later. This mismatch results in delays for admitted patients waiting in the ED and PACU. Organization leaders identified increasing early discharges (EDCs) as one initiative to contribute to improved patient flow.

Our organization aimed to increase the number of discharges before 11 am across the acute care units from an average of 8% in the 17 months prior to May 2015 to 25% by December 2016. Based on the average number and timing of planned admissions, they hypothesized that 25% of EDCs would decrease ED and PACU wait times.

Setting

We focused our EDC interventions on the 87 acute care beds at LPCHS. All patients discharged from these beds were included in the study. We excluded patients discharged from intensive care, maternity, and nursery. Acute care includes five units, one focused on hematology/oncology (Unit A), one focused on cardiology (Unit B), and the others with a surgical and medical pediatric patient mix (Units C, D, and E). Although physician teams have primary units, due to unit size, patients on teams other than cardiology and hematology/oncology are often spread across multiple units wherever there is a bed (including Units A and B). Most of the frontline care physicians are residents supervised by attendings; however, a minority of patients are cared for by nurse practitioners (NPs) or physician assistants (PAs).

Improvement Team

In early 2015, we formed a multidisciplinary group inclusive of a case manager, frontline nurses, nurse management, pediatric residents, and hospitalist physicians with support from performance improvement. We periodically included physician leaders from other specialties to help initiate changes within their own clinical areas. Our group used Lean A3 thinking¹⁶ to gather information about the current state, formulate the problem statement, analyze the problem, and consider interventions implemented in three Plan–Do-Check-Act (PDCA) cycles. The A3 is a structured tool to analyze problems before jumping to solutions and communicate with stakeholders. We interviewed leaders, nurses, residents, case managers, etc. and observed work processes around discharge. We met weekly to follow data, assess results of interventions, and problem solve.

Barriers and Interventions

The first barrier we identified and addressed was poor identification and shared team mental model of potential EDC patients and lack of preparation when an EDC was identified. In intervention one starting May 2015, charge nurses on Units C, D, and E were each asked to identify one EDC for the following day. The identified patient was discussed at the previously existing afternoon daily unit huddle¹⁷ attended by nurse management, case management, and hospitalist leaders. Following the huddle, the resident, NP, or PA responsible for the patient was paged regarding the EDC plan and tasked with medication reconciliation and discharge paperwork. Others were asked to address their specific area of patient care for discharge (eg, case manager–supplies, nursing–education). The patient was identified on the unit white board with a yellow magnet (use of a visual control¹⁸), so that all would be aware of the EDC. An e-mail was sent to case management, nurse leaders, and patient placement coordinators regarding the planned EDCs. Finally, the EDCs were discussed during regularly scheduled huddles throughout the evening and into the next day.¹⁷

Despite this first intervention, we noted that progress toward increased EDCs was slow. Thus, we spent approximately seven days (spread over one month) further observing the work processes.¹⁹ Over five days, we asked each unit’s charge nurse every hour which patients were waiting to be discharged and the primary reason for waiting. From this information, we created a pareto chart demonstrating that rounds were the highest contributor to waiting (Appendix A). Thus, our second intervention was a daily physician morning huddle that the four nonsurgical physician teams (excluding cardiology, hematology/oncology) implemented one team at a time between November 2015 and February 2016. At the huddle, previously identified EDCs (located on any of the five units) were confirmed and preparatory work was completed (inclusive of the discharge order) before rounds. Further, the attending and resident physicians were to see the patient before or at the start of rounds.

Our working group still observed slow EDC improvement and sought feedback from all providers. EDC was described as “extra” work, apart from routine practices and culture. In addition, our interventions had not addressed most discharges on Units A and B. Consequently, our third intervention in February 2016 aimed to recognize and incentivize teams, units, and individuals for EDC successes. Units and/or physician teams that met 25% of EDCs the previous week were acknowledged through hospital-wide screensavers and certificates of appreciation signed by the Chief Nursing Officer. Units and/or physician teams that met 25% of EDC the previous month were acknowledged with a trophy. Residents received coffee cards for each EDC (though not without controversy among the improvement group as we acknowledged that all providers contributed to EDCs). Finally, weekly, we shared an EDC dashboard displaying unit, team, and organization performance at the hospital-wide leader huddle. We also e-mailed the dashboard regularly to division chiefs, medical directors, and nursing leaders.

Measures

Our primary outcome was percentage of EDCs (based on the time the patient left the room) across acute care. Secondary outcome measures were median wait times for an inpatient bed from the ED (time bed requested to the time patient left the ED) and the average PACU wait time (time the patient is ready to leave the PACU to time the patient left the PACU) per admitted patient. We also assessed balancing measures, including discharge satisfaction, seven-day readmission rates, and LOS. We obtained the mean discharge satisfaction score from the organization’s Press Ganey survey results across acute care (the three discharge questions’ mean – “degree … you felt ready to have your child discharged,” “speed of discharge process …,” and “instructions… to care for your child…”). We obtained seven-day readmission rates from acute care discharges using the hospital’s regularly reported data. We assessed patient characteristics, including sex, age, case mix index (CMI; >2 vs <2), insurance type (nongovernment vs government), day of discharge (weekend vs weekday), and LOS from those patients categorized as inpatients. Complete patient characteristics were not available for observation (InterQual^® criteria) status patients.

Analysis

We used descriptive statistics to describe the inpatient population characteristics by analyzing differences when EDC did and did not occur using chi-square and the Mann–Whitney U tests. Patients with missing data were removed from analyses that incorporated patient factors.

To assess our primary outcome, we used an interrupted time series analysis assessing the percentage of EDC in the total population before any intervention (May 2015) and after the last intervention (March 2016). We used the Durbin–Watson statistic to assess autocorrelation of errors in our regression models. As we had only patient characteristics for the inpatient population, we repeated the analysis including only inpatients and accounting for patient factors significantly associated with EDC.

As units and physician teams had differential exposure to the interventions, we performed a subanalysis (using interrupted time series) creating groups based on the combination of interventions to which a patient’s discharge was exposed (based on unit and physician team at discharge). Patient discharges from group 1 (medical patients on Units C, D, and E) were exposed to all three interventions, group 2 patient discharges (medical patients on Units A and B) were exposed to interventions 2 and 3, group 3 (cardiology, hematology/oncology, surgical patients on Units A and B) were exposed to intervention 3, and group 4 (surgical, cardiology, hematology/oncology patients on Units C, D, and E) were exposed to interventions 1 and 3 (Figure 1). Interrupted time series models were fit using the R Statistical Software Package.²⁰

There were 16,175 discharges on acute care from January 2014 through December 2016. Across all acute care units, EDCs increased from an average of 8.8% before the start of interventions (May 2015) to 15.8% after all interventions (February 2016). From the estimated trend in the preintervention period, there was a jump of 3.9% to the start of the postintervention trend (P = .02; Figure 2). Furthermore, there was an increase of 0.48% (95% CI 0.15-0.82%; P < .01) per month in the trend of the slope between the pre- and postintervention. The autocorrelation function and the Durbin–Watson test did not show evidence of autocorrelation (P = .85). Lack of evidence for autocorrelation in this and each of our subsequent fitted models led to excluding an autocorrelation parameter from our models.

From 16,175 discharges, 1,764 (11%) were assigned to observation status. Among inpatients (14,411), patients with missing values (CMI, insurance status) were also excluded (n = 66, 0.5%). Among the remaining 14,345 inpatients, 54% were males, 50% were government-insured, and 1,645 (11.5%) were discharged early. The average age was 8.5 years, the average LOS was seven days, and the median CMI was 2.2. Children who were younger, had shorter LOS, CMI <2, and nongovernment insurance were more likely to be discharged early (P < .01 for all). For each of these variables, F-tests were performed to determine whether there was a statistically significant reduction in variation by adding the variable to our initial model. None of the variables alone or in combination led to a statistically significant reduction in variation. Including these factors in the interrupted time series did not change the significance of the results (jump at postintervention start 3.6%, 95% CI 0.7%-7.2%; P = .02, slope increased by 0.59% per month, 95% CI 0.29-0.89%; P < .01).

In the subgroup analysis, we did not account for patient factors as they did not change the results in the analysis of total population. Though each group had a greater percentage of EDCs in the postintervention period, the changes in slopes and jumps were primarily nonsignificant (Figure 3). Only the change in slope in group 4 was significant (1.1%, 95% CI 0.3-1.9%; P = .01).

The percentage of patients who left the hospital prior to 11 am significantly improved after a number of interventions aimed at emphasizing EDC and discharge task completion earlier within the hospital stay. Our EDC improvement was associated with improved ED and PACU wait times without negatively impacting discharge satisfaction, seven-day readmissions, or LOS.

It is difficult to compare our EDC improvements to those of previous studies, as we are unaware of published data on pediatric EDC efforts across an entire hospital. In addition, studies have reported discharges prior to different times in the day (noon, 1 pm, etc).^{12, 13} Our interventions were similar to those of Wertheimer et al.,¹¹ including the use of interdisciplinary rounds, identification of potential EDCs the afternoon before discharge, and “reward and recognition.” Wertheimer also sent an e-mail about EDCs to a multidisciplinary group, which was then updated as conditions changed. Unlike Wertheimer, we did not include physicians in our e-mail due to the large number and frequently changing physician teams. Our EDC rate prior to 11 am was lower than their achieved rate of 35% prior to 12 pm. When we assessed our discharges using 12 pm, our rate was still lower (22%-28%), but a direct comparison was complicated by different patient populations. Still, our study adds to the evidence that interdisciplinary rounds and reward and recognition lead to earlier discharge. In addition, this study builds upon Wertheimer’s results as although they later assessed the timing of ED admissions as a result of their EDC improvements, they did not directly assess inpatient bed wait times as we did in our study.¹⁴

As providers of all types were aware of the constant push for beds due to canceled surgeries, delayed admissions and intensive care transfers, and the inability to accept admission, it is difficult to compare the subgroups directly. Furthermore, although physician teams and units are distinct, individuals (nurses, case managers, trainees) may rotate through different units and teams and we cannot account for individual influences on EDCs depending on exposure to interventions over time. Although all groups improved, the improvement in slope in group 4 (exposed to interventions 1 and 3) was the only significant change. As group 4 contained a large number of surgical patients who often have more predictable hospital stays, perhaps this group was more responsive to the interventions.

Our EDC improvements were associated with a decrease in ED and PACU bed wait times. Importantly, we did not address potential confounding factors impacting these times such as total hospital admission volumes, ED and PACU patient complexity, and distribution of ED and PACU admission requests throughout the day. Modeling has suggested that EDCs could also improve ED flow,⁷ but studies implementing EDC have not necessarily assessed this outcome.^10-15 One study retrospectively evaluated ED boarding times in the context of an EDC improvement effort and found a decrease in boarding times.²¹ This decrease is important as ED boarders may be at a higher risk for adverse events, a longer LOS, and more readmissions.^3,7 Less is known about prolonged PACU wait times; however, studies have reported delays in receiving patients from the operating room (OR), which could presumably impact timeliness of other scheduled procedures and patient satisfaction.^22-24 It is worth noting that OR holds as a result of PACU backups happened more frequently at our institution before our EDC work.

Our limitations include that individual providers in the various groups were not completely blind to the interventions and groups often comprised distinct patient populations. Second, LPCHS has a high CMI and LOS relative to most other children’s hospitals, complicating comparison with patient populations at other children’s hospitals. In addition, our work was done at this single institution. However, since a higher CMI was associated with a lower probability of EDC, hospitals with a lower CMI may have a greater opportunity for EDC improvements. Third, hospital systems are more impacted by low EDCs when operating at high occupancy (as we were at LPCHS); thus, improvements in ED and PACU wait times for inpatient beds might not be noted for hospitals operating with a >10% inventory of beds.²⁵ Importantly, our hospital had multiple daily management structures in place, which we harnessed for our interventions, and better patient flow was a key hospital initiative garnering improvement of resources. Hospitals without these resources may have more difficulty implementing similar interventions. Finally, other work to improve patient flow was concurrently implemented, including matching numbers of scheduled OR admissions with anticipated capacity, which probably also contributed to the decrease in ED and PACU wait times.

We found that a multimodal intervention was associated with more EDCs and improved ED and PACU bed wait times. We observed no impact on discharge satisfaction or readmissions. Our EDC improvement efforts may guide institutions operating at high capacity and aiming to improve EDCs to improve patient flow.

Acknowledgments

The authors would like to acknowledge all those engaged in the early discharge work at LPCHS. They would like to particularly acknowledge Ava Rezvani for her engagement and work in helping to implement the interventions.

Disclosures

The authors have no conflicts of interest relevant to this article to disclose. The authors have no financial relationships relevant to this article to disclose.

Funding

This project was accomplished without specific funding. Funding for incentives was provided by the Lucile Packard Children’s Hospital Stanford.

Patient flow throughout the hospital has been shown to be adversely affected by discharge delays.¹ When hospitals are operating at peak capacity, these delays impact throughput, length of stay (LOS), and cost of care and block patients from the emergency department (ED), postanesthesia recovery unit (PACU), or home awaiting inpatient beds.^2-5 As patients wait in locations not ideal for inpatient care, they may suffer from adverse events and poor satisfaction.^3,6 Several studies have analyzed discharge timing as it relates to ED boarding of admitted patients and demonstrated that early discharges (EDCs) can impact boarding times.^7-9 A number of recent improvement efforts directed at moving discharges earlier in the day have been published.^10-15 However, these improvements are often targeted at specific units or teams within a larger hospital setting and only one is in the pediatric setting.

Lucile Packard Children’s Hospital Stanford (LPCHS) is a 311-bed quaternary care academic women and children’s hospital in Northern California. As our organization expanded, the demand for hospital beds often exceeded capacity. The challenge of overall demand was regularly compounded by a mismatch in bed availability timing – bed demand is early in the day and bed availability is later. This mismatch results in delays for admitted patients waiting in the ED and PACU. Organization leaders identified increasing early discharges (EDCs) as one initiative to contribute to improved patient flow.

Our organization aimed to increase the number of discharges before 11 am across the acute care units from an average of 8% in the 17 months prior to May 2015 to 25% by December 2016. Based on the average number and timing of planned admissions, they hypothesized that 25% of EDCs would decrease ED and PACU wait times.

Setting

We focused our EDC interventions on the 87 acute care beds at LPCHS. All patients discharged from these beds were included in the study. We excluded patients discharged from intensive care, maternity, and nursery. Acute care includes five units, one focused on hematology/oncology (Unit A), one focused on cardiology (Unit B), and the others with a surgical and medical pediatric patient mix (Units C, D, and E). Although physician teams have primary units, due to unit size, patients on teams other than cardiology and hematology/oncology are often spread across multiple units wherever there is a bed (including Units A and B). Most of the frontline care physicians are residents supervised by attendings; however, a minority of patients are cared for by nurse practitioners (NPs) or physician assistants (PAs).

Improvement Team

In early 2015, we formed a multidisciplinary group inclusive of a case manager, frontline nurses, nurse management, pediatric residents, and hospitalist physicians with support from performance improvement. We periodically included physician leaders from other specialties to help initiate changes within their own clinical areas. Our group used Lean A3 thinking¹⁶ to gather information about the current state, formulate the problem statement, analyze the problem, and consider interventions implemented in three Plan–Do-Check-Act (PDCA) cycles. The A3 is a structured tool to analyze problems before jumping to solutions and communicate with stakeholders. We interviewed leaders, nurses, residents, case managers, etc. and observed work processes around discharge. We met weekly to follow data, assess results of interventions, and problem solve.

Barriers and Interventions

The first barrier we identified and addressed was poor identification and shared team mental model of potential EDC patients and lack of preparation when an EDC was identified. In intervention one starting May 2015, charge nurses on Units C, D, and E were each asked to identify one EDC for the following day. The identified patient was discussed at the previously existing afternoon daily unit huddle¹⁷ attended by nurse management, case management, and hospitalist leaders. Following the huddle, the resident, NP, or PA responsible for the patient was paged regarding the EDC plan and tasked with medication reconciliation and discharge paperwork. Others were asked to address their specific area of patient care for discharge (eg, case manager–supplies, nursing–education). The patient was identified on the unit white board with a yellow magnet (use of a visual control¹⁸), so that all would be aware of the EDC. An e-mail was sent to case management, nurse leaders, and patient placement coordinators regarding the planned EDCs. Finally, the EDCs were discussed during regularly scheduled huddles throughout the evening and into the next day.¹⁷

Despite this first intervention, we noted that progress toward increased EDCs was slow. Thus, we spent approximately seven days (spread over one month) further observing the work processes.¹⁹ Over five days, we asked each unit’s charge nurse every hour which patients were waiting to be discharged and the primary reason for waiting. From this information, we created a pareto chart demonstrating that rounds were the highest contributor to waiting (Appendix A). Thus, our second intervention was a daily physician morning huddle that the four nonsurgical physician teams (excluding cardiology, hematology/oncology) implemented one team at a time between November 2015 and February 2016. At the huddle, previously identified EDCs (located on any of the five units) were confirmed and preparatory work was completed (inclusive of the discharge order) before rounds. Further, the attending and resident physicians were to see the patient before or at the start of rounds.

Our working group still observed slow EDC improvement and sought feedback from all providers. EDC was described as “extra” work, apart from routine practices and culture. In addition, our interventions had not addressed most discharges on Units A and B. Consequently, our third intervention in February 2016 aimed to recognize and incentivize teams, units, and individuals for EDC successes. Units and/or physician teams that met 25% of EDCs the previous week were acknowledged through hospital-wide screensavers and certificates of appreciation signed by the Chief Nursing Officer. Units and/or physician teams that met 25% of EDC the previous month were acknowledged with a trophy. Residents received coffee cards for each EDC (though not without controversy among the improvement group as we acknowledged that all providers contributed to EDCs). Finally, weekly, we shared an EDC dashboard displaying unit, team, and organization performance at the hospital-wide leader huddle. We also e-mailed the dashboard regularly to division chiefs, medical directors, and nursing leaders.

Measures

Our primary outcome was percentage of EDCs (based on the time the patient left the room) across acute care. Secondary outcome measures were median wait times for an inpatient bed from the ED (time bed requested to the time patient left the ED) and the average PACU wait time (time the patient is ready to leave the PACU to time the patient left the PACU) per admitted patient. We also assessed balancing measures, including discharge satisfaction, seven-day readmission rates, and LOS. We obtained the mean discharge satisfaction score from the organization’s Press Ganey survey results across acute care (the three discharge questions’ mean – “degree … you felt ready to have your child discharged,” “speed of discharge process …,” and “instructions… to care for your child…”). We obtained seven-day readmission rates from acute care discharges using the hospital’s regularly reported data. We assessed patient characteristics, including sex, age, case mix index (CMI; >2 vs <2), insurance type (nongovernment vs government), day of discharge (weekend vs weekday), and LOS from those patients categorized as inpatients. Complete patient characteristics were not available for observation (InterQual^® criteria) status patients.

Analysis

We used descriptive statistics to describe the inpatient population characteristics by analyzing differences when EDC did and did not occur using chi-square and the Mann–Whitney U tests. Patients with missing data were removed from analyses that incorporated patient factors.

To assess our primary outcome, we used an interrupted time series analysis assessing the percentage of EDC in the total population before any intervention (May 2015) and after the last intervention (March 2016). We used the Durbin–Watson statistic to assess autocorrelation of errors in our regression models. As we had only patient characteristics for the inpatient population, we repeated the analysis including only inpatients and accounting for patient factors significantly associated with EDC.

As units and physician teams had differential exposure to the interventions, we performed a subanalysis (using interrupted time series) creating groups based on the combination of interventions to which a patient’s discharge was exposed (based on unit and physician team at discharge). Patient discharges from group 1 (medical patients on Units C, D, and E) were exposed to all three interventions, group 2 patient discharges (medical patients on Units A and B) were exposed to interventions 2 and 3, group 3 (cardiology, hematology/oncology, surgical patients on Units A and B) were exposed to intervention 3, and group 4 (surgical, cardiology, hematology/oncology patients on Units C, D, and E) were exposed to interventions 1 and 3 (Figure 1). Interrupted time series models were fit using the R Statistical Software Package.²⁰

There were 16,175 discharges on acute care from January 2014 through December 2016. Across all acute care units, EDCs increased from an average of 8.8% before the start of interventions (May 2015) to 15.8% after all interventions (February 2016). From the estimated trend in the preintervention period, there was a jump of 3.9% to the start of the postintervention trend (P = .02; Figure 2). Furthermore, there was an increase of 0.48% (95% CI 0.15-0.82%; P < .01) per month in the trend of the slope between the pre- and postintervention. The autocorrelation function and the Durbin–Watson test did not show evidence of autocorrelation (P = .85). Lack of evidence for autocorrelation in this and each of our subsequent fitted models led to excluding an autocorrelation parameter from our models.

From 16,175 discharges, 1,764 (11%) were assigned to observation status. Among inpatients (14,411), patients with missing values (CMI, insurance status) were also excluded (n = 66, 0.5%). Among the remaining 14,345 inpatients, 54% were males, 50% were government-insured, and 1,645 (11.5%) were discharged early. The average age was 8.5 years, the average LOS was seven days, and the median CMI was 2.2. Children who were younger, had shorter LOS, CMI <2, and nongovernment insurance were more likely to be discharged early (P < .01 for all). For each of these variables, F-tests were performed to determine whether there was a statistically significant reduction in variation by adding the variable to our initial model. None of the variables alone or in combination led to a statistically significant reduction in variation. Including these factors in the interrupted time series did not change the significance of the results (jump at postintervention start 3.6%, 95% CI 0.7%-7.2%; P = .02, slope increased by 0.59% per month, 95% CI 0.29-0.89%; P < .01).

In the subgroup analysis, we did not account for patient factors as they did not change the results in the analysis of total population. Though each group had a greater percentage of EDCs in the postintervention period, the changes in slopes and jumps were primarily nonsignificant (Figure 3). Only the change in slope in group 4 was significant (1.1%, 95% CI 0.3-1.9%; P = .01).

The percentage of patients who left the hospital prior to 11 am significantly improved after a number of interventions aimed at emphasizing EDC and discharge task completion earlier within the hospital stay. Our EDC improvement was associated with improved ED and PACU wait times without negatively impacting discharge satisfaction, seven-day readmissions, or LOS.

It is difficult to compare our EDC improvements to those of previous studies, as we are unaware of published data on pediatric EDC efforts across an entire hospital. In addition, studies have reported discharges prior to different times in the day (noon, 1 pm, etc).^{12, 13} Our interventions were similar to those of Wertheimer et al.,¹¹ including the use of interdisciplinary rounds, identification of potential EDCs the afternoon before discharge, and “reward and recognition.” Wertheimer also sent an e-mail about EDCs to a multidisciplinary group, which was then updated as conditions changed. Unlike Wertheimer, we did not include physicians in our e-mail due to the large number and frequently changing physician teams. Our EDC rate prior to 11 am was lower than their achieved rate of 35% prior to 12 pm. When we assessed our discharges using 12 pm, our rate was still lower (22%-28%), but a direct comparison was complicated by different patient populations. Still, our study adds to the evidence that interdisciplinary rounds and reward and recognition lead to earlier discharge. In addition, this study builds upon Wertheimer’s results as although they later assessed the timing of ED admissions as a result of their EDC improvements, they did not directly assess inpatient bed wait times as we did in our study.¹⁴

As providers of all types were aware of the constant push for beds due to canceled surgeries, delayed admissions and intensive care transfers, and the inability to accept admission, it is difficult to compare the subgroups directly. Furthermore, although physician teams and units are distinct, individuals (nurses, case managers, trainees) may rotate through different units and teams and we cannot account for individual influences on EDCs depending on exposure to interventions over time. Although all groups improved, the improvement in slope in group 4 (exposed to interventions 1 and 3) was the only significant change. As group 4 contained a large number of surgical patients who often have more predictable hospital stays, perhaps this group was more responsive to the interventions.

Our EDC improvements were associated with a decrease in ED and PACU bed wait times. Importantly, we did not address potential confounding factors impacting these times such as total hospital admission volumes, ED and PACU patient complexity, and distribution of ED and PACU admission requests throughout the day. Modeling has suggested that EDCs could also improve ED flow,⁷ but studies implementing EDC have not necessarily assessed this outcome.^10-15 One study retrospectively evaluated ED boarding times in the context of an EDC improvement effort and found a decrease in boarding times.²¹ This decrease is important as ED boarders may be at a higher risk for adverse events, a longer LOS, and more readmissions.^3,7 Less is known about prolonged PACU wait times; however, studies have reported delays in receiving patients from the operating room (OR), which could presumably impact timeliness of other scheduled procedures and patient satisfaction.^22-24 It is worth noting that OR holds as a result of PACU backups happened more frequently at our institution before our EDC work.

Our limitations include that individual providers in the various groups were not completely blind to the interventions and groups often comprised distinct patient populations. Second, LPCHS has a high CMI and LOS relative to most other children’s hospitals, complicating comparison with patient populations at other children’s hospitals. In addition, our work was done at this single institution. However, since a higher CMI was associated with a lower probability of EDC, hospitals with a lower CMI may have a greater opportunity for EDC improvements. Third, hospital systems are more impacted by low EDCs when operating at high occupancy (as we were at LPCHS); thus, improvements in ED and PACU wait times for inpatient beds might not be noted for hospitals operating with a >10% inventory of beds.²⁵ Importantly, our hospital had multiple daily management structures in place, which we harnessed for our interventions, and better patient flow was a key hospital initiative garnering improvement of resources. Hospitals without these resources may have more difficulty implementing similar interventions. Finally, other work to improve patient flow was concurrently implemented, including matching numbers of scheduled OR admissions with anticipated capacity, which probably also contributed to the decrease in ED and PACU wait times.

We found that a multimodal intervention was associated with more EDCs and improved ED and PACU bed wait times. We observed no impact on discharge satisfaction or readmissions. Our EDC improvement efforts may guide institutions operating at high capacity and aiming to improve EDCs to improve patient flow.

Acknowledgments

The authors would like to acknowledge all those engaged in the early discharge work at LPCHS. They would like to particularly acknowledge Ava Rezvani for her engagement and work in helping to implement the interventions.

Disclosures

The authors have no conflicts of interest relevant to this article to disclose. The authors have no financial relationships relevant to this article to disclose.

Funding

This project was accomplished without specific funding. Funding for incentives was provided by the Lucile Packard Children’s Hospital Stanford.

Many hospitals and emergency departments (EDs) face challenges posed by overcrowding and hospital throughput. Slow ED throughput has been associated with worse patient outcomes.¹ One strategy increasingly employed to improve hospital throughput is to increase the rate of inpatient discharges earlier in the day, which is often defined as discharges before noon (DCBNs). The hypothesis behind DCBN is that earlier hospital discharges will allow for earlier ED admissions and thus mitigate ED overcrowding while optimizing inpatient hospital flow. Previous quality improvement efforts to increase the percentage of DCBNs have been successfully implemented. For example, Wertheimer et al. implemented a process for earlier discharges and reported a 27-percentage point (11% to 38%) increase in DCBN on general medicine units.² In a recent survey among leaders in hospital medicine programs, a majority reported early discharge as an important institutional goal.³

Studies of the effectiveness of DCBN initiatives on improving throughput and shortening length of stay (LOS) in adult patients have had mixed results. Computer modeling has supported the idea that earlier inpatient discharges would shorten ED patient boarding time.⁴Wertheimer et al. performed a retrospective analysis of a DCBN intervention on two inpatient medicine units and reported an association between slightly shorter observed versus expected inpatient LOS² and earlier arrival time of inpatient admissions from the ED.⁵ In contrast, Rajkomar et al. conducted a retrospective analysis of the association of DCBN and LOS in a predominantly surgical services population and reported a longer LOS for DCBN patients when controlling for patient characteristics and comorbidities.⁶ These mixed findings have led some authors to question the value of DCBN initiatives and created concern for the potential of prolonged patient hospitalizations as a result of institutional DCBN goals.⁷ The impact of DCBN in pediatric patients is much less studied.

A question of interest for hospitals is if DCBN is a good indicator of shorter LOS, or is DCBN an arbitrary indicator, as morning discharges might just be the result of a delayed discharge of a patient ready for discharge the prior afternoon/evening. Our study objectives were: (1) to determine whether DCBN is associated with a shorter LOS in a pediatric population at an academic medical center, and (2) to examine separately this association in medical and surgical patients given the different provider workflow and patient clinical characteristics in those groups.

Patients and Settings

This retrospective cohort analysis included pediatric medical and surgical inpatient admissions from a single academic medical center from May 2014 to April 2017. The University of North Carolina (UNC) Children’s Hospital is a 175-bed tertiary care ‘hospital within a hospital’ in an academic setting with multiple residencies. UNC Children’s Hospital contains three units providing inpatient pediatric care. Each unit occupies a floor of the Children’s hospital and are loosely regionalized, as follows: (1) Unit 7 is focused on surgical patients; (2) Unit 6 is focused on general, neurologic, and renal patients; and (3) Unit 5 is focused on hematology/oncology and pulmonary patients. Extending the entire study period, Unit 6 initiated a quality improvement effort to discharge patients earlier in the day, specifically before 1 pm; however, the initiative did not extend beyond this one unit.

We included patients 21 years or younger with an inpatient admission to any of the following pediatric medical or surgical services: cardiac surgery, cardiology, endocrinology, gastroenterology, general services, hematology/oncology, nephrology, orthopedics, otolaryngology, plastic surgery, pulmonology, and urology. Patients whose stay did not extend beyond one midnight were excluded because discharge time of day for these short stays was strongly related to the time of admission. We also excluded patients whose stay extended beyond two standard deviations of the average LOS for the discharge service under the assumption that these patients represented atypical circumstances. Finally, we excluded patients who died or left against medical advice. A consortium diagram of all exclusion criteria can be found in Supplemental Figure 1. Discharge data were extracted from the Carolina Database Warehouse, a data repository of the University of North Carolina Health System. The University of North Carolina Institutional Review Board reviewed and approved this study (IRB 17-0500).

Measures

The outcome of interest was LOS, defined as discharge date and time minus admission date and time, and thus a continuous measure of time in the hospital rather than a number of midnights. Rajkomar et al. used the same definition of LOS.⁶ The independent variable of interest was whether the discharge occurred before noon. Because discharges between midnight and 8:00 am are likely unplanned and not attributable to any particular workflow, we followed a similar definition of DCBN used by Rajkomer et al. and defined DCBN as a patient leaving between 8:00 am and 11:59 am (pre-8:00 am discharges accounted for less than one half of one percent of discharges).⁶

All model covariates were collected at the patient level (Table 1), including demographic characteristics such as age, sex, race, and ethnicity. We also collected covariates describing the patient’s hospitalization as follows: (1) whether the patient was discharged on a weekend versus weekday; (2) hospital service at time of discharge (dichotomized to a surgical or medical service); (3) whether the patient was discharged from the unit that had a DCBN quality improvement initiative; (4) discharge disposition (home with self-care, assisted living or home health, or other); (5) insurance type during hospitalization (commercial, Medicaid, no insurance, or other); and (6) case mix index (CMI), a measure of hospital resource intensity of a patient’s principal diagnosis. Covariate selection was made on the basis of a priori knowledge of causal pathways.⁸

Statistical Analysis

Student t tests and χ² statistics were used to compare baseline characteristics of hospitalizations of patients DCBN and after noon. We used ordinary least squares (OLS) regression models to assess the association between DCBN and LOS. Because DCBN may be correlated with patient characteristics, we used propensity score weighted models. Propensity scores were estimated using a logistic regression predicting DCBN using the variables given in Table 1 (excluding the outcome variable LOS). To estimate the average treatment effect on the entire sample for each model, we weighted each observation by the inverse-probability of treatment as per recent propensity score methods detailed by Garrido et al.⁹ In the inverse-probability weighted models, we clustered on attending physician to adjust for the autocorrelation caused by unobservable similarities of discharges by the same attending. We tested for multicollinearity using the variance inflation factor (VIF). To test our secondary hypothesis that there was a difference in the relationship between DCBN and LOS based on service type (medical versus surgical), we tested if the service type moderated any of the coefficients using a joint Wald test on the 10 coefficients interacted with the service type.

For our sensitivity analysis, we reran all surgical and medical discharges models changing the LOS outlier exclusion criteria to greater than three and then four standard deviations. Statistical modeling and analysis were completed using Stata version 14 (StataCorp, College Station, Texas).

Our study sample comprised 8,226 pediatric hospitalizations with a LOS mean of 5.10 and a median of 3.91 days respectively (range, 1.25-32.83 days). There were 1,531 (18.6%) DCBNs. Compared to those discharged after noon, patients with DCBN had a higher probability of being surgical patients, having commercial insurance, discharge home with self-care, discharge on the weekend, and discharge from a nonquality improvement unit (Table 1). Patients with DCBN were also more likely to be white, non-Hispanic, and male.

Our propensity score weighted ordinary least score (OLS) LOS regression results are presented in Table 2. In the bivariate analysis, DCBN was associated with an average 0.40 day, or roughly 10 hours, shorter LOS (P < .001). In the multivariate model of all discharges, we found that DCBN was associated with a mean of 0.27 day (P = .010) shorter LOS when compared to discharge in the afternoon when controlling for age, race, ethnicity, weekend discharge, discharge from quality improvement unit, discharge service type, CMI, insurance type, and discharge disposition. In the multivariate analysis, weekend discharge, surgical discharge, and discharge disposition of home with self-care, compared to assisted living or home health were associated with shorter LOS.

In our study of over 8,000 pediatric discharges during a three-year period, DCBN was associated with shorter LOS for medical pediatric patients, but this finding was not consistent for surgical patients. Among medical discharges, DCBN was associated with shorter LOS, an effect robust enough to include or exclude outliers (for LOS, outliers are an important subset because there are always, in general, a few patients with very long lengths of stay). Discharge before noon showed no association with LOS for surgical patients unless we included outlier values.

The differential effect of DCBN on LOS in surgical and medical discharges suggests that the relationship between DCBN and LOS may be related to provider team workflow. For example, surgical teams may tend to round one time per day early in the morning before spending the entire day in the operating room, and thus completing more early morning discharge orders compared to medical teams. However, if a patient on a surgical service is not ready for discharge first thing in the morning, the patient may be more likely to wait until the following morning for a discharge order. On medical services, physician schedules may allow for more flexibility for rounding and responding with a discharge order when a patient becomes ready; however, medical services may round later in the day compared to surgeons and for a longer period of time, delaying discharges beyond noon that could have been made earlier. Another possibility, given UNC pediatric services are loosely regionalized with surgical patients concentrated more in one unit, is that unit-level differences in how staff processed discharges could have contributed to the difference observed between medical and surgical patients, particularly as there was a unit-level quality improvement effort for decreasing discharge time on one of two medical floors. However, we analyzed for differences based on the discharging unit and found no association. The influence of outliers on the association between DCBN and LOS increases also suggests that this group of children who have extremely long hospital stays might need further exploration.

Our study has some similar and some contrasting results with prior studies in adult patients. Our findings support the modeling literature that suggests DCBN may improve discharge efficiency by shortening patient LOS for some discharges.⁴ These findings contrast with Rajkomar et al., who reported that DCBN was associated with a longer LOS in adult patients.⁶ The contrasting findings could be due to differences in pediatric versus adult patients. Additionally, the population Rajkomar et al. studied was predominantly surgical patients, whose discharges may differ from medical patients’ in many aspects. Another possible explanation is that the Rajkomar et al. study was performed in a setting with clearly set institutional targets for DCBN, whereas, our institution lacked any hospital-wide DCBN initiatives or standards to which providers were held accountable. Some authors have argued setting DCBN as a measure of hospital quality perhaps creates the unintended consequence of providers holding potential afternoon or evening discharges until the next day so that they can be DCBN.^7,10 In that scenario, perhaps there would be a relationship between DCBN and longer LOS compared to patients who are reevaluated in the afternoon or evening and discharged. We did not find evidence of these effects in our analysis, however, understanding the potential for this is important when designing quality improvement efforts aimed at increasing discharge efficiency.

While shorter LOS can be an indicator of high-value care, the relationship between LOS, DCBN, and efficiency of discharge processes remains unclear. Prior studies have found evidence that multidisciplinary care teams with frequent care coordination rounds and integration of electronic admission order sets can be effective in improving discharge efficiency as measured by discharge within two hours of meeting discharge goals.^11,12 Measuring discharge efficiency on an ongoing basis is very difficult; however, easy-to-measure targets such as discharge before noon may be used as a proxy measure of efficiency. These targets also have “face validity,” and because of these two factors, measures like DCBN have been widely implemented even though evidence to support their validity is minimal.

Our study has several limitations. While we controlled for observable characteristics using covariates and propensity score weighted analyses, there are likely unobservable characteristics that confound our analysis. We did not measure other factors that may affect discharge time of day such as high occupancy, staffing levels, patient transportation availability, and patient and family preferences. Given these limitations, we caution against interpreting a causal relationship between independent variables and the outcome. Finally, this analysis was conducted at a single tertiary care, academic medical center. The majority of pediatric admissions at this institution are either transferred from other hospitals or scheduled admissions for medical or surgical care. A smaller proportion of discharges are acute, unplanned admissions through our emergency department in children with or without underlying medical complexity. These factors plus the exclusion of observation, extended recovery, and all the less than two-day stays in this study contribute to a relatively higher average LOS. These factors potentially limit generalizability to other care settings. Additionally, the majority of the care teams involve care by resident physicians, and they are often the primary caregivers and write the majority of orders in patient charts such as discharge orders. While we were not able to control for within resident physician similarities between patients, we did control for autocorrelation at the attending level.

The results of our study suggest that DCBN is associated with a decreased LOS for medical but not surgical pediatric patients. DCBN may not be an appropriate measure for all services. Further research should be done to identify other feasible but more valid indicators for shorter LOS.

Disclosures

The authors have no financial relationships relative to this article to disclose. The authors have no conflicts of interest relevant to this article to disclose.

Funding

There were no external sources of funding for this work.

Many hospitals and emergency departments (EDs) face challenges posed by overcrowding and hospital throughput. Slow ED throughput has been associated with worse patient outcomes.¹ One strategy increasingly employed to improve hospital throughput is to increase the rate of inpatient discharges earlier in the day, which is often defined as discharges before noon (DCBNs). The hypothesis behind DCBN is that earlier hospital discharges will allow for earlier ED admissions and thus mitigate ED overcrowding while optimizing inpatient hospital flow. Previous quality improvement efforts to increase the percentage of DCBNs have been successfully implemented. For example, Wertheimer et al. implemented a process for earlier discharges and reported a 27-percentage point (11% to 38%) increase in DCBN on general medicine units.² In a recent survey among leaders in hospital medicine programs, a majority reported early discharge as an important institutional goal.³

Studies of the effectiveness of DCBN initiatives on improving throughput and shortening length of stay (LOS) in adult patients have had mixed results. Computer modeling has supported the idea that earlier inpatient discharges would shorten ED patient boarding time.⁴Wertheimer et al. performed a retrospective analysis of a DCBN intervention on two inpatient medicine units and reported an association between slightly shorter observed versus expected inpatient LOS² and earlier arrival time of inpatient admissions from the ED.⁵ In contrast, Rajkomar et al. conducted a retrospective analysis of the association of DCBN and LOS in a predominantly surgical services population and reported a longer LOS for DCBN patients when controlling for patient characteristics and comorbidities.⁶ These mixed findings have led some authors to question the value of DCBN initiatives and created concern for the potential of prolonged patient hospitalizations as a result of institutional DCBN goals.⁷ The impact of DCBN in pediatric patients is much less studied.

A question of interest for hospitals is if DCBN is a good indicator of shorter LOS, or is DCBN an arbitrary indicator, as morning discharges might just be the result of a delayed discharge of a patient ready for discharge the prior afternoon/evening. Our study objectives were: (1) to determine whether DCBN is associated with a shorter LOS in a pediatric population at an academic medical center, and (2) to examine separately this association in medical and surgical patients given the different provider workflow and patient clinical characteristics in those groups.

Patients and Settings

This retrospective cohort analysis included pediatric medical and surgical inpatient admissions from a single academic medical center from May 2014 to April 2017. The University of North Carolina (UNC) Children’s Hospital is a 175-bed tertiary care ‘hospital within a hospital’ in an academic setting with multiple residencies. UNC Children’s Hospital contains three units providing inpatient pediatric care. Each unit occupies a floor of the Children’s hospital and are loosely regionalized, as follows: (1) Unit 7 is focused on surgical patients; (2) Unit 6 is focused on general, neurologic, and renal patients; and (3) Unit 5 is focused on hematology/oncology and pulmonary patients. Extending the entire study period, Unit 6 initiated a quality improvement effort to discharge patients earlier in the day, specifically before 1 pm; however, the initiative did not extend beyond this one unit.

We included patients 21 years or younger with an inpatient admission to any of the following pediatric medical or surgical services: cardiac surgery, cardiology, endocrinology, gastroenterology, general services, hematology/oncology, nephrology, orthopedics, otolaryngology, plastic surgery, pulmonology, and urology. Patients whose stay did not extend beyond one midnight were excluded because discharge time of day for these short stays was strongly related to the time of admission. We also excluded patients whose stay extended beyond two standard deviations of the average LOS for the discharge service under the assumption that these patients represented atypical circumstances. Finally, we excluded patients who died or left against medical advice. A consortium diagram of all exclusion criteria can be found in Supplemental Figure 1. Discharge data were extracted from the Carolina Database Warehouse, a data repository of the University of North Carolina Health System. The University of North Carolina Institutional Review Board reviewed and approved this study (IRB 17-0500).

Measures

The outcome of interest was LOS, defined as discharge date and time minus admission date and time, and thus a continuous measure of time in the hospital rather than a number of midnights. Rajkomar et al. used the same definition of LOS.⁶ The independent variable of interest was whether the discharge occurred before noon. Because discharges between midnight and 8:00 am are likely unplanned and not attributable to any particular workflow, we followed a similar definition of DCBN used by Rajkomer et al. and defined DCBN as a patient leaving between 8:00 am and 11:59 am (pre-8:00 am discharges accounted for less than one half of one percent of discharges).⁶

All model covariates were collected at the patient level (Table 1), including demographic characteristics such as age, sex, race, and ethnicity. We also collected covariates describing the patient’s hospitalization as follows: (1) whether the patient was discharged on a weekend versus weekday; (2) hospital service at time of discharge (dichotomized to a surgical or medical service); (3) whether the patient was discharged from the unit that had a DCBN quality improvement initiative; (4) discharge disposition (home with self-care, assisted living or home health, or other); (5) insurance type during hospitalization (commercial, Medicaid, no insurance, or other); and (6) case mix index (CMI), a measure of hospital resource intensity of a patient’s principal diagnosis. Covariate selection was made on the basis of a priori knowledge of causal pathways.⁸

Statistical Analysis

Student t tests and χ² statistics were used to compare baseline characteristics of hospitalizations of patients DCBN and after noon. We used ordinary least squares (OLS) regression models to assess the association between DCBN and LOS. Because DCBN may be correlated with patient characteristics, we used propensity score weighted models. Propensity scores were estimated using a logistic regression predicting DCBN using the variables given in Table 1 (excluding the outcome variable LOS). To estimate the average treatment effect on the entire sample for each model, we weighted each observation by the inverse-probability of treatment as per recent propensity score methods detailed by Garrido et al.⁹ In the inverse-probability weighted models, we clustered on attending physician to adjust for the autocorrelation caused by unobservable similarities of discharges by the same attending. We tested for multicollinearity using the variance inflation factor (VIF). To test our secondary hypothesis that there was a difference in the relationship between DCBN and LOS based on service type (medical versus surgical), we tested if the service type moderated any of the coefficients using a joint Wald test on the 10 coefficients interacted with the service type.

For our sensitivity analysis, we reran all surgical and medical discharges models changing the LOS outlier exclusion criteria to greater than three and then four standard deviations. Statistical modeling and analysis were completed using Stata version 14 (StataCorp, College Station, Texas).

Our study sample comprised 8,226 pediatric hospitalizations with a LOS mean of 5.10 and a median of 3.91 days respectively (range, 1.25-32.83 days). There were 1,531 (18.6%) DCBNs. Compared to those discharged after noon, patients with DCBN had a higher probability of being surgical patients, having commercial insurance, discharge home with self-care, discharge on the weekend, and discharge from a nonquality improvement unit (Table 1). Patients with DCBN were also more likely to be white, non-Hispanic, and male.

Our propensity score weighted ordinary least score (OLS) LOS regression results are presented in Table 2. In the bivariate analysis, DCBN was associated with an average 0.40 day, or roughly 10 hours, shorter LOS (P < .001). In the multivariate model of all discharges, we found that DCBN was associated with a mean of 0.27 day (P = .010) shorter LOS when compared to discharge in the afternoon when controlling for age, race, ethnicity, weekend discharge, discharge from quality improvement unit, discharge service type, CMI, insurance type, and discharge disposition. In the multivariate analysis, weekend discharge, surgical discharge, and discharge disposition of home with self-care, compared to assisted living or home health were associated with shorter LOS.

In our study of over 8,000 pediatric discharges during a three-year period, DCBN was associated with shorter LOS for medical pediatric patients, but this finding was not consistent for surgical patients. Among medical discharges, DCBN was associated with shorter LOS, an effect robust enough to include or exclude outliers (for LOS, outliers are an important subset because there are always, in general, a few patients with very long lengths of stay). Discharge before noon showed no association with LOS for surgical patients unless we included outlier values.

The differential effect of DCBN on LOS in surgical and medical discharges suggests that the relationship between DCBN and LOS may be related to provider team workflow. For example, surgical teams may tend to round one time per day early in the morning before spending the entire day in the operating room, and thus completing more early morning discharge orders compared to medical teams. However, if a patient on a surgical service is not ready for discharge first thing in the morning, the patient may be more likely to wait until the following morning for a discharge order. On medical services, physician schedules may allow for more flexibility for rounding and responding with a discharge order when a patient becomes ready; however, medical services may round later in the day compared to surgeons and for a longer period of time, delaying discharges beyond noon that could have been made earlier. Another possibility, given UNC pediatric services are loosely regionalized with surgical patients concentrated more in one unit, is that unit-level differences in how staff processed discharges could have contributed to the difference observed between medical and surgical patients, particularly as there was a unit-level quality improvement effort for decreasing discharge time on one of two medical floors. However, we analyzed for differences based on the discharging unit and found no association. The influence of outliers on the association between DCBN and LOS increases also suggests that this group of children who have extremely long hospital stays might need further exploration.

Our study has some similar and some contrasting results with prior studies in adult patients. Our findings support the modeling literature that suggests DCBN may improve discharge efficiency by shortening patient LOS for some discharges.⁴ These findings contrast with Rajkomar et al., who reported that DCBN was associated with a longer LOS in adult patients.⁶ The contrasting findings could be due to differences in pediatric versus adult patients. Additionally, the population Rajkomar et al. studied was predominantly surgical patients, whose discharges may differ from medical patients’ in many aspects. Another possible explanation is that the Rajkomar et al. study was performed in a setting with clearly set institutional targets for DCBN, whereas, our institution lacked any hospital-wide DCBN initiatives or standards to which providers were held accountable. Some authors have argued setting DCBN as a measure of hospital quality perhaps creates the unintended consequence of providers holding potential afternoon or evening discharges until the next day so that they can be DCBN.^7,10 In that scenario, perhaps there would be a relationship between DCBN and longer LOS compared to patients who are reevaluated in the afternoon or evening and discharged. We did not find evidence of these effects in our analysis, however, understanding the potential for this is important when designing quality improvement efforts aimed at increasing discharge efficiency.

While shorter LOS can be an indicator of high-value care, the relationship between LOS, DCBN, and efficiency of discharge processes remains unclear. Prior studies have found evidence that multidisciplinary care teams with frequent care coordination rounds and integration of electronic admission order sets can be effective in improving discharge efficiency as measured by discharge within two hours of meeting discharge goals.^11,12 Measuring discharge efficiency on an ongoing basis is very difficult; however, easy-to-measure targets such as discharge before noon may be used as a proxy measure of efficiency. These targets also have “face validity,” and because of these two factors, measures like DCBN have been widely implemented even though evidence to support their validity is minimal.

Our study has several limitations. While we controlled for observable characteristics using covariates and propensity score weighted analyses, there are likely unobservable characteristics that confound our analysis. We did not measure other factors that may affect discharge time of day such as high occupancy, staffing levels, patient transportation availability, and patient and family preferences. Given these limitations, we caution against interpreting a causal relationship between independent variables and the outcome. Finally, this analysis was conducted at a single tertiary care, academic medical center. The majority of pediatric admissions at this institution are either transferred from other hospitals or scheduled admissions for medical or surgical care. A smaller proportion of discharges are acute, unplanned admissions through our emergency department in children with or without underlying medical complexity. These factors plus the exclusion of observation, extended recovery, and all the less than two-day stays in this study contribute to a relatively higher average LOS. These factors potentially limit generalizability to other care settings. Additionally, the majority of the care teams involve care by resident physicians, and they are often the primary caregivers and write the majority of orders in patient charts such as discharge orders. While we were not able to control for within resident physician similarities between patients, we did control for autocorrelation at the attending level.

The results of our study suggest that DCBN is associated with a decreased LOS for medical but not surgical pediatric patients. DCBN may not be an appropriate measure for all services. Further research should be done to identify other feasible but more valid indicators for shorter LOS.

Disclosures

The authors have no financial relationships relative to this article to disclose. The authors have no conflicts of interest relevant to this article to disclose.

Funding

There were no external sources of funding for this work.

Many hospitals and emergency departments (EDs) face challenges posed by overcrowding and hospital throughput. Slow ED throughput has been associated with worse patient outcomes.¹ One strategy increasingly employed to improve hospital throughput is to increase the rate of inpatient discharges earlier in the day, which is often defined as discharges before noon (DCBNs). The hypothesis behind DCBN is that earlier hospital discharges will allow for earlier ED admissions and thus mitigate ED overcrowding while optimizing inpatient hospital flow. Previous quality improvement efforts to increase the percentage of DCBNs have been successfully implemented. For example, Wertheimer et al. implemented a process for earlier discharges and reported a 27-percentage point (11% to 38%) increase in DCBN on general medicine units.² In a recent survey among leaders in hospital medicine programs, a majority reported early discharge as an important institutional goal.³

Studies of the effectiveness of DCBN initiatives on improving throughput and shortening length of stay (LOS) in adult patients have had mixed results. Computer modeling has supported the idea that earlier inpatient discharges would shorten ED patient boarding time.⁴Wertheimer et al. performed a retrospective analysis of a DCBN intervention on two inpatient medicine units and reported an association between slightly shorter observed versus expected inpatient LOS² and earlier arrival time of inpatient admissions from the ED.⁵ In contrast, Rajkomar et al. conducted a retrospective analysis of the association of DCBN and LOS in a predominantly surgical services population and reported a longer LOS for DCBN patients when controlling for patient characteristics and comorbidities.⁶ These mixed findings have led some authors to question the value of DCBN initiatives and created concern for the potential of prolonged patient hospitalizations as a result of institutional DCBN goals.⁷ The impact of DCBN in pediatric patients is much less studied.

A question of interest for hospitals is if DCBN is a good indicator of shorter LOS, or is DCBN an arbitrary indicator, as morning discharges might just be the result of a delayed discharge of a patient ready for discharge the prior afternoon/evening. Our study objectives were: (1) to determine whether DCBN is associated with a shorter LOS in a pediatric population at an academic medical center, and (2) to examine separately this association in medical and surgical patients given the different provider workflow and patient clinical characteristics in those groups.

Patients and Settings

This retrospective cohort analysis included pediatric medical and surgical inpatient admissions from a single academic medical center from May 2014 to April 2017. The University of North Carolina (UNC) Children’s Hospital is a 175-bed tertiary care ‘hospital within a hospital’ in an academic setting with multiple residencies. UNC Children’s Hospital contains three units providing inpatient pediatric care. Each unit occupies a floor of the Children’s hospital and are loosely regionalized, as follows: (1) Unit 7 is focused on surgical patients; (2) Unit 6 is focused on general, neurologic, and renal patients; and (3) Unit 5 is focused on hematology/oncology and pulmonary patients. Extending the entire study period, Unit 6 initiated a quality improvement effort to discharge patients earlier in the day, specifically before 1 pm; however, the initiative did not extend beyond this one unit.

We included patients 21 years or younger with an inpatient admission to any of the following pediatric medical or surgical services: cardiac surgery, cardiology, endocrinology, gastroenterology, general services, hematology/oncology, nephrology, orthopedics, otolaryngology, plastic surgery, pulmonology, and urology. Patients whose stay did not extend beyond one midnight were excluded because discharge time of day for these short stays was strongly related to the time of admission. We also excluded patients whose stay extended beyond two standard deviations of the average LOS for the discharge service under the assumption that these patients represented atypical circumstances. Finally, we excluded patients who died or left against medical advice. A consortium diagram of all exclusion criteria can be found in Supplemental Figure 1. Discharge data were extracted from the Carolina Database Warehouse, a data repository of the University of North Carolina Health System. The University of North Carolina Institutional Review Board reviewed and approved this study (IRB 17-0500).

Measures

The outcome of interest was LOS, defined as discharge date and time minus admission date and time, and thus a continuous measure of time in the hospital rather than a number of midnights. Rajkomar et al. used the same definition of LOS.⁶ The independent variable of interest was whether the discharge occurred before noon. Because discharges between midnight and 8:00 am are likely unplanned and not attributable to any particular workflow, we followed a similar definition of DCBN used by Rajkomer et al. and defined DCBN as a patient leaving between 8:00 am and 11:59 am (pre-8:00 am discharges accounted for less than one half of one percent of discharges).⁶

All model covariates were collected at the patient level (Table 1), including demographic characteristics such as age, sex, race, and ethnicity. We also collected covariates describing the patient’s hospitalization as follows: (1) whether the patient was discharged on a weekend versus weekday; (2) hospital service at time of discharge (dichotomized to a surgical or medical service); (3) whether the patient was discharged from the unit that had a DCBN quality improvement initiative; (4) discharge disposition (home with self-care, assisted living or home health, or other); (5) insurance type during hospitalization (commercial, Medicaid, no insurance, or other); and (6) case mix index (CMI), a measure of hospital resource intensity of a patient’s principal diagnosis. Covariate selection was made on the basis of a priori knowledge of causal pathways.⁸

Statistical Analysis

Student t tests and χ² statistics were used to compare baseline characteristics of hospitalizations of patients DCBN and after noon. We used ordinary least squares (OLS) regression models to assess the association between DCBN and LOS. Because DCBN may be correlated with patient characteristics, we used propensity score weighted models. Propensity scores were estimated using a logistic regression predicting DCBN using the variables given in Table 1 (excluding the outcome variable LOS). To estimate the average treatment effect on the entire sample for each model, we weighted each observation by the inverse-probability of treatment as per recent propensity score methods detailed by Garrido et al.⁹ In the inverse-probability weighted models, we clustered on attending physician to adjust for the autocorrelation caused by unobservable similarities of discharges by the same attending. We tested for multicollinearity using the variance inflation factor (VIF). To test our secondary hypothesis that there was a difference in the relationship between DCBN and LOS based on service type (medical versus surgical), we tested if the service type moderated any of the coefficients using a joint Wald test on the 10 coefficients interacted with the service type.

For our sensitivity analysis, we reran all surgical and medical discharges models changing the LOS outlier exclusion criteria to greater than three and then four standard deviations. Statistical modeling and analysis were completed using Stata version 14 (StataCorp, College Station, Texas).

Our study sample comprised 8,226 pediatric hospitalizations with a LOS mean of 5.10 and a median of 3.91 days respectively (range, 1.25-32.83 days). There were 1,531 (18.6%) DCBNs. Compared to those discharged after noon, patients with DCBN had a higher probability of being surgical patients, having commercial insurance, discharge home with self-care, discharge on the weekend, and discharge from a nonquality improvement unit (Table 1). Patients with DCBN were also more likely to be white, non-Hispanic, and male.

Our propensity score weighted ordinary least score (OLS) LOS regression results are presented in Table 2. In the bivariate analysis, DCBN was associated with an average 0.40 day, or roughly 10 hours, shorter LOS (P < .001). In the multivariate model of all discharges, we found that DCBN was associated with a mean of 0.27 day (P = .010) shorter LOS when compared to discharge in the afternoon when controlling for age, race, ethnicity, weekend discharge, discharge from quality improvement unit, discharge service type, CMI, insurance type, and discharge disposition. In the multivariate analysis, weekend discharge, surgical discharge, and discharge disposition of home with self-care, compared to assisted living or home health were associated with shorter LOS.

In our study of over 8,000 pediatric discharges during a three-year period, DCBN was associated with shorter LOS for medical pediatric patients, but this finding was not consistent for surgical patients. Among medical discharges, DCBN was associated with shorter LOS, an effect robust enough to include or exclude outliers (for LOS, outliers are an important subset because there are always, in general, a few patients with very long lengths of stay). Discharge before noon showed no association with LOS for surgical patients unless we included outlier values.

The differential effect of DCBN on LOS in surgical and medical discharges suggests that the relationship between DCBN and LOS may be related to provider team workflow. For example, surgical teams may tend to round one time per day early in the morning before spending the entire day in the operating room, and thus completing more early morning discharge orders compared to medical teams. However, if a patient on a surgical service is not ready for discharge first thing in the morning, the patient may be more likely to wait until the following morning for a discharge order. On medical services, physician schedules may allow for more flexibility for rounding and responding with a discharge order when a patient becomes ready; however, medical services may round later in the day compared to surgeons and for a longer period of time, delaying discharges beyond noon that could have been made earlier. Another possibility, given UNC pediatric services are loosely regionalized with surgical patients concentrated more in one unit, is that unit-level differences in how staff processed discharges could have contributed to the difference observed between medical and surgical patients, particularly as there was a unit-level quality improvement effort for decreasing discharge time on one of two medical floors. However, we analyzed for differences based on the discharging unit and found no association. The influence of outliers on the association between DCBN and LOS increases also suggests that this group of children who have extremely long hospital stays might need further exploration.

Our study has some similar and some contrasting results with prior studies in adult patients. Our findings support the modeling literature that suggests DCBN may improve discharge efficiency by shortening patient LOS for some discharges.⁴ These findings contrast with Rajkomar et al., who reported that DCBN was associated with a longer LOS in adult patients.⁶ The contrasting findings could be due to differences in pediatric versus adult patients. Additionally, the population Rajkomar et al. studied was predominantly surgical patients, whose discharges may differ from medical patients’ in many aspects. Another possible explanation is that the Rajkomar et al. study was performed in a setting with clearly set institutional targets for DCBN, whereas, our institution lacked any hospital-wide DCBN initiatives or standards to which providers were held accountable. Some authors have argued setting DCBN as a measure of hospital quality perhaps creates the unintended consequence of providers holding potential afternoon or evening discharges until the next day so that they can be DCBN.^7,10 In that scenario, perhaps there would be a relationship between DCBN and longer LOS compared to patients who are reevaluated in the afternoon or evening and discharged. We did not find evidence of these effects in our analysis, however, understanding the potential for this is important when designing quality improvement efforts aimed at increasing discharge efficiency.

While shorter LOS can be an indicator of high-value care, the relationship between LOS, DCBN, and efficiency of discharge processes remains unclear. Prior studies have found evidence that multidisciplinary care teams with frequent care coordination rounds and integration of electronic admission order sets can be effective in improving discharge efficiency as measured by discharge within two hours of meeting discharge goals.^11,12 Measuring discharge efficiency on an ongoing basis is very difficult; however, easy-to-measure targets such as discharge before noon may be used as a proxy measure of efficiency. These targets also have “face validity,” and because of these two factors, measures like DCBN have been widely implemented even though evidence to support their validity is minimal.

Our study has several limitations. While we controlled for observable characteristics using covariates and propensity score weighted analyses, there are likely unobservable characteristics that confound our analysis. We did not measure other factors that may affect discharge time of day such as high occupancy, staffing levels, patient transportation availability, and patient and family preferences. Given these limitations, we caution against interpreting a causal relationship between independent variables and the outcome. Finally, this analysis was conducted at a single tertiary care, academic medical center. The majority of pediatric admissions at this institution are either transferred from other hospitals or scheduled admissions for medical or surgical care. A smaller proportion of discharges are acute, unplanned admissions through our emergency department in children with or without underlying medical complexity. These factors plus the exclusion of observation, extended recovery, and all the less than two-day stays in this study contribute to a relatively higher average LOS. These factors potentially limit generalizability to other care settings. Additionally, the majority of the care teams involve care by resident physicians, and they are often the primary caregivers and write the majority of orders in patient charts such as discharge orders. While we were not able to control for within resident physician similarities between patients, we did control for autocorrelation at the attending level.

The results of our study suggest that DCBN is associated with a decreased LOS for medical but not surgical pediatric patients. DCBN may not be an appropriate measure for all services. Further research should be done to identify other feasible but more valid indicators for shorter LOS.

Disclosures

The authors have no financial relationships relative to this article to disclose. The authors have no conflicts of interest relevant to this article to disclose.

Funding

There were no external sources of funding for this work.

Community-acquired pneumonia (CAP) is the most common infection in hospitalized patients and the eighth most common cause of death in the United States.¹ Mortality from CAP is estimated to be 5.1% in the outpatient population,13.6% in hospitalized patients, and 35.1% in patients admitted to the intensive care unit.^2,3 CAP accounts for more than 50,000 deaths annually in the United States.² There are multiple risk factors for CAP, including tobacco use, malnutrition, chronic obstructive pulmonary disease (COPD), bronchiectasis, cystic fibrosis, and mechanical bronchial obstruction. Underlying immunodeficiency, specifically humoral immunodeficiency, is also a risk factor for CAP.

Primary immunodeficiency (PIDD) is estimated to affect one in 1,800 individuals in the United States.⁴ The National Institutes of Health (NIH) estimates that only one out of three individuals with PIDD are appropriately diagnosed. Based on probability calculations on known PIDD patients versus incidence of disease, the NIH estimates that more than 500,000 individuals with PIDD remain undiagnosed in the United States.⁴ Further, there exists an average diagnostic delay of at least five years. This delay increases both morbidity and mortality and leads to increased healthcare utilization.^5,6

The most common form of primary immunodeficiency is due to humoral immunodeficiency, including selective IgA deficiency, specific antibody deficiency, and common variable immunodeficiency. Specific antibody deficiency is defined as a lack of response to polysaccharide antigens in the setting of low to normal Ig levels and an intact response to peptide antigens.⁷ Selective IgA deficiency is defined as the isolated deficiency of serum IgA in the setting of normal serum levels of IgG and IgM in an individual older than four years in whom other causes of hypogammaglobinemia have been excluded.⁸ Common variable immunodeficiency (CVID) is defined as a decreased serum concentration of IgG in combination with low levels of IgA and/or IgM with a poor or absent response to immunization in the absence of other defined immunodeficiency state.⁹ In addition to experiencing recurrent infections—namely bronchitis, sinusitis, otitis, and pneumonia—patients with CVID are also at increased risk of autoimmunity and malignancy. In adults, secondary immunodeficiency is more common than primary immunodeficiency. Secondary immunodeficiency occurs commonly with disease states like HIV infection, diabetes, cirrhosis, malnutrition, and autoimmune conditions.¹⁰ Additional causes of secondary immune defects due to humoral immunodeficiency include immune-modulating drugs—such as rituximab and ibrutinib—and hematologic malignancies, including chronic lymphocytic leukemia and multiple myeloma. Recurrent infections remain the leading cause of morbidity and mortality in patients with both primary and secondary immunodeficiency.^11,12

Evaluation of the humoral immune system begins with measurement of serum immunoglobulin (Ig) levels. Although abnormal Ig levels are not diagnostic of immunodeficiency, abnormal results may prompt additional evaluation. Screening strategies may assist in making an earlier diagnoses, potentially decreasing morbidity and mortality in patients with immunodeficiency.^13-15 To date, there have been no studies evaluating the utility of screening Ig levels to evaluate for underlying humoral immunodeficiency in patients hospitalized for CAP.

Study Design

This was a prospective cohort study conducted at Rochester General Hospital, a 528-bed tertiary care medical center, from February 2017 to April 2017. We enrolled 100 consecutive patients admitted to the inpatient internal medicine service with a physician diagnosis of CAP. Written consent was obtained from each patient. The study was approved by the institutional review board at Rochester General Hospital.

Case Definition

The following criteria were used to diagnose CAP: (1) Respiratory symptoms of productive cough or pleuritic chest pain, (2) Fever >38°C before or at the time of admission, and (3) chest imaging with infiltrate. Exclusion criteria included a diagnosis of hospital-acquired pneumonia, prior diagnosis of primary immunodeficiency, immunosuppression due to an underlying condition, such as HIV or malignancy, therapy with immunosuppressive medications including chemotherapy, Ig replacement within the past six months, or treatment with >10 mg prednisone for greater than 14 days before hospital admission.

Patients underwent an additional evaluation by a clinical immunologist if they met one of the following criteria: any hypergammaglobinemia (elevated IgG, IgM, or IgA), IgG hypogammaglobinemia <550 mg/dL, undetectable IgM or IgA, or if IgG, IgM, and IgA were all below the lower limit of normal.

CURB-65 was used for estimation of the severity of illness with CAP. The components of the score include age ≥65, confusion, BUN >19 mg/dl, respiratory rate ≥30 breaths per minute and systolic blood pressure <90 mm Hg or diastolic blood pressure ≤60 mm Hg. Each component is scored zero if absent or one if present. Predicted mortality ranges from 0.6% for a score of zero to 27.8% for a score of 5.

Data Collection

Patient health information including age, race, gender, medical history, admission notes, results of chest imaging studies, and relevant laboratory studies including serum levels of IgG, IgM, IgA, IgE on admission was obtained from the electronic medical health record. An additional evaluation by the immunologist occurred within three months of hospital discharge and included repeat Ig levels, pre- and postvaccination titers of polysaccharide and peptide antigens, serum protein electrophoresis, and B & T cell panels.

Description of Normal Levels

The normal levels of immunoglobulins were defined based on standard reference ranges at the laboratory at Rochester General Hospital; IgG (700-1,600 mg/dl), IgM (50-300 mg/dl), IgA (70-400 mg/dl), and IgE (0-378 IU/ml). Although there is no established classification regarding the degree of IgG hypogammaglobinemia,¹⁶ clinical immunologists commonly classify the severity of IgG hypogammaglobinemia as follows: mild (550-699 mg/dL), moderate (400-549 mg/dL), and severe (<400 mg/dL) IgG hypogammaglobinemia.

Statistical Analysis

Statistical analysis was performed using STATA software (StataCorp LLC, College Station, Texas). We conducted a Wilcoxon rank-sum test to compare the median difference in length of stay between groups with a low versus normal range of immunoglobulins. A Kruskal–Wallis test was performed to check for the median difference in IgG levels across degrees of illness severity (CURB-65 score categories). We conducted a simple linear regression analysis using the logarithmic data of the length of stay and IgG level variables. A chi-square test was used to determine the association between comorbidities and Ig levels.

Baseline Characteristics

There were 100 patients with CAP enrolled in this study with a median age of 65.04 ± 18.8, and 53% were female. Forty-seven patients reported a previous history of pneumonia and 18 reported a history of recurrent sinusitis or otitis media. Of the 100 enrolled patients, 46 had received pneumococcal polysaccharide vaccine (PPV23), 26 had received the 13-valent pneumococcal conjugate vaccine (PCV13), and 22 had received both (Table 1). The mean white blood cell count on admission was 12.9 ± 7 × 10³/uL with 75 ± 12.5% neutrophils. Total protein (6.5 ± 0.8) and albumin (3.7 ± 0.5) were within the normal range for the study population.

Immunoglobulin Analyses

The prevalence of hypogammaglobinemia in the study was 38% (95% CI: 28.47% to 48.25%). The median values of Ig levels for the entire study population and in patients with hypogammaglobinemia are summarized in Table 2.

• IgG hypogammaglobinemia (<700 mg/dl) was found in 27/100 patients, with a median level of 598 mg/dL, IQ range: 459-654. The median age in this group was 76.5 years, and 13 were female. Of these 27 patients, 10 had low IgM, four had low IgA, and four had an elevated IgE. In this group, 11 patients had received PPSV23, nine had received PCV13, and six had received both PPV23 and PCV13 before the index hospital admission.
• IgG hypergammaglobinemia (>1,600 mg/dl) was found in 9/100 patients, with a median level of 1,381 mg/dL, IQ range: 1,237-1,627. The median age was 61 years, and six were female. Of these nine patients, three had low IgM, one had low IgA, and four had elevated IgE.
• IgM hypogammaglobinemia (<50 mg/dl) was found in 23/100 patients with a median level of 38 mg/dL, IQ range: 25-43. In this group, the median age was 69 years, and 10 were female. Of these 23 patients, 10 had low IgG, and three had an elevated IgG.
• IgM hypergammaglobinemia (>300 mg/dl) was noted in two patients, with a median level of 491 mg/dL, IQ range: 418-564. Both patients were female, and one had elevated IgG.
• IgA hypogammaglobinemia (<70 mg/dl) was discovered in six patients, with a median level of 36 mg/dL, IQ range: 18-50. In this group, four patients had low IgG, four had low IgM, one had elevated IgE, and one had elevated IgG.
• IgA hypergammaglobinemia (>400 mg/dl) was noted in five patients, with a median level of 561 mg/dL, IQ range: 442-565: Two patients were female. Of these five patients, one had high IgG, and one had low IgG.

Length of Stay and Severity of Pneumonia

The median length of stay in the hospital for the entire study population was three days (IQ range: 2-5.5 days). Among patients with IgG hypogammaglobinemia, the median length of stay was two days longer as compared with patients who had IgG levels in the normal range (5 days, IQ range [3-10] vs 3days, IQ range [2-5], P = .0085).

The median CURB-65 score for the entire study population was two (IQ range: 1-3). The median CURB-65 score did not differ between patients with low and normal ranges of IgG levels (Median: 2, IQ range [1-3] vs Median: 1, IQ range [0-3], P = .2922). The CURB-65 score was not correlated with IgG levels (ρ = −0.0776, P = .4428). Length of stay, however, was positively correlated with CURB-65 score (ρ = .4673, P = .000)

A simple linear regression analysis using the logarithmic transformation of both length of stay and IgG level revealed a linear relationship between serum IgG levels and hospital length of stay (P = .0335, [R² = .0453]).

Comorbidities and New Diagnoses

No significant association was found between smoking status, obesity, COPD, asthma, diabetes mellitus, and hypogammaglobinemia.

Fourteen patients with abnormal Ig levels as defined by (1) the presence of hypergammaglobinemia (elevated IgG, IgM, or IgA), (2) IgG levels <550, (3) undetectable IgA or IgM, and (4) either IgG or both IgM and IgA below the lower limit of normal underwent further evaluation. Of these 14 patients, one was diagnosed with multiple myeloma, one with selective IgA deficiency, and three with specific antibody deficiency (Table 3).

Community-acquired pneumonia (CAP) is the most common infection in hospitalized patients and the eighth most common cause of death in the United States.¹ Mortality from CAP is estimated to be 5.1% in the outpatient population,13.6% in hospitalized patients, and 35.1% in patients admitted to the intensive care unit.^2,3 CAP accounts for more than 50,000 deaths annually in the United States.² There are multiple risk factors for CAP, including tobacco use, malnutrition, chronic obstructive pulmonary disease (COPD), bronchiectasis, cystic fibrosis, and mechanical bronchial obstruction. Underlying immunodeficiency, specifically humoral immunodeficiency, is also a risk factor for CAP.

Primary immunodeficiency (PIDD) is estimated to affect one in 1,800 individuals in the United States.⁴ The National Institutes of Health (NIH) estimates that only one out of three individuals with PIDD are appropriately diagnosed. Based on probability calculations on known PIDD patients versus incidence of disease, the NIH estimates that more than 500,000 individuals with PIDD remain undiagnosed in the United States.⁴ Further, there exists an average diagnostic delay of at least five years. This delay increases both morbidity and mortality and leads to increased healthcare utilization.^5,6

The most common form of primary immunodeficiency is due to humoral immunodeficiency, including selective IgA deficiency, specific antibody deficiency, and common variable immunodeficiency. Specific antibody deficiency is defined as a lack of response to polysaccharide antigens in the setting of low to normal Ig levels and an intact response to peptide antigens.⁷ Selective IgA deficiency is defined as the isolated deficiency of serum IgA in the setting of normal serum levels of IgG and IgM in an individual older than four years in whom other causes of hypogammaglobinemia have been excluded.⁸ Common variable immunodeficiency (CVID) is defined as a decreased serum concentration of IgG in combination with low levels of IgA and/or IgM with a poor or absent response to immunization in the absence of other defined immunodeficiency state.⁹ In addition to experiencing recurrent infections—namely bronchitis, sinusitis, otitis, and pneumonia—patients with CVID are also at increased risk of autoimmunity and malignancy. In adults, secondary immunodeficiency is more common than primary immunodeficiency. Secondary immunodeficiency occurs commonly with disease states like HIV infection, diabetes, cirrhosis, malnutrition, and autoimmune conditions.¹⁰ Additional causes of secondary immune defects due to humoral immunodeficiency include immune-modulating drugs—such as rituximab and ibrutinib—and hematologic malignancies, including chronic lymphocytic leukemia and multiple myeloma. Recurrent infections remain the leading cause of morbidity and mortality in patients with both primary and secondary immunodeficiency.^11,12

Evaluation of the humoral immune system begins with measurement of serum immunoglobulin (Ig) levels. Although abnormal Ig levels are not diagnostic of immunodeficiency, abnormal results may prompt additional evaluation. Screening strategies may assist in making an earlier diagnoses, potentially decreasing morbidity and mortality in patients with immunodeficiency.^13-15 To date, there have been no studies evaluating the utility of screening Ig levels to evaluate for underlying humoral immunodeficiency in patients hospitalized for CAP.

Study Design

This was a prospective cohort study conducted at Rochester General Hospital, a 528-bed tertiary care medical center, from February 2017 to April 2017. We enrolled 100 consecutive patients admitted to the inpatient internal medicine service with a physician diagnosis of CAP. Written consent was obtained from each patient. The study was approved by the institutional review board at Rochester General Hospital.

Case Definition

The following criteria were used to diagnose CAP: (1) Respiratory symptoms of productive cough or pleuritic chest pain, (2) Fever >38°C before or at the time of admission, and (3) chest imaging with infiltrate. Exclusion criteria included a diagnosis of hospital-acquired pneumonia, prior diagnosis of primary immunodeficiency, immunosuppression due to an underlying condition, such as HIV or malignancy, therapy with immunosuppressive medications including chemotherapy, Ig replacement within the past six months, or treatment with >10 mg prednisone for greater than 14 days before hospital admission.

Patients underwent an additional evaluation by a clinical immunologist if they met one of the following criteria: any hypergammaglobinemia (elevated IgG, IgM, or IgA), IgG hypogammaglobinemia <550 mg/dL, undetectable IgM or IgA, or if IgG, IgM, and IgA were all below the lower limit of normal.

CURB-65 was used for estimation of the severity of illness with CAP. The components of the score include age ≥65, confusion, BUN >19 mg/dl, respiratory rate ≥30 breaths per minute and systolic blood pressure <90 mm Hg or diastolic blood pressure ≤60 mm Hg. Each component is scored zero if absent or one if present. Predicted mortality ranges from 0.6% for a score of zero to 27.8% for a score of 5.

Data Collection

Patient health information including age, race, gender, medical history, admission notes, results of chest imaging studies, and relevant laboratory studies including serum levels of IgG, IgM, IgA, IgE on admission was obtained from the electronic medical health record. An additional evaluation by the immunologist occurred within three months of hospital discharge and included repeat Ig levels, pre- and postvaccination titers of polysaccharide and peptide antigens, serum protein electrophoresis, and B & T cell panels.

Description of Normal Levels

The normal levels of immunoglobulins were defined based on standard reference ranges at the laboratory at Rochester General Hospital; IgG (700-1,600 mg/dl), IgM (50-300 mg/dl), IgA (70-400 mg/dl), and IgE (0-378 IU/ml). Although there is no established classification regarding the degree of IgG hypogammaglobinemia,¹⁶ clinical immunologists commonly classify the severity of IgG hypogammaglobinemia as follows: mild (550-699 mg/dL), moderate (400-549 mg/dL), and severe (<400 mg/dL) IgG hypogammaglobinemia.

Statistical Analysis

Statistical analysis was performed using STATA software (StataCorp LLC, College Station, Texas). We conducted a Wilcoxon rank-sum test to compare the median difference in length of stay between groups with a low versus normal range of immunoglobulins. A Kruskal–Wallis test was performed to check for the median difference in IgG levels across degrees of illness severity (CURB-65 score categories). We conducted a simple linear regression analysis using the logarithmic data of the length of stay and IgG level variables. A chi-square test was used to determine the association between comorbidities and Ig levels.

Baseline Characteristics

There were 100 patients with CAP enrolled in this study with a median age of 65.04 ± 18.8, and 53% were female. Forty-seven patients reported a previous history of pneumonia and 18 reported a history of recurrent sinusitis or otitis media. Of the 100 enrolled patients, 46 had received pneumococcal polysaccharide vaccine (PPV23), 26 had received the 13-valent pneumococcal conjugate vaccine (PCV13), and 22 had received both (Table 1). The mean white blood cell count on admission was 12.9 ± 7 × 10³/uL with 75 ± 12.5% neutrophils. Total protein (6.5 ± 0.8) and albumin (3.7 ± 0.5) were within the normal range for the study population.

Immunoglobulin Analyses

The prevalence of hypogammaglobinemia in the study was 38% (95% CI: 28.47% to 48.25%). The median values of Ig levels for the entire study population and in patients with hypogammaglobinemia are summarized in Table 2.

• IgG hypogammaglobinemia (<700 mg/dl) was found in 27/100 patients, with a median level of 598 mg/dL, IQ range: 459-654. The median age in this group was 76.5 years, and 13 were female. Of these 27 patients, 10 had low IgM, four had low IgA, and four had an elevated IgE. In this group, 11 patients had received PPSV23, nine had received PCV13, and six had received both PPV23 and PCV13 before the index hospital admission.
• IgG hypergammaglobinemia (>1,600 mg/dl) was found in 9/100 patients, with a median level of 1,381 mg/dL, IQ range: 1,237-1,627. The median age was 61 years, and six were female. Of these nine patients, three had low IgM, one had low IgA, and four had elevated IgE.
• IgM hypogammaglobinemia (<50 mg/dl) was found in 23/100 patients with a median level of 38 mg/dL, IQ range: 25-43. In this group, the median age was 69 years, and 10 were female. Of these 23 patients, 10 had low IgG, and three had an elevated IgG.
• IgM hypergammaglobinemia (>300 mg/dl) was noted in two patients, with a median level of 491 mg/dL, IQ range: 418-564. Both patients were female, and one had elevated IgG.
• IgA hypogammaglobinemia (<70 mg/dl) was discovered in six patients, with a median level of 36 mg/dL, IQ range: 18-50. In this group, four patients had low IgG, four had low IgM, one had elevated IgE, and one had elevated IgG.
• IgA hypergammaglobinemia (>400 mg/dl) was noted in five patients, with a median level of 561 mg/dL, IQ range: 442-565: Two patients were female. Of these five patients, one had high IgG, and one had low IgG.

Length of Stay and Severity of Pneumonia

The median length of stay in the hospital for the entire study population was three days (IQ range: 2-5.5 days). Among patients with IgG hypogammaglobinemia, the median length of stay was two days longer as compared with patients who had IgG levels in the normal range (5 days, IQ range [3-10] vs 3days, IQ range [2-5], P = .0085).

The median CURB-65 score for the entire study population was two (IQ range: 1-3). The median CURB-65 score did not differ between patients with low and normal ranges of IgG levels (Median: 2, IQ range [1-3] vs Median: 1, IQ range [0-3], P = .2922). The CURB-65 score was not correlated with IgG levels (ρ = −0.0776, P = .4428). Length of stay, however, was positively correlated with CURB-65 score (ρ = .4673, P = .000)

A simple linear regression analysis using the logarithmic transformation of both length of stay and IgG level revealed a linear relationship between serum IgG levels and hospital length of stay (P = .0335, [R² = .0453]).

Comorbidities and New Diagnoses

No significant association was found between smoking status, obesity, COPD, asthma, diabetes mellitus, and hypogammaglobinemia.

Fourteen patients with abnormal Ig levels as defined by (1) the presence of hypergammaglobinemia (elevated IgG, IgM, or IgA), (2) IgG levels <550, (3) undetectable IgA or IgM, and (4) either IgG or both IgM and IgA below the lower limit of normal underwent further evaluation. Of these 14 patients, one was diagnosed with multiple myeloma, one with selective IgA deficiency, and three with specific antibody deficiency (Table 3).

Community-acquired pneumonia (CAP) is the most common infection in hospitalized patients and the eighth most common cause of death in the United States.¹ Mortality from CAP is estimated to be 5.1% in the outpatient population,13.6% in hospitalized patients, and 35.1% in patients admitted to the intensive care unit.^2,3 CAP accounts for more than 50,000 deaths annually in the United States.² There are multiple risk factors for CAP, including tobacco use, malnutrition, chronic obstructive pulmonary disease (COPD), bronchiectasis, cystic fibrosis, and mechanical bronchial obstruction. Underlying immunodeficiency, specifically humoral immunodeficiency, is also a risk factor for CAP.

Primary immunodeficiency (PIDD) is estimated to affect one in 1,800 individuals in the United States.⁴ The National Institutes of Health (NIH) estimates that only one out of three individuals with PIDD are appropriately diagnosed. Based on probability calculations on known PIDD patients versus incidence of disease, the NIH estimates that more than 500,000 individuals with PIDD remain undiagnosed in the United States.⁴ Further, there exists an average diagnostic delay of at least five years. This delay increases both morbidity and mortality and leads to increased healthcare utilization.^5,6

The most common form of primary immunodeficiency is due to humoral immunodeficiency, including selective IgA deficiency, specific antibody deficiency, and common variable immunodeficiency. Specific antibody deficiency is defined as a lack of response to polysaccharide antigens in the setting of low to normal Ig levels and an intact response to peptide antigens.⁷ Selective IgA deficiency is defined as the isolated deficiency of serum IgA in the setting of normal serum levels of IgG and IgM in an individual older than four years in whom other causes of hypogammaglobinemia have been excluded.⁸ Common variable immunodeficiency (CVID) is defined as a decreased serum concentration of IgG in combination with low levels of IgA and/or IgM with a poor or absent response to immunization in the absence of other defined immunodeficiency state.⁹ In addition to experiencing recurrent infections—namely bronchitis, sinusitis, otitis, and pneumonia—patients with CVID are also at increased risk of autoimmunity and malignancy. In adults, secondary immunodeficiency is more common than primary immunodeficiency. Secondary immunodeficiency occurs commonly with disease states like HIV infection, diabetes, cirrhosis, malnutrition, and autoimmune conditions.¹⁰ Additional causes of secondary immune defects due to humoral immunodeficiency include immune-modulating drugs—such as rituximab and ibrutinib—and hematologic malignancies, including chronic lymphocytic leukemia and multiple myeloma. Recurrent infections remain the leading cause of morbidity and mortality in patients with both primary and secondary immunodeficiency.^11,12

Evaluation of the humoral immune system begins with measurement of serum immunoglobulin (Ig) levels. Although abnormal Ig levels are not diagnostic of immunodeficiency, abnormal results may prompt additional evaluation. Screening strategies may assist in making an earlier diagnoses, potentially decreasing morbidity and mortality in patients with immunodeficiency.^13-15 To date, there have been no studies evaluating the utility of screening Ig levels to evaluate for underlying humoral immunodeficiency in patients hospitalized for CAP.

Study Design

This was a prospective cohort study conducted at Rochester General Hospital, a 528-bed tertiary care medical center, from February 2017 to April 2017. We enrolled 100 consecutive patients admitted to the inpatient internal medicine service with a physician diagnosis of CAP. Written consent was obtained from each patient. The study was approved by the institutional review board at Rochester General Hospital.

Case Definition

The following criteria were used to diagnose CAP: (1) Respiratory symptoms of productive cough or pleuritic chest pain, (2) Fever >38°C before or at the time of admission, and (3) chest imaging with infiltrate. Exclusion criteria included a diagnosis of hospital-acquired pneumonia, prior diagnosis of primary immunodeficiency, immunosuppression due to an underlying condition, such as HIV or malignancy, therapy with immunosuppressive medications including chemotherapy, Ig replacement within the past six months, or treatment with >10 mg prednisone for greater than 14 days before hospital admission.

Patients underwent an additional evaluation by a clinical immunologist if they met one of the following criteria: any hypergammaglobinemia (elevated IgG, IgM, or IgA), IgG hypogammaglobinemia <550 mg/dL, undetectable IgM or IgA, or if IgG, IgM, and IgA were all below the lower limit of normal.

CURB-65 was used for estimation of the severity of illness with CAP. The components of the score include age ≥65, confusion, BUN >19 mg/dl, respiratory rate ≥30 breaths per minute and systolic blood pressure <90 mm Hg or diastolic blood pressure ≤60 mm Hg. Each component is scored zero if absent or one if present. Predicted mortality ranges from 0.6% for a score of zero to 27.8% for a score of 5.

Data Collection

Patient health information including age, race, gender, medical history, admission notes, results of chest imaging studies, and relevant laboratory studies including serum levels of IgG, IgM, IgA, IgE on admission was obtained from the electronic medical health record. An additional evaluation by the immunologist occurred within three months of hospital discharge and included repeat Ig levels, pre- and postvaccination titers of polysaccharide and peptide antigens, serum protein electrophoresis, and B & T cell panels.

Description of Normal Levels

The normal levels of immunoglobulins were defined based on standard reference ranges at the laboratory at Rochester General Hospital; IgG (700-1,600 mg/dl), IgM (50-300 mg/dl), IgA (70-400 mg/dl), and IgE (0-378 IU/ml). Although there is no established classification regarding the degree of IgG hypogammaglobinemia,¹⁶ clinical immunologists commonly classify the severity of IgG hypogammaglobinemia as follows: mild (550-699 mg/dL), moderate (400-549 mg/dL), and severe (<400 mg/dL) IgG hypogammaglobinemia.

Statistical Analysis

Statistical analysis was performed using STATA software (StataCorp LLC, College Station, Texas). We conducted a Wilcoxon rank-sum test to compare the median difference in length of stay between groups with a low versus normal range of immunoglobulins. A Kruskal–Wallis test was performed to check for the median difference in IgG levels across degrees of illness severity (CURB-65 score categories). We conducted a simple linear regression analysis using the logarithmic data of the length of stay and IgG level variables. A chi-square test was used to determine the association between comorbidities and Ig levels.

Baseline Characteristics

There were 100 patients with CAP enrolled in this study with a median age of 65.04 ± 18.8, and 53% were female. Forty-seven patients reported a previous history of pneumonia and 18 reported a history of recurrent sinusitis or otitis media. Of the 100 enrolled patients, 46 had received pneumococcal polysaccharide vaccine (PPV23), 26 had received the 13-valent pneumococcal conjugate vaccine (PCV13), and 22 had received both (Table 1). The mean white blood cell count on admission was 12.9 ± 7 × 10³/uL with 75 ± 12.5% neutrophils. Total protein (6.5 ± 0.8) and albumin (3.7 ± 0.5) were within the normal range for the study population.

Immunoglobulin Analyses

The prevalence of hypogammaglobinemia in the study was 38% (95% CI: 28.47% to 48.25%). The median values of Ig levels for the entire study population and in patients with hypogammaglobinemia are summarized in Table 2.

• IgG hypogammaglobinemia (<700 mg/dl) was found in 27/100 patients, with a median level of 598 mg/dL, IQ range: 459-654. The median age in this group was 76.5 years, and 13 were female. Of these 27 patients, 10 had low IgM, four had low IgA, and four had an elevated IgE. In this group, 11 patients had received PPSV23, nine had received PCV13, and six had received both PPV23 and PCV13 before the index hospital admission.
• IgG hypergammaglobinemia (>1,600 mg/dl) was found in 9/100 patients, with a median level of 1,381 mg/dL, IQ range: 1,237-1,627. The median age was 61 years, and six were female. Of these nine patients, three had low IgM, one had low IgA, and four had elevated IgE.
• IgM hypogammaglobinemia (<50 mg/dl) was found in 23/100 patients with a median level of 38 mg/dL, IQ range: 25-43. In this group, the median age was 69 years, and 10 were female. Of these 23 patients, 10 had low IgG, and three had an elevated IgG.
• IgM hypergammaglobinemia (>300 mg/dl) was noted in two patients, with a median level of 491 mg/dL, IQ range: 418-564. Both patients were female, and one had elevated IgG.
• IgA hypogammaglobinemia (<70 mg/dl) was discovered in six patients, with a median level of 36 mg/dL, IQ range: 18-50. In this group, four patients had low IgG, four had low IgM, one had elevated IgE, and one had elevated IgG.
• IgA hypergammaglobinemia (>400 mg/dl) was noted in five patients, with a median level of 561 mg/dL, IQ range: 442-565: Two patients were female. Of these five patients, one had high IgG, and one had low IgG.

Length of Stay and Severity of Pneumonia

The median length of stay in the hospital for the entire study population was three days (IQ range: 2-5.5 days). Among patients with IgG hypogammaglobinemia, the median length of stay was two days longer as compared with patients who had IgG levels in the normal range (5 days, IQ range [3-10] vs 3days, IQ range [2-5], P = .0085).

The median CURB-65 score for the entire study population was two (IQ range: 1-3). The median CURB-65 score did not differ between patients with low and normal ranges of IgG levels (Median: 2, IQ range [1-3] vs Median: 1, IQ range [0-3], P = .2922). The CURB-65 score was not correlated with IgG levels (ρ = −0.0776, P = .4428). Length of stay, however, was positively correlated with CURB-65 score (ρ = .4673, P = .000)

A simple linear regression analysis using the logarithmic transformation of both length of stay and IgG level revealed a linear relationship between serum IgG levels and hospital length of stay (P = .0335, [R² = .0453]).

Comorbidities and New Diagnoses

No significant association was found between smoking status, obesity, COPD, asthma, diabetes mellitus, and hypogammaglobinemia.

Fourteen patients with abnormal Ig levels as defined by (1) the presence of hypergammaglobinemia (elevated IgG, IgM, or IgA), (2) IgG levels <550, (3) undetectable IgA or IgM, and (4) either IgG or both IgM and IgA below the lower limit of normal underwent further evaluation. Of these 14 patients, one was diagnosed with multiple myeloma, one with selective IgA deficiency, and three with specific antibody deficiency (Table 3).

Hospitals are complex adaptive systems within which practitioners, technology, physical resources, and other components adapt interdependently to attempt to best meet the needs of patients.¹ Hospitals must provide a stable, dependable level of care while also surging to respond to times of high demand, such as patient emergencies or swells in patient volume. Given the critical and resource-intensive nature of this work, optimizing the system is essential; however, because of the complexity of the system, making changes can result in unexpected and possibly deleterious effects. We need to approach change in hospital processes carefully and thoughtfully.

The Institute of Medicine, the National Academy of Engineering, and the President’s Council of Advisors on Science and Technology have recommended the application of systems engineering approaches to improve health care delivery.^2,3 Systems engineering seeks to coordinate, synchronize, and integrate complex systems of people, information, materials, technology, and financial resources.^4,5 To determine how complex systems can be improved, engineers apply analytic methods to describe how such systems operate and what the impact of changes might be. These methodologies have improved patient care and reduced costs at several hospitals.⁶ For example, a decision support system that combined simulation, optimization, and machine learning methods in an emergency department (ED) resulted in a 33% reduction in length of stay (LOS) and a 28% decrease in ED readmissions.⁷ Other strategies to improve patient flow include shaping demand (decreasing variation in surgical scheduling, relocating low acuity care ED visit to primary care, etc.), redesigning systems (early discharges, improving efficiency, and coordination of hospital discharge process, decreasing care variation, etc.), or aligning capacity and demand. Another approach, real-time demand capacity (RTDC), is based on management principles and queuing and constraint theory and has been implemented successfully in a variety of health care organizations. RTDC represents a promising approach to improve hospitalwide patient flow and can be integrated into current bed management processes.⁸ Unfortunately, many of these approaches are not well known to clinicians and would benefit from greater awareness and input from healthcare practitioners.

One systems engineering tool that can be used to describe, analyze, and evaluate proposed changes in care is simulation.⁹ Simulation creates a model within which what-if scenarios (ie, adjusting various inputs into the simulation) allow researchers to define the likelihood of consequences from various courses of action and determine the optimal change to a system. Such analyses can predict the impact of a proposed change on patients and healthcare practitioners.^10-13

A critical concern for hospitals that simulation may help address is managing the volume of inpatients. A high inpatient census is necessary for financial solvency, yet too high a census of inpatients or an unexpected surge in acuity can overwhelm hospital resources. Many hospitals, pressured by growing numbers of increasingly complex patients, have seen medical inpatients spread across multiple nonmedical nursing units (NUs) of their institution such that a particular medical team may have only a couple patients assigned to each nursing unit.¹⁴ This dispersion may hinder communication between physicians and nurses and limits the time physicians have to interact with patients.¹⁵ Additionally, coordination of care may become more challenging for discharge planning.¹⁶ Aligning medical teams with NUs may benefit the quality and efficiency of care or may create a barrier to patient flow, which worsens these problems.^15,17 Alternatively, hospitals might meet the increasing demands for care by choosing to add capacity by opening new NUs or hiring additional healthcare providers. We identified no studies in the literature that applied simulation modeling to general medicine inpatients to evaluate the impact of these different decisions.

This article describes the application of simulation to model the interconnected variables and subsequent future states created by several possible strategic decisions around the care of general medicine inpatients. Through the application of systems engineering techniques, we modeled four future states that illustrate the following: (1) the complexities of a large health delivery system, (2) the intended and unintended consequences of implementing different changes in the process of care delivery, and (3) how the simulation modeling might be used to inform decision making.

Setting and Present State

Virginia Commonwealth University (VCU) is a 865-bed tertiary academic medical center, with inpatient care activities spread between four connected buildings and 50 different NUs. The occupancy rate had been over 92% during the time period of this project with admission volume limited primarily by the capacity of the facility. Three of the NUs were primarily allocated to general medicine (GIM) patients. However, over the years, GIM inpatients grew to over 7500 admissions annually, resulting in nearly 50% of GIM patients being admitted to a non-GIM nursing unit.

Additionally, patients on each medical team had a high degree of spread across NUs due to several factors. Admissions and discharges from the hospital did not align across the day. While discharges clumped in the late afternoon, admission occurred throughout the day with a surge in the later afternoon. This mismatch frequently led to patients waiting in the ED for a bed, medical team, or both, and patients were typically assigned to the first available bed and team. For medical team assignments, newly admitted patients were distributed relatively equally across five hospitalist teams and five housestaff teams (that include residents, interns, and medical students). This steady distribution of patients through the day supported meeting housestaff work-hour restrictions of 80 hours each week.¹⁸ Yet, as a result of the high occupancy rate, the patterns of patient admissions and discharges, and the distribution of patients among medical teams and across NUs, medical teams and NUs rarely shared more than a few patients.

Leaders at our institution outlined several possible options to address these challenges, including aligning medical teams with NU, adding an additional hospitalist team, or adding an additional nursing unit. In addition, institutional leaders were concerned about the impact of continued growth in admission volume and the impact of patient dispersion on trainees and students. The overall goal of creating a simulation model was to determine the impact of an increased volume of patients and these possible strategic decisions on operational metrics, including number of patients waiting in the ED, ED boarding time per patient, time in system per patient (ED boarding time plus inpatient LOS), team utilization, and rounding travel time.

Simulation Modeling

To model the impact of some possible system changes on patient care, we applied Kelton and Law’s simulation study framework,¹⁹ including data collection; model building and validation; and what-if scenario testing (Figure 1).

Data Collection

Process Flow Map

We created a complex process flow map of patient care activities on medical teams. The map was developed by four general medicine physicians (R.C., H.M., V.M., and S.P.T.) who all provided medical care on the hospital-based services and ensured expert input on the patient care activities captured by the simulation modeling.

Time and Motion Studies

Time and motion study is a well-established technique used to evaluate the efficiency of work processes.^20,21 Originally applied to increase productivity in manufacturing, this technique uses first-hand observations to measure the time allotted to different work tasks to systematically analyze workflow.²² Workflow in healthcare, like manufacturing tasks, tends to have a repetitive pattern, making time and motion studies a highly applicable tool.

A research assistant observed a total of 30 hospitalist work cycles to describe the work of our inpatient clinicians. A work cycle, defined as one complete process flow,²³ began when the hospitalist started a daytime shift of patient care and concluded after the physician “signed out” to the physician who was assuming responsibility for ongoing medical care of the patients (ie, cross-coverage). Time spent on different activities identified by the process flow map was captured throughout the cycle. These activities included time spent traveling to evaluate patients located on different NUs. To minimize disruptions in patient care and adhere to privacy standards, no observations were conducted in patient rooms, and details of computer work were not recorded. To ensure stable estimates of the mean and standard deviation of the time spent at each step, at least 30 cycles of observation are recommended. Thus, 300 hours of observations over the course of 30 separate days were collected.

Hospital Data

We extracted admission and discharge data from the electronic health records (EHR) for general medicine patients admitted from the ED for the calendar year 2013. These records were used to establish means and standard deviations for admission date and time, distribution of patients across NUs, and LOS.

Model Building and Internal Validation

On the basis of these data inputs and using SIMIO® Simulation Software version 7, we constructed a discrete event simulation (DES) model representing the patient care activities of general medicine teams. Each patient was assigned a bed on a nursing unit through a probability distribution based on prior EHR data and then randomly assigned to a general medicine team. We replicated the model 200 times, and each model ran for 365 days. Each team was limited to 16 assigned patients, the maximum number of patients per housestaff team allowed by VCU protocol; henceforth, this number is referred to as team-patient capacity. The model assumed patients remained on the assigned nursing unit and medical team for the entirety of their hospital stay and that each patient was seen by their assigned medical team every day. The results of the present state model, including mean number of patients on each nursing unit, mean team census, patient dispersion (ie, the number of NUs on which each medical team had patients), and team utilization (ie, mean team census divided by team patient capacity), were compared with actual data from 2013 to internally validate the model.

What-If Scenario Testing

We constructed four what-if scenarios based on possible strategic directions identified by leadership. These models evaluated:

• constraining patients on housestaff (but not hospitalist) teams to the three general medicine NUs (Future State 1),
• increasing bed capacity for general medicine patients by adding one additional nursing unit of 26 beds (Future State 2),
• increasing the number of general medicine teams by adding one additional hospitalist team of up to 16 patients (Future State 3),
• modeling the impact of increased patient admissions from 21 per day to 25 per day while also adding a nursing unit and an additional medical team (Future State 4).

For Future States 1-3, admission volume was held constant. The model generated nursing unit LOS using a random continuous exponential probability distribution with a mean of 133 hours to match the LOS distribution derived from health system data. As patients entered the system for admission, the model assigned a bed to the patient, but the patient could not move to the assigned bed until a bed and care team were both available. We were only interested in the steady-state behavior of the system, so collecting performance statistics only after the model had been populated and steady state had been achieved was important.

Table 1 summarizes the input data, fixed, and dynamic variable for each future state model.

Time and Motion Data

The mean time spent with each patient was nine minutes. The mean time traveling between NUs Healthcare Quality for Children and Adolescents with Suicidality Admitted to Acute Care Hospitals in the United States was five minutes. Average rounding time was noted to be two hours, 53 minutes. Thirty-seven minutes, about ~21% of the time, was wasted in traveling. Each team, on average, traveled to seven different NUs to round on their daily census, averaging 1.6 patients in each nursing unit.

Hospital Data

Between January 1, 2011 to December 31, 2013, a total of 7,902 patients were admitted to the general medicine teams, spanning 23 NU. The average number of admissions per day was 21.6, and the average nursing unit LOS was 133 hours. Average team census was derived from historical data across all GIM team for 2013 and was noted to be 11.5 patients per team, and these patients were spread over seven NU.

Model Validation

The mean number of patients admitted to different NUs was estimated from the simulation model then compared with the EHR data from 2013. None were statistically different (P > .05), which signified that the validated simulation model is similar to the EHR data from 2013 despite the underlying assumptions.

Model Outputs

Analysis of the models indicated that steady-state (based upon hospital census) was realized at approximately 800 hours or after 680 patients were admitted to the GIM teams. Statistics collection, therefore, was started after 800 hours of simulated time and reflected the admission of the remaining 7222 patients in the model validation sample (Table 2).

In the model, the total daily general medicine patient census was 119.26. Average time in the system per patient was noted to be 147.37 hours, which was 14.37 hours more than the average nursing unit LOS of 133 hours. Average number of patients waiting for a bed was noted to be 11.31, while the average wait time for a patient to get a bed was 12.39 hours.

Average housestaff team and hospitalist team utilization were 76.06% and 73.02%, respectively, with average team utilization of 74.54% (range: 72.88%-76.19%). Housestaff team and hospitalist team averaged 12.17 and 11.68 patients per care team, respectively. General medicine teams had patients on 7.30 NUs on average. GIM teams rounding travel time was 36.5 minutes.

What-If Scenario Testing

Simulation outputs for the four future states are summarized in Table 2. With Future State 1, through which patients were selectively assigned to housestaff teams aligned with three NUs, the average time in the system per patient increased by 2.35 hours, with 1.87 more patients waiting for a bed and waiting for 2.03 more hours as compared with the present state. A marked disparity was observed in hospitalist and housestaff team utilization of 62.22% and 86.55% respectively. Patient dispersion to various NUs significantly decreased, and rounding time correspondingly decreased by approximately 41%.

Future State 2, adding a nursing unit, decreased average time in the system per patient by 9.86 hours, with 9.32 fewer patients waiting for a bed as compared with the present state. A slight increase in patient dispersion and rounding time was observed. Overall, patients spent 137.51 hours in the system, which demonstrated improved efficiency of the system.

Future State 3, adding an additional medical team, interestingly did not have a significant effect on patients’ average time in system or the number of patients waiting for a bed even though a decrease occurred in average team census, team utilization, and patient dispersion.

Finally, Future State 4, increasing admissions while also adding a nursing unit and a hospitalist team, resulted in an increase in admission volume while maintaining similar utilization rates for teams and NU. Patients spent about 2.48 hours less in the system, while only 9.94 patients were noted to be waiting for a bed as compared with 11.21 patients in the present state model. The total daily general medicine patient census was noted to be 137.19. Average team census and average team utilization were noted to be similar to those of the present state model, while admissions were up by approximately 1,080 per year. Both patient dispersion and rounding were slightly worsened.

Sensitivity Analysis

Overall, average time in system was most affected by the number of patient arrivals. This became particularly significant as the volume of patient arrivals approached and exceeded the capacity of the rounding teams. Adding a nursing unit had more impact on decreasing average time in the system than adding a medical team or aligning teams with NUs under the conditions defined by the model. However, under different conditions, such as increasing admission volume, the relative benefit of different approaches may vary.

Given that hospitals are large, complex systems,² the impact of system-level changes can have unpredictable and potentially deleterious effects. Simulation provides a technique for modeling the impact of changes to understand the ramifications of these interventions more thoroughly.³ In this study, we describe the process of building a simulation model for the admission and discharge of patients from general medicine services in a tertiary care hospital, internally validating this model, and examining the outcomes from several potential changes to the system.

The outcomes for these what-if scenarios provided some important insights about the secondary effect of system changes and the need for multiple, simultaneous interventions. Given that hospitals often function at near capacity, adding a hospitalist team or nursing unit might be seen as a reasonable strategy to improve the system metrics, number of patient discharges, or average LOS. On the basis of our analysis, adding a nursing unit would have more benefit than adding a hospitalist team. Leaders who want to increase capacity may need to consider both adding a hospitalist team and a nursing unit, and model the impact of each choice as described with a simulation.

Additionally, assigning patients to medical teams aligned with NUs seems theoretically appealing to improve interprofessional communication and decrease the time spent in transit between patients by physicians. While our findings supported a decrease in rounding time and patient dispersion, the teams not aligned with a nursing unit (ie, the hospitalists) exceeded 80% utilization, the threshold at which efficiency is known to decrease.²⁴ Potentially, benefits resulting from teams being aligned with NUs were offset by decrements in performance of the teams not aligned with NU. If medical teams and NUs become aligned, then a higher number of teams may be necessary to maintain patient throughput.

Simulation models identify these unexpected consequences prior to investing resources in a significant change; however, modeling is not simple. Simulation models depend on the characteristics of the model and the quality of the input data. For example, we used an expert approach to map physician workflow as an underpinning of the model, but we may have missed an important variation in physician workflow. Understanding this variation could strengthen the model and provide some testable variables for future study. Likewise, understanding nursing workflow and how variation in physician workflow shapes nursing workflow, and vice versa, is worth exploring.

Other data could also be added to, and help interpret, the outputs of this model. For example, the impact of various levels of team and unit utilization on diversion time for the hospital ED may help determine whether adding team capacity or unit capacity is more beneficial for the system. Likewise, aligning medical teams with NUs seems to hinder patient throughput on this analysis, but benefits in patient satisfaction or decreased readmissions might improve reimbursement and outweigh the revenue lost from throughput. Underpinning each of these types of decisions is a need to model the system well and thoughtfully choose the inputs, processes, and outputs. Pursuing a new strategic decision usually involves cost; simulation modeling provides data to help leaders weigh the benefits in terms of the needed investment.

The major limitations of the study stem from these choices. Our study focused on matching capacity and demand while limiting other changes in the system, such as changes in nursing unit LOS. Future work to quantify the relationship of other variables on parameters, such as the impact of decreased team dispersion on LOS, early discharges, and decreasing care variation, would make future models more robust. This model does not consider other strategies to improve patient flow, such as shaping demand, adaptive team assignment algorithms, or creating surge capacity. We also used only hospitalist time and motion data in our model; housestaff workflow is likely different. In addition, we modeled all patients as having a general level of nursing care and did not account for admissions or transfers to intensive care units or other services. These parameters could be added in future iterations. Finally, the biggest limitation in any simulation is the underlying assumptions made to construct the model. While we validated the model retrospectively, prospective validation and refinement should also be performed with attention to how the model functions under extreme conditions, such as a very high patient load.

Major system changes are expensive and must be made carefully. Systems engineering techniques, such as DES, provide techniques to estimate the impact of changes on pertinent care delivery variables. Results from this study underscore the complexity of patient care delivery and how simulation models can integrate multiple system components to provide a data-driven approach to inform decision making in a complex system.

Acknowledgments

The simulation software used in this study was awarded as an educational software grant from SIMIO®. We would like to acknowledge support from the Department of Internal Medicine at Virginia Commonwealth University for this project and thank Lena Rivera for her assistance with the manuscript preparation.

Dislosures

Dr. Heim recived a consulting fee for programming guidance from Virginia Commonwealth University. All other authors have nothing to disclose.

Hospitals are complex adaptive systems within which practitioners, technology, physical resources, and other components adapt interdependently to attempt to best meet the needs of patients.¹ Hospitals must provide a stable, dependable level of care while also surging to respond to times of high demand, such as patient emergencies or swells in patient volume. Given the critical and resource-intensive nature of this work, optimizing the system is essential; however, because of the complexity of the system, making changes can result in unexpected and possibly deleterious effects. We need to approach change in hospital processes carefully and thoughtfully.

The Institute of Medicine, the National Academy of Engineering, and the President’s Council of Advisors on Science and Technology have recommended the application of systems engineering approaches to improve health care delivery.^2,3 Systems engineering seeks to coordinate, synchronize, and integrate complex systems of people, information, materials, technology, and financial resources.^4,5 To determine how complex systems can be improved, engineers apply analytic methods to describe how such systems operate and what the impact of changes might be. These methodologies have improved patient care and reduced costs at several hospitals.⁶ For example, a decision support system that combined simulation, optimization, and machine learning methods in an emergency department (ED) resulted in a 33% reduction in length of stay (LOS) and a 28% decrease in ED readmissions.⁷ Other strategies to improve patient flow include shaping demand (decreasing variation in surgical scheduling, relocating low acuity care ED visit to primary care, etc.), redesigning systems (early discharges, improving efficiency, and coordination of hospital discharge process, decreasing care variation, etc.), or aligning capacity and demand. Another approach, real-time demand capacity (RTDC), is based on management principles and queuing and constraint theory and has been implemented successfully in a variety of health care organizations. RTDC represents a promising approach to improve hospitalwide patient flow and can be integrated into current bed management processes.⁸ Unfortunately, many of these approaches are not well known to clinicians and would benefit from greater awareness and input from healthcare practitioners.

One systems engineering tool that can be used to describe, analyze, and evaluate proposed changes in care is simulation.⁹ Simulation creates a model within which what-if scenarios (ie, adjusting various inputs into the simulation) allow researchers to define the likelihood of consequences from various courses of action and determine the optimal change to a system. Such analyses can predict the impact of a proposed change on patients and healthcare practitioners.^10-13

A critical concern for hospitals that simulation may help address is managing the volume of inpatients. A high inpatient census is necessary for financial solvency, yet too high a census of inpatients or an unexpected surge in acuity can overwhelm hospital resources. Many hospitals, pressured by growing numbers of increasingly complex patients, have seen medical inpatients spread across multiple nonmedical nursing units (NUs) of their institution such that a particular medical team may have only a couple patients assigned to each nursing unit.¹⁴ This dispersion may hinder communication between physicians and nurses and limits the time physicians have to interact with patients.¹⁵ Additionally, coordination of care may become more challenging for discharge planning.¹⁶ Aligning medical teams with NUs may benefit the quality and efficiency of care or may create a barrier to patient flow, which worsens these problems.^15,17 Alternatively, hospitals might meet the increasing demands for care by choosing to add capacity by opening new NUs or hiring additional healthcare providers. We identified no studies in the literature that applied simulation modeling to general medicine inpatients to evaluate the impact of these different decisions.

This article describes the application of simulation to model the interconnected variables and subsequent future states created by several possible strategic decisions around the care of general medicine inpatients. Through the application of systems engineering techniques, we modeled four future states that illustrate the following: (1) the complexities of a large health delivery system, (2) the intended and unintended consequences of implementing different changes in the process of care delivery, and (3) how the simulation modeling might be used to inform decision making.

Setting and Present State

Virginia Commonwealth University (VCU) is a 865-bed tertiary academic medical center, with inpatient care activities spread between four connected buildings and 50 different NUs. The occupancy rate had been over 92% during the time period of this project with admission volume limited primarily by the capacity of the facility. Three of the NUs were primarily allocated to general medicine (GIM) patients. However, over the years, GIM inpatients grew to over 7500 admissions annually, resulting in nearly 50% of GIM patients being admitted to a non-GIM nursing unit.

Additionally, patients on each medical team had a high degree of spread across NUs due to several factors. Admissions and discharges from the hospital did not align across the day. While discharges clumped in the late afternoon, admission occurred throughout the day with a surge in the later afternoon. This mismatch frequently led to patients waiting in the ED for a bed, medical team, or both, and patients were typically assigned to the first available bed and team. For medical team assignments, newly admitted patients were distributed relatively equally across five hospitalist teams and five housestaff teams (that include residents, interns, and medical students). This steady distribution of patients through the day supported meeting housestaff work-hour restrictions of 80 hours each week.¹⁸ Yet, as a result of the high occupancy rate, the patterns of patient admissions and discharges, and the distribution of patients among medical teams and across NUs, medical teams and NUs rarely shared more than a few patients.

Leaders at our institution outlined several possible options to address these challenges, including aligning medical teams with NU, adding an additional hospitalist team, or adding an additional nursing unit. In addition, institutional leaders were concerned about the impact of continued growth in admission volume and the impact of patient dispersion on trainees and students. The overall goal of creating a simulation model was to determine the impact of an increased volume of patients and these possible strategic decisions on operational metrics, including number of patients waiting in the ED, ED boarding time per patient, time in system per patient (ED boarding time plus inpatient LOS), team utilization, and rounding travel time.

Simulation Modeling

To model the impact of some possible system changes on patient care, we applied Kelton and Law’s simulation study framework,¹⁹ including data collection; model building and validation; and what-if scenario testing (Figure 1).

Data Collection

Process Flow Map

We created a complex process flow map of patient care activities on medical teams. The map was developed by four general medicine physicians (R.C., H.M., V.M., and S.P.T.) who all provided medical care on the hospital-based services and ensured expert input on the patient care activities captured by the simulation modeling.

Time and Motion Studies

Time and motion study is a well-established technique used to evaluate the efficiency of work processes.^20,21 Originally applied to increase productivity in manufacturing, this technique uses first-hand observations to measure the time allotted to different work tasks to systematically analyze workflow.²² Workflow in healthcare, like manufacturing tasks, tends to have a repetitive pattern, making time and motion studies a highly applicable tool.

A research assistant observed a total of 30 hospitalist work cycles to describe the work of our inpatient clinicians. A work cycle, defined as one complete process flow,²³ began when the hospitalist started a daytime shift of patient care and concluded after the physician “signed out” to the physician who was assuming responsibility for ongoing medical care of the patients (ie, cross-coverage). Time spent on different activities identified by the process flow map was captured throughout the cycle. These activities included time spent traveling to evaluate patients located on different NUs. To minimize disruptions in patient care and adhere to privacy standards, no observations were conducted in patient rooms, and details of computer work were not recorded. To ensure stable estimates of the mean and standard deviation of the time spent at each step, at least 30 cycles of observation are recommended. Thus, 300 hours of observations over the course of 30 separate days were collected.

Hospital Data

We extracted admission and discharge data from the electronic health records (EHR) for general medicine patients admitted from the ED for the calendar year 2013. These records were used to establish means and standard deviations for admission date and time, distribution of patients across NUs, and LOS.

Model Building and Internal Validation

On the basis of these data inputs and using SIMIO® Simulation Software version 7, we constructed a discrete event simulation (DES) model representing the patient care activities of general medicine teams. Each patient was assigned a bed on a nursing unit through a probability distribution based on prior EHR data and then randomly assigned to a general medicine team. We replicated the model 200 times, and each model ran for 365 days. Each team was limited to 16 assigned patients, the maximum number of patients per housestaff team allowed by VCU protocol; henceforth, this number is referred to as team-patient capacity. The model assumed patients remained on the assigned nursing unit and medical team for the entirety of their hospital stay and that each patient was seen by their assigned medical team every day. The results of the present state model, including mean number of patients on each nursing unit, mean team census, patient dispersion (ie, the number of NUs on which each medical team had patients), and team utilization (ie, mean team census divided by team patient capacity), were compared with actual data from 2013 to internally validate the model.

What-If Scenario Testing

We constructed four what-if scenarios based on possible strategic directions identified by leadership. These models evaluated:

• constraining patients on housestaff (but not hospitalist) teams to the three general medicine NUs (Future State 1),
• increasing bed capacity for general medicine patients by adding one additional nursing unit of 26 beds (Future State 2),
• increasing the number of general medicine teams by adding one additional hospitalist team of up to 16 patients (Future State 3),
• modeling the impact of increased patient admissions from 21 per day to 25 per day while also adding a nursing unit and an additional medical team (Future State 4).

For Future States 1-3, admission volume was held constant. The model generated nursing unit LOS using a random continuous exponential probability distribution with a mean of 133 hours to match the LOS distribution derived from health system data. As patients entered the system for admission, the model assigned a bed to the patient, but the patient could not move to the assigned bed until a bed and care team were both available. We were only interested in the steady-state behavior of the system, so collecting performance statistics only after the model had been populated and steady state had been achieved was important.

Table 1 summarizes the input data, fixed, and dynamic variable for each future state model.

Time and Motion Data

The mean time spent with each patient was nine minutes. The mean time traveling between NUs Healthcare Quality for Children and Adolescents with Suicidality Admitted to Acute Care Hospitals in the United States was five minutes. Average rounding time was noted to be two hours, 53 minutes. Thirty-seven minutes, about ~21% of the time, was wasted in traveling. Each team, on average, traveled to seven different NUs to round on their daily census, averaging 1.6 patients in each nursing unit.

Hospital Data

Between January 1, 2011 to December 31, 2013, a total of 7,902 patients were admitted to the general medicine teams, spanning 23 NU. The average number of admissions per day was 21.6, and the average nursing unit LOS was 133 hours. Average team census was derived from historical data across all GIM team for 2013 and was noted to be 11.5 patients per team, and these patients were spread over seven NU.

Model Validation

The mean number of patients admitted to different NUs was estimated from the simulation model then compared with the EHR data from 2013. None were statistically different (P > .05), which signified that the validated simulation model is similar to the EHR data from 2013 despite the underlying assumptions.

Model Outputs

Analysis of the models indicated that steady-state (based upon hospital census) was realized at approximately 800 hours or after 680 patients were admitted to the GIM teams. Statistics collection, therefore, was started after 800 hours of simulated time and reflected the admission of the remaining 7222 patients in the model validation sample (Table 2).

In the model, the total daily general medicine patient census was 119.26. Average time in the system per patient was noted to be 147.37 hours, which was 14.37 hours more than the average nursing unit LOS of 133 hours. Average number of patients waiting for a bed was noted to be 11.31, while the average wait time for a patient to get a bed was 12.39 hours.

Average housestaff team and hospitalist team utilization were 76.06% and 73.02%, respectively, with average team utilization of 74.54% (range: 72.88%-76.19%). Housestaff team and hospitalist team averaged 12.17 and 11.68 patients per care team, respectively. General medicine teams had patients on 7.30 NUs on average. GIM teams rounding travel time was 36.5 minutes.

What-If Scenario Testing

Simulation outputs for the four future states are summarized in Table 2. With Future State 1, through which patients were selectively assigned to housestaff teams aligned with three NUs, the average time in the system per patient increased by 2.35 hours, with 1.87 more patients waiting for a bed and waiting for 2.03 more hours as compared with the present state. A marked disparity was observed in hospitalist and housestaff team utilization of 62.22% and 86.55% respectively. Patient dispersion to various NUs significantly decreased, and rounding time correspondingly decreased by approximately 41%.

Future State 2, adding a nursing unit, decreased average time in the system per patient by 9.86 hours, with 9.32 fewer patients waiting for a bed as compared with the present state. A slight increase in patient dispersion and rounding time was observed. Overall, patients spent 137.51 hours in the system, which demonstrated improved efficiency of the system.

Future State 3, adding an additional medical team, interestingly did not have a significant effect on patients’ average time in system or the number of patients waiting for a bed even though a decrease occurred in average team census, team utilization, and patient dispersion.

Finally, Future State 4, increasing admissions while also adding a nursing unit and a hospitalist team, resulted in an increase in admission volume while maintaining similar utilization rates for teams and NU. Patients spent about 2.48 hours less in the system, while only 9.94 patients were noted to be waiting for a bed as compared with 11.21 patients in the present state model. The total daily general medicine patient census was noted to be 137.19. Average team census and average team utilization were noted to be similar to those of the present state model, while admissions were up by approximately 1,080 per year. Both patient dispersion and rounding were slightly worsened.

Sensitivity Analysis

Overall, average time in system was most affected by the number of patient arrivals. This became particularly significant as the volume of patient arrivals approached and exceeded the capacity of the rounding teams. Adding a nursing unit had more impact on decreasing average time in the system than adding a medical team or aligning teams with NUs under the conditions defined by the model. However, under different conditions, such as increasing admission volume, the relative benefit of different approaches may vary.

Given that hospitals are large, complex systems,² the impact of system-level changes can have unpredictable and potentially deleterious effects. Simulation provides a technique for modeling the impact of changes to understand the ramifications of these interventions more thoroughly.³ In this study, we describe the process of building a simulation model for the admission and discharge of patients from general medicine services in a tertiary care hospital, internally validating this model, and examining the outcomes from several potential changes to the system.

The outcomes for these what-if scenarios provided some important insights about the secondary effect of system changes and the need for multiple, simultaneous interventions. Given that hospitals often function at near capacity, adding a hospitalist team or nursing unit might be seen as a reasonable strategy to improve the system metrics, number of patient discharges, or average LOS. On the basis of our analysis, adding a nursing unit would have more benefit than adding a hospitalist team. Leaders who want to increase capacity may need to consider both adding a hospitalist team and a nursing unit, and model the impact of each choice as described with a simulation.

Additionally, assigning patients to medical teams aligned with NUs seems theoretically appealing to improve interprofessional communication and decrease the time spent in transit between patients by physicians. While our findings supported a decrease in rounding time and patient dispersion, the teams not aligned with a nursing unit (ie, the hospitalists) exceeded 80% utilization, the threshold at which efficiency is known to decrease.²⁴ Potentially, benefits resulting from teams being aligned with NUs were offset by decrements in performance of the teams not aligned with NU. If medical teams and NUs become aligned, then a higher number of teams may be necessary to maintain patient throughput.

Simulation models identify these unexpected consequences prior to investing resources in a significant change; however, modeling is not simple. Simulation models depend on the characteristics of the model and the quality of the input data. For example, we used an expert approach to map physician workflow as an underpinning of the model, but we may have missed an important variation in physician workflow. Understanding this variation could strengthen the model and provide some testable variables for future study. Likewise, understanding nursing workflow and how variation in physician workflow shapes nursing workflow, and vice versa, is worth exploring.

Other data could also be added to, and help interpret, the outputs of this model. For example, the impact of various levels of team and unit utilization on diversion time for the hospital ED may help determine whether adding team capacity or unit capacity is more beneficial for the system. Likewise, aligning medical teams with NUs seems to hinder patient throughput on this analysis, but benefits in patient satisfaction or decreased readmissions might improve reimbursement and outweigh the revenue lost from throughput. Underpinning each of these types of decisions is a need to model the system well and thoughtfully choose the inputs, processes, and outputs. Pursuing a new strategic decision usually involves cost; simulation modeling provides data to help leaders weigh the benefits in terms of the needed investment.

The major limitations of the study stem from these choices. Our study focused on matching capacity and demand while limiting other changes in the system, such as changes in nursing unit LOS. Future work to quantify the relationship of other variables on parameters, such as the impact of decreased team dispersion on LOS, early discharges, and decreasing care variation, would make future models more robust. This model does not consider other strategies to improve patient flow, such as shaping demand, adaptive team assignment algorithms, or creating surge capacity. We also used only hospitalist time and motion data in our model; housestaff workflow is likely different. In addition, we modeled all patients as having a general level of nursing care and did not account for admissions or transfers to intensive care units or other services. These parameters could be added in future iterations. Finally, the biggest limitation in any simulation is the underlying assumptions made to construct the model. While we validated the model retrospectively, prospective validation and refinement should also be performed with attention to how the model functions under extreme conditions, such as a very high patient load.

Major system changes are expensive and must be made carefully. Systems engineering techniques, such as DES, provide techniques to estimate the impact of changes on pertinent care delivery variables. Results from this study underscore the complexity of patient care delivery and how simulation models can integrate multiple system components to provide a data-driven approach to inform decision making in a complex system.

Acknowledgments

The simulation software used in this study was awarded as an educational software grant from SIMIO®. We would like to acknowledge support from the Department of Internal Medicine at Virginia Commonwealth University for this project and thank Lena Rivera for her assistance with the manuscript preparation.

Dislosures

Dr. Heim recived a consulting fee for programming guidance from Virginia Commonwealth University. All other authors have nothing to disclose.

Hospitals are complex adaptive systems within which practitioners, technology, physical resources, and other components adapt interdependently to attempt to best meet the needs of patients.¹ Hospitals must provide a stable, dependable level of care while also surging to respond to times of high demand, such as patient emergencies or swells in patient volume. Given the critical and resource-intensive nature of this work, optimizing the system is essential; however, because of the complexity of the system, making changes can result in unexpected and possibly deleterious effects. We need to approach change in hospital processes carefully and thoughtfully.

The Institute of Medicine, the National Academy of Engineering, and the President’s Council of Advisors on Science and Technology have recommended the application of systems engineering approaches to improve health care delivery.^2,3 Systems engineering seeks to coordinate, synchronize, and integrate complex systems of people, information, materials, technology, and financial resources.^4,5 To determine how complex systems can be improved, engineers apply analytic methods to describe how such systems operate and what the impact of changes might be. These methodologies have improved patient care and reduced costs at several hospitals.⁶ For example, a decision support system that combined simulation, optimization, and machine learning methods in an emergency department (ED) resulted in a 33% reduction in length of stay (LOS) and a 28% decrease in ED readmissions.⁷ Other strategies to improve patient flow include shaping demand (decreasing variation in surgical scheduling, relocating low acuity care ED visit to primary care, etc.), redesigning systems (early discharges, improving efficiency, and coordination of hospital discharge process, decreasing care variation, etc.), or aligning capacity and demand. Another approach, real-time demand capacity (RTDC), is based on management principles and queuing and constraint theory and has been implemented successfully in a variety of health care organizations. RTDC represents a promising approach to improve hospitalwide patient flow and can be integrated into current bed management processes.⁸ Unfortunately, many of these approaches are not well known to clinicians and would benefit from greater awareness and input from healthcare practitioners.

One systems engineering tool that can be used to describe, analyze, and evaluate proposed changes in care is simulation.⁹ Simulation creates a model within which what-if scenarios (ie, adjusting various inputs into the simulation) allow researchers to define the likelihood of consequences from various courses of action and determine the optimal change to a system. Such analyses can predict the impact of a proposed change on patients and healthcare practitioners.^10-13

A critical concern for hospitals that simulation may help address is managing the volume of inpatients. A high inpatient census is necessary for financial solvency, yet too high a census of inpatients or an unexpected surge in acuity can overwhelm hospital resources. Many hospitals, pressured by growing numbers of increasingly complex patients, have seen medical inpatients spread across multiple nonmedical nursing units (NUs) of their institution such that a particular medical team may have only a couple patients assigned to each nursing unit.¹⁴ This dispersion may hinder communication between physicians and nurses and limits the time physicians have to interact with patients.¹⁵ Additionally, coordination of care may become more challenging for discharge planning.¹⁶ Aligning medical teams with NUs may benefit the quality and efficiency of care or may create a barrier to patient flow, which worsens these problems.^15,17 Alternatively, hospitals might meet the increasing demands for care by choosing to add capacity by opening new NUs or hiring additional healthcare providers. We identified no studies in the literature that applied simulation modeling to general medicine inpatients to evaluate the impact of these different decisions.

This article describes the application of simulation to model the interconnected variables and subsequent future states created by several possible strategic decisions around the care of general medicine inpatients. Through the application of systems engineering techniques, we modeled four future states that illustrate the following: (1) the complexities of a large health delivery system, (2) the intended and unintended consequences of implementing different changes in the process of care delivery, and (3) how the simulation modeling might be used to inform decision making.

Setting and Present State

Virginia Commonwealth University (VCU) is a 865-bed tertiary academic medical center, with inpatient care activities spread between four connected buildings and 50 different NUs. The occupancy rate had been over 92% during the time period of this project with admission volume limited primarily by the capacity of the facility. Three of the NUs were primarily allocated to general medicine (GIM) patients. However, over the years, GIM inpatients grew to over 7500 admissions annually, resulting in nearly 50% of GIM patients being admitted to a non-GIM nursing unit.

Additionally, patients on each medical team had a high degree of spread across NUs due to several factors. Admissions and discharges from the hospital did not align across the day. While discharges clumped in the late afternoon, admission occurred throughout the day with a surge in the later afternoon. This mismatch frequently led to patients waiting in the ED for a bed, medical team, or both, and patients were typically assigned to the first available bed and team. For medical team assignments, newly admitted patients were distributed relatively equally across five hospitalist teams and five housestaff teams (that include residents, interns, and medical students). This steady distribution of patients through the day supported meeting housestaff work-hour restrictions of 80 hours each week.¹⁸ Yet, as a result of the high occupancy rate, the patterns of patient admissions and discharges, and the distribution of patients among medical teams and across NUs, medical teams and NUs rarely shared more than a few patients.

Leaders at our institution outlined several possible options to address these challenges, including aligning medical teams with NU, adding an additional hospitalist team, or adding an additional nursing unit. In addition, institutional leaders were concerned about the impact of continued growth in admission volume and the impact of patient dispersion on trainees and students. The overall goal of creating a simulation model was to determine the impact of an increased volume of patients and these possible strategic decisions on operational metrics, including number of patients waiting in the ED, ED boarding time per patient, time in system per patient (ED boarding time plus inpatient LOS), team utilization, and rounding travel time.

Simulation Modeling

To model the impact of some possible system changes on patient care, we applied Kelton and Law’s simulation study framework,¹⁹ including data collection; model building and validation; and what-if scenario testing (Figure 1).

Data Collection

Process Flow Map

We created a complex process flow map of patient care activities on medical teams. The map was developed by four general medicine physicians (R.C., H.M., V.M., and S.P.T.) who all provided medical care on the hospital-based services and ensured expert input on the patient care activities captured by the simulation modeling.

Time and Motion Studies

Time and motion study is a well-established technique used to evaluate the efficiency of work processes.^20,21 Originally applied to increase productivity in manufacturing, this technique uses first-hand observations to measure the time allotted to different work tasks to systematically analyze workflow.²² Workflow in healthcare, like manufacturing tasks, tends to have a repetitive pattern, making time and motion studies a highly applicable tool.

A research assistant observed a total of 30 hospitalist work cycles to describe the work of our inpatient clinicians. A work cycle, defined as one complete process flow,²³ began when the hospitalist started a daytime shift of patient care and concluded after the physician “signed out” to the physician who was assuming responsibility for ongoing medical care of the patients (ie, cross-coverage). Time spent on different activities identified by the process flow map was captured throughout the cycle. These activities included time spent traveling to evaluate patients located on different NUs. To minimize disruptions in patient care and adhere to privacy standards, no observations were conducted in patient rooms, and details of computer work were not recorded. To ensure stable estimates of the mean and standard deviation of the time spent at each step, at least 30 cycles of observation are recommended. Thus, 300 hours of observations over the course of 30 separate days were collected.

Hospital Data

We extracted admission and discharge data from the electronic health records (EHR) for general medicine patients admitted from the ED for the calendar year 2013. These records were used to establish means and standard deviations for admission date and time, distribution of patients across NUs, and LOS.

Model Building and Internal Validation

On the basis of these data inputs and using SIMIO® Simulation Software version 7, we constructed a discrete event simulation (DES) model representing the patient care activities of general medicine teams. Each patient was assigned a bed on a nursing unit through a probability distribution based on prior EHR data and then randomly assigned to a general medicine team. We replicated the model 200 times, and each model ran for 365 days. Each team was limited to 16 assigned patients, the maximum number of patients per housestaff team allowed by VCU protocol; henceforth, this number is referred to as team-patient capacity. The model assumed patients remained on the assigned nursing unit and medical team for the entirety of their hospital stay and that each patient was seen by their assigned medical team every day. The results of the present state model, including mean number of patients on each nursing unit, mean team census, patient dispersion (ie, the number of NUs on which each medical team had patients), and team utilization (ie, mean team census divided by team patient capacity), were compared with actual data from 2013 to internally validate the model.

What-If Scenario Testing

We constructed four what-if scenarios based on possible strategic directions identified by leadership. These models evaluated:

• constraining patients on housestaff (but not hospitalist) teams to the three general medicine NUs (Future State 1),
• increasing bed capacity for general medicine patients by adding one additional nursing unit of 26 beds (Future State 2),
• increasing the number of general medicine teams by adding one additional hospitalist team of up to 16 patients (Future State 3),
• modeling the impact of increased patient admissions from 21 per day to 25 per day while also adding a nursing unit and an additional medical team (Future State 4).

For Future States 1-3, admission volume was held constant. The model generated nursing unit LOS using a random continuous exponential probability distribution with a mean of 133 hours to match the LOS distribution derived from health system data. As patients entered the system for admission, the model assigned a bed to the patient, but the patient could not move to the assigned bed until a bed and care team were both available. We were only interested in the steady-state behavior of the system, so collecting performance statistics only after the model had been populated and steady state had been achieved was important.

Table 1 summarizes the input data, fixed, and dynamic variable for each future state model.

Time and Motion Data

The mean time spent with each patient was nine minutes. The mean time traveling between NUs Healthcare Quality for Children and Adolescents with Suicidality Admitted to Acute Care Hospitals in the United States was five minutes. Average rounding time was noted to be two hours, 53 minutes. Thirty-seven minutes, about ~21% of the time, was wasted in traveling. Each team, on average, traveled to seven different NUs to round on their daily census, averaging 1.6 patients in each nursing unit.

Hospital Data

Between January 1, 2011 to December 31, 2013, a total of 7,902 patients were admitted to the general medicine teams, spanning 23 NU. The average number of admissions per day was 21.6, and the average nursing unit LOS was 133 hours. Average team census was derived from historical data across all GIM team for 2013 and was noted to be 11.5 patients per team, and these patients were spread over seven NU.

Model Validation

The mean number of patients admitted to different NUs was estimated from the simulation model then compared with the EHR data from 2013. None were statistically different (P > .05), which signified that the validated simulation model is similar to the EHR data from 2013 despite the underlying assumptions.

Model Outputs

Analysis of the models indicated that steady-state (based upon hospital census) was realized at approximately 800 hours or after 680 patients were admitted to the GIM teams. Statistics collection, therefore, was started after 800 hours of simulated time and reflected the admission of the remaining 7222 patients in the model validation sample (Table 2).

In the model, the total daily general medicine patient census was 119.26. Average time in the system per patient was noted to be 147.37 hours, which was 14.37 hours more than the average nursing unit LOS of 133 hours. Average number of patients waiting for a bed was noted to be 11.31, while the average wait time for a patient to get a bed was 12.39 hours.

Average housestaff team and hospitalist team utilization were 76.06% and 73.02%, respectively, with average team utilization of 74.54% (range: 72.88%-76.19%). Housestaff team and hospitalist team averaged 12.17 and 11.68 patients per care team, respectively. General medicine teams had patients on 7.30 NUs on average. GIM teams rounding travel time was 36.5 minutes.

What-If Scenario Testing

Simulation outputs for the four future states are summarized in Table 2. With Future State 1, through which patients were selectively assigned to housestaff teams aligned with three NUs, the average time in the system per patient increased by 2.35 hours, with 1.87 more patients waiting for a bed and waiting for 2.03 more hours as compared with the present state. A marked disparity was observed in hospitalist and housestaff team utilization of 62.22% and 86.55% respectively. Patient dispersion to various NUs significantly decreased, and rounding time correspondingly decreased by approximately 41%.

Future State 2, adding a nursing unit, decreased average time in the system per patient by 9.86 hours, with 9.32 fewer patients waiting for a bed as compared with the present state. A slight increase in patient dispersion and rounding time was observed. Overall, patients spent 137.51 hours in the system, which demonstrated improved efficiency of the system.

Future State 3, adding an additional medical team, interestingly did not have a significant effect on patients’ average time in system or the number of patients waiting for a bed even though a decrease occurred in average team census, team utilization, and patient dispersion.

Finally, Future State 4, increasing admissions while also adding a nursing unit and a hospitalist team, resulted in an increase in admission volume while maintaining similar utilization rates for teams and NU. Patients spent about 2.48 hours less in the system, while only 9.94 patients were noted to be waiting for a bed as compared with 11.21 patients in the present state model. The total daily general medicine patient census was noted to be 137.19. Average team census and average team utilization were noted to be similar to those of the present state model, while admissions were up by approximately 1,080 per year. Both patient dispersion and rounding were slightly worsened.

Sensitivity Analysis

Overall, average time in system was most affected by the number of patient arrivals. This became particularly significant as the volume of patient arrivals approached and exceeded the capacity of the rounding teams. Adding a nursing unit had more impact on decreasing average time in the system than adding a medical team or aligning teams with NUs under the conditions defined by the model. However, under different conditions, such as increasing admission volume, the relative benefit of different approaches may vary.

Given that hospitals are large, complex systems,² the impact of system-level changes can have unpredictable and potentially deleterious effects. Simulation provides a technique for modeling the impact of changes to understand the ramifications of these interventions more thoroughly.³ In this study, we describe the process of building a simulation model for the admission and discharge of patients from general medicine services in a tertiary care hospital, internally validating this model, and examining the outcomes from several potential changes to the system.

The outcomes for these what-if scenarios provided some important insights about the secondary effect of system changes and the need for multiple, simultaneous interventions. Given that hospitals often function at near capacity, adding a hospitalist team or nursing unit might be seen as a reasonable strategy to improve the system metrics, number of patient discharges, or average LOS. On the basis of our analysis, adding a nursing unit would have more benefit than adding a hospitalist team. Leaders who want to increase capacity may need to consider both adding a hospitalist team and a nursing unit, and model the impact of each choice as described with a simulation.

Additionally, assigning patients to medical teams aligned with NUs seems theoretically appealing to improve interprofessional communication and decrease the time spent in transit between patients by physicians. While our findings supported a decrease in rounding time and patient dispersion, the teams not aligned with a nursing unit (ie, the hospitalists) exceeded 80% utilization, the threshold at which efficiency is known to decrease.²⁴ Potentially, benefits resulting from teams being aligned with NUs were offset by decrements in performance of the teams not aligned with NU. If medical teams and NUs become aligned, then a higher number of teams may be necessary to maintain patient throughput.

Simulation models identify these unexpected consequences prior to investing resources in a significant change; however, modeling is not simple. Simulation models depend on the characteristics of the model and the quality of the input data. For example, we used an expert approach to map physician workflow as an underpinning of the model, but we may have missed an important variation in physician workflow. Understanding this variation could strengthen the model and provide some testable variables for future study. Likewise, understanding nursing workflow and how variation in physician workflow shapes nursing workflow, and vice versa, is worth exploring.

Other data could also be added to, and help interpret, the outputs of this model. For example, the impact of various levels of team and unit utilization on diversion time for the hospital ED may help determine whether adding team capacity or unit capacity is more beneficial for the system. Likewise, aligning medical teams with NUs seems to hinder patient throughput on this analysis, but benefits in patient satisfaction or decreased readmissions might improve reimbursement and outweigh the revenue lost from throughput. Underpinning each of these types of decisions is a need to model the system well and thoughtfully choose the inputs, processes, and outputs. Pursuing a new strategic decision usually involves cost; simulation modeling provides data to help leaders weigh the benefits in terms of the needed investment.

The major limitations of the study stem from these choices. Our study focused on matching capacity and demand while limiting other changes in the system, such as changes in nursing unit LOS. Future work to quantify the relationship of other variables on parameters, such as the impact of decreased team dispersion on LOS, early discharges, and decreasing care variation, would make future models more robust. This model does not consider other strategies to improve patient flow, such as shaping demand, adaptive team assignment algorithms, or creating surge capacity. We also used only hospitalist time and motion data in our model; housestaff workflow is likely different. In addition, we modeled all patients as having a general level of nursing care and did not account for admissions or transfers to intensive care units or other services. These parameters could be added in future iterations. Finally, the biggest limitation in any simulation is the underlying assumptions made to construct the model. While we validated the model retrospectively, prospective validation and refinement should also be performed with attention to how the model functions under extreme conditions, such as a very high patient load.

Major system changes are expensive and must be made carefully. Systems engineering techniques, such as DES, provide techniques to estimate the impact of changes on pertinent care delivery variables. Results from this study underscore the complexity of patient care delivery and how simulation models can integrate multiple system components to provide a data-driven approach to inform decision making in a complex system.

Acknowledgments

The simulation software used in this study was awarded as an educational software grant from SIMIO®. We would like to acknowledge support from the Department of Internal Medicine at Virginia Commonwealth University for this project and thank Lena Rivera for her assistance with the manuscript preparation.

Dislosures

Dr. Heim recived a consulting fee for programming guidance from Virginia Commonwealth University. All other authors have nothing to disclose.

ABSTRACT

Systemic corticosteroids are used to treat a number of medical conditions; however, they are associated with numerous adverse effects. The impact of preoperative chronic corticosteroid use on postoperative outcomes following total hip arthroplasty (THA) is unclear. The purpose of this study was to assess the independent effect of chronic systemic preoperative steroid use on short-term perioperative complications and readmissions after THA.

All patients undergoing primary THA in the American College of Surgeons National Surgical Quality Improvement Program registry from 2005 to -–2015 were identified. Patients were considered chronic steroid users if they used any dosage of oral or parenteral steroids for >10 of the preceding 30 days before THA. Two equally sized propensity-matched groups based on preoperative steroid use were generated to account for differences in operative and baseline characteristics between the groups. Thirty-day complications and hospital readmissions rates were compared using bivariate analysis.

Of 101,532 THA patients who underwent primary THA, 3714 (3.7%) were identified as chronic corticosteroid users. Comparison of propensity-matched cohorts identified an increased rate of any complication (odds ratio [OR] 1.30, P = .003), sepsis (OR 2.07, P = .022), urinary tract infection (OR 1.61, P = .020), superficial surgical site infection (OR 1.73, P = .038), and hospital readmission (OR 1.50, P < .001) in patients who used systemic steroids preoperatively. Readmissions in preoperative steroid users were most commonly for infectious reasons.

Patients prescribed chronic corticosteroids are at a significantly increased risk of both 30-day periopative complications and hospital readmissions. This finding has important implications for pre- and postoperative patient counseling as well as preoperative risk stratification.

Continue to: Corticosteroids are powerful...

Corticosteroids are powerful anti-inflammatory steroid hormones that have many indications in the treatment of medical diseases, including advanced or poorly controlled asthma, chronic obstructive pulmonary disease (COPD), inflammatory bowel disease, allergic conditions, among other indications.^1-4 In orthopedics and rheumatology, systemic steroids are, at times, used in patients with rheumatoid arthritis, systemic lupus erythematosus, and vasculitides.^5-7 Overman and colleagues,⁸ using data from the National Health and Nutrition Examination Survey between 1999 and 2008 identified both a 1.2% prevalence of chronic corticosteroid usage in the United States across all age groups and a positive correlation between steroid use prevalence and increasing age. In that study, nearly two-thirds of survey respondents reported using corticosteroids chronically for >90 days. Another observational study in the United Kingdom found that long-term steroid prescriptions increased between 1989 to 2008 and that 13.6% of patients with rheumatoid arthritis and 66.5% of patients with polymyalgia rheumatica or giant cell arteritis used long-term steroids.⁹

Enterally- or parenterally-administered corticosteroids have numerous systemic effects that are of particular relevance to orthopedic surgeons. Corticosteroids induce osteoporosis by preferentially inducing osteoclastic activity while inhibiting the differentiation of osteoblasts, ultimately leading to decreased bone quality and mass.¹⁰ As a consequence, patients who have previously used corticosteroids are more than twice as likely to have a hip fracture.¹¹ Steroids also increase the risk of both osteonecrosis and myopathy, among other musculoskeletal effects.¹² In addition to orthopedic complications, steroids have broad inhibitory effects on both acquired and innate immunity, which significantly increases the risk of infections.¹³ This increased risk of infection is dose-dependent¹⁴ and synergistic with other immunosuppressive drugs.¹⁵

Patients with hip pain may receive localized corticosteroid hip joint injections during the nonoperative management of various hip pathologies, including arthritis, bursitis, and labral tears.^16,17 Outcomes of patients who received intra-articular corticosteroid injections before total hip arthroplasty (THA) were evaluated in a systematic review of 9 studies by Pereira and colleagues.¹⁷ These authors found that the infection rate (both superficial and deep surgical site infections [SSI]) after THA in patients who received local steroid injection into the hip before surgery was between 0% and 30%.¹⁷ However, similar studies assessing the impact that systemic steroids have on outcomes after THA are lacking. Patients who undergo THA for conditions associated with higher lifetime steroid usage have worse outcomes than those who do not. For instance, in patients undergoing THA for rheumatoid arthritis, the rates of both postoperative periprosthetic joint infection and hip dislocation are higher, when compared with osteoarthritis.^18,19 However, it is unclear how much of this difference in outcomes is due to the underlying disease, adverse effects of steroids, or both. Given the high prevalence of chronic systemic steroid use, it is essential to elucidate more clearly the impact that these medications have on perioperative outcomes after THA.  

Therefore, the purpose of this study was to characterize short-term perioperative outcomes, including complication and readmission rates in patients undergoing THA while taking chronic preoperative corticosteroids. We also sought to identify the most common reasons for hospital readmission in patients who did and did not use long-term steroids.

MATERIALS AND METHODS

STUDY DESIGN AND SETTING

This investigation was a retrospective cohort study that utilized the American College of Surgeons National Surgical Quality Improvement Program (ACS-NSQIP) registry.²⁰ The ACS-NSQIP is a prospectively collected, multi-institutional database that collects demographical information, operative variables, and both postoperative complications and hospital readmission data. Data is collected for up to 30 days after the index procedure, and patients are contacted by telephone if they are discharged before 30 days. Patient data is entered by specially trained surgical clinical reviewers and is routinely audited by the ACS-NSQIP, leading to more accurate data when compared with administrative research databases.^21,22 The ACS-NSQIP has been used in orthopedic surgery outcomes-based studies.^23-25

Continue to: All patients undergoing...

All patients undergoing THA between 2005 and 2015 were identified in the registry using primary Current Procedural Terminology code 27130. Patients were split into 2 groups based on whether or not they chronically used corticosteroids preoperatively for a medical condition. A patient was considered a chronic corticosteroid user if he/she used oral or parenteral corticosteroids within 30 days before the index procedure for >10 of the preceding 30 days. Those who received a 1-time steroid pulse or those who used topical or inhaled steroids were not considered as steroid users in this study.

BASELINE CHARACTERISTICS AND PERIOPERATIVE OUTCOMES

Baseline patient and operative characteristics, including patient age, gender, body mass index (BMI), functional status, American Society of Anesthesiologists (ASA) class, anesthesia type, operative duration, and medical comorbidities including hypertension, COPD, diabetes mellitus, and smoking history, were compared between both groups. Perioperative outcomes that were assessed in this study include death, renal, respiratory, and cardiac complications, deep vein thrombosis or pulmonary embolism, stroke, sepsis, return to the operating room, urinary tract infection (UTI), wound dehiscence, superficial and deep SSI, need for a blood transfusion within 72 hours of index surgical procedure, and hospital readmissions. Renal complications were defined as acute or progressive renal insufficiency; respiratory complications were defined as failure to wean from the ventilator, need for intubation after the index procedure, and the occurrence of pneumonia; and cardiac complications were defined as myocardial infarction or cardiac arrest requiring cardiopulmonary resuscitation. Patients were excluded if they had missing baseline or operative characteristic data, an unclean wound classification at the time of admission, or if their THA was considered emergent.

STATISTICAL ANALYSIS

A propensity score-matched comparison was performed to adjust for differences in baseline and operative characteristics between the 2 cohorts in this study. In the current study, the propensity score was defined as the conditional probability that a patient chronically used preoperative corticosteroids for a medical condition, as a function of age, BMI, gender, ASA class, functional status, medical comorbidities, anesthesia type, and operative duration. A 1:1 matching with tight calipers (0.0001), and nearest-neighbor matching was used to generate 2 equally-sized, propensity-matched cohorts based on steroid status.²⁶ Nearest-neighbor matching identifies patients in both cohorts with the closest propensity scores for inclusion in propensity-matched cohorts. This matching is continued until 1 group runs out of patients to match. Baseline patient and operative characteristics for the unadjusted and propensity-matched groups were compared using Pearson’s χ² analysis. Outcomes after THA by steroid status were also compared in both unadjusted and propensity-matched groups. Finally, all patients who were readmitted were identified, and the reason for readmission was determined using the International Classification of Disease Ninth (ICD-9) and Tenth (ICD-10) edition codes. Patients were classified as having an infectious readmission only if the ICD code clearly stated an infectious etiology. For instance, a patient with an intestinal infection due to Clostridium difficile (ICD-9 008.45) was counted as a gastrointestinal infection, whereas diarrhea without a distinctly specified etiology (ICD-9 787.91, ICD-10 R19.7) was counted as a gastrointestinal medical complication. Readmission data was only available in ACS-NSQIP from 2011 to 2015, constituting 92.5% of all patients included in this study. We used SPSS version 23 (IBM Corporation) for all statistical analyses, and defined a significant P value as <.05.

RESULTS

BASELINE PATIENTS AND OPERATIVE CHARACTERISTICS

In total, we identified 101,532 patients who underwent THA (Table 1). O these, 3714 (3.7%) chronically used corticosteroids preoperatively, whereas 97,818 (96.3%) did not. 

When the unadjusted cohorts were compared, patients using corticosteroids were more likely to be female, less likely to obese, more likely to have hypertension, diabetes mellitus, COPD, higher ASA class, undergone THA with general anesthesia, and have a dependent functional status (P < .001 for all comparisons). After propensity matching, 2 equally sized cohorts of 3618 patients each were generated based on steroid status and no differences in baseline and operative characteristics were identified between the 2 groups.

Continue to: CLINICAL OUTCOMES BY STEROID STATUS

CLINCIAL OUTCOMES BY STEROID STATUS

A comparison of unadjusted cohorts showed that patients who used preoperative steroids had an increased rate of any complication (7.89%) when compared with those who did not (4.87%) (Table 2). 

Similarly, those who used corticosteroids preoperatively had an increased rate of renal complications, respiratory complications, return to the operating room, sepsis, UTI, superficial and deep SSI, and perioperative blood transfusions. They also were more likely to have a 30-day hospital readmission (P < .05 for all comparisons).

When propensity-matched cohorts were compared, patients who used steroids preoperatively were found to have higher rates of any complication (odds Ratio [OR] 1.30, P = .003), sepsis (OR 2.07, P = .022), UTI (OR 1.61, P = .020), superficial SSI (OR 1.73, P = .038), and hospital readmission (OR 1.50, P < .001; Table 3).

REASONS FOR HOSPITAL READMISSION

In total, 3397 patients were readmitted to the hospital within thirty days. Of these, 226 used steroids preoperatively, and 3171 did not (Table 4). 

The most common reason for hospital readmission in patients who used preoperative corticosteroids was infectious complications (72 patients, 31.9% of all readmitted patients in this cohort), followed by medical complications (59 patients, 26.1%), and hip-related complications (48 patients, 21.2%). In those who did not use steroids preoperatively, the most common reason for hospital readmission was medical complications (932 patients, 29.4% of all readmitted patients in this cohort), followed by infectious complications (792 patients, 25.0%), and hip-related complications (763 patients, 24.1%).

Continue to: DISCUSSION

Nearly 3% of individuals >80 years in the US population chronically use corticosteroids for a medical condition,⁸ and this rate is likely higher in specific subsets of patients, such as those with rheumatoid arthritis.⁹ While some studies have assessed the impact of intra-articular corticosteroid hip injections on perioperative outcomes in THA,¹⁷ similar studies assessing systemic corticosteroid usage are lacking. The purpose of this study was to characterize short-term perioperative outcomes in patients undergoing THA who chronically use systemic steroids when compared with those who do not. We found that the prevalence of preoperative chronic steroid use in this cohort of THA patients was 3.7%. We also identified increased rates of infectious complications, including sepsis, UTI, and superficial SSI, in patients who used preoperative corticosteroids. Furthermore, we found an increased rate of hospital readmissions in corticosteroid users and identified the most common reason for hospital readmission as infectious complications in this cohort.

The primary finding of this study was an increase in postoperative infections in patients who use preoperative steroids chronically for medical conditions. Immunosuppression has previously been identified as a risk factor for developing periprosthetic joint infections. Tannenbaum and colleagues²⁷ performed a retrospective study of 19 patients who underwent either a kidney or liver transplant and were maintained on an induction regimen of either prednisone and azathioprine or cyclosporine. These 19 patients also underwent either a THA or total knee arthroplasty, and 5 of these patients (26.3%) developed a periprosthetic joint infection after an average of 3.4 years following the arthroplasty procedure. In another study of 37 renal transplant and dialysis patients who underwent a total of 45 THA procedures, there were 3 instances of superficial SSI and 2 instances of deep SSI.²⁸ However, reported infection rates in transplant patients undergoing THA vary significantly, and studies have been unable to assess the true impact that chronic immunosuppression has on perioperative infection rates.²⁹ In this study, patients who used preoperative corticosteroids chronically were at increased risk of perioperative infections, including sepsis, UTI, and superficial SSI.

Deep vein thrombosis is another postoperative complication that has been associated with chronic steroid use.³⁰ In a case-control study of 38,765 patients who developed a venous thromboembolism and 387,650 control patients who did not, Johannesdottir and colleagues³⁰ found an increased thromboembolic risk in current users of systemic glucocorticoids, but not former users, as well as an increased risk as the dose of glucocorticoids increased. We were not able to identify a similar increase in DVT/PE in chronic corticosteroid users, perhaps due to our sample size, or because we could not do subgroup analyses based on the type or dosage of steroid that a patient was taking. Future studies that identify the highest risk patients among those using systemic corticosteroids are important because parenteral corticosteroids are being increasingly used in THA to alleviate postoperative pain as an opioid-sparing measure.^31,32

Finally, we also found that patients who use chronic, systemic corticosteroids are at an increased risk for hospital readmission, when compared with those patients who are not using steroids and are most likely to be readmitted for an infectious complication. Schairer and colleagues³³ assessed readmission rates after THA and found 30- and 90-day readmission rate of 4% and 7%, respectively. These authors also found that medical complications accounted for approximately 25% of readmissions, and hip-related complications (eg, dislocation, SSI) accounted for >50%. In our study, we found a 30-day readmission rate in non-steroid users of 3.53% and a rate of 6.52% in chronic steroid users. More than 30% of patients using a steroid were readmitted for infectious complications. As THA is becoming increasingly reimbursed under a bundled payments model by Medicare and Medicaid,^34-36 reducing short-term readmissions is imperative. Therefore, discharge counseling that emphasizes how to recognize both the signs and symptoms of infection as well as how to prevent infections, such as reducing SSIs through appropriate wound care, may be warranted in higher risk chronic steroid users.

This study has a number of limitations that are inherent to ACS-NSQIP. First, we lacked specific information on a patient’s steroid history, including which corticosteroid they were using, dosage, frequency, and the indication for corticosteroid therapy. Therefore, we were unable to establish a dose-dependent relationship between steroid exposure and postoperative complications after THA. Second, we were able to assess only 30-day rates of complications and readmissions, and therefore, we were unable to identify intermediate- and long-term effects of systemic corticosteroid use on THA. Finally, we could not determine orthopedic- or hip-specific postoperative outcomes, such as functional scores and range of motion.

Continue to: CONCLUSION

In conclusion, this study quantified the increased risk for perioperative complications and hospital readmissions in patients who chronically use corticosteroids and are undergoing THA, when compared with those who do not use corticosteroids. These results suggest that patients who are on long-term steroids are at an increased risk for complications, primarily infectious complications. This finding has important implications for patient counseling, preoperative risk stratification, and suggests that higher risk patients, such as chronic steroid users, may benefit from improved discharge care to decrease complication rates.

ABSTRACT

Systemic corticosteroids are used to treat a number of medical conditions; however, they are associated with numerous adverse effects. The impact of preoperative chronic corticosteroid use on postoperative outcomes following total hip arthroplasty (THA) is unclear. The purpose of this study was to assess the independent effect of chronic systemic preoperative steroid use on short-term perioperative complications and readmissions after THA.

All patients undergoing primary THA in the American College of Surgeons National Surgical Quality Improvement Program registry from 2005 to -–2015 were identified. Patients were considered chronic steroid users if they used any dosage of oral or parenteral steroids for >10 of the preceding 30 days before THA. Two equally sized propensity-matched groups based on preoperative steroid use were generated to account for differences in operative and baseline characteristics between the groups. Thirty-day complications and hospital readmissions rates were compared using bivariate analysis.

Of 101,532 THA patients who underwent primary THA, 3714 (3.7%) were identified as chronic corticosteroid users. Comparison of propensity-matched cohorts identified an increased rate of any complication (odds ratio [OR] 1.30, P = .003), sepsis (OR 2.07, P = .022), urinary tract infection (OR 1.61, P = .020), superficial surgical site infection (OR 1.73, P = .038), and hospital readmission (OR 1.50, P < .001) in patients who used systemic steroids preoperatively. Readmissions in preoperative steroid users were most commonly for infectious reasons.

Patients prescribed chronic corticosteroids are at a significantly increased risk of both 30-day periopative complications and hospital readmissions. This finding has important implications for pre- and postoperative patient counseling as well as preoperative risk stratification.

Continue to: Corticosteroids are powerful...

Corticosteroids are powerful anti-inflammatory steroid hormones that have many indications in the treatment of medical diseases, including advanced or poorly controlled asthma, chronic obstructive pulmonary disease (COPD), inflammatory bowel disease, allergic conditions, among other indications.^1-4 In orthopedics and rheumatology, systemic steroids are, at times, used in patients with rheumatoid arthritis, systemic lupus erythematosus, and vasculitides.^5-7 Overman and colleagues,⁸ using data from the National Health and Nutrition Examination Survey between 1999 and 2008 identified both a 1.2% prevalence of chronic corticosteroid usage in the United States across all age groups and a positive correlation between steroid use prevalence and increasing age. In that study, nearly two-thirds of survey respondents reported using corticosteroids chronically for >90 days. Another observational study in the United Kingdom found that long-term steroid prescriptions increased between 1989 to 2008 and that 13.6% of patients with rheumatoid arthritis and 66.5% of patients with polymyalgia rheumatica or giant cell arteritis used long-term steroids.⁹

Enterally- or parenterally-administered corticosteroids have numerous systemic effects that are of particular relevance to orthopedic surgeons. Corticosteroids induce osteoporosis by preferentially inducing osteoclastic activity while inhibiting the differentiation of osteoblasts, ultimately leading to decreased bone quality and mass.¹⁰ As a consequence, patients who have previously used corticosteroids are more than twice as likely to have a hip fracture.¹¹ Steroids also increase the risk of both osteonecrosis and myopathy, among other musculoskeletal effects.¹² In addition to orthopedic complications, steroids have broad inhibitory effects on both acquired and innate immunity, which significantly increases the risk of infections.¹³ This increased risk of infection is dose-dependent¹⁴ and synergistic with other immunosuppressive drugs.¹⁵

Patients with hip pain may receive localized corticosteroid hip joint injections during the nonoperative management of various hip pathologies, including arthritis, bursitis, and labral tears.^16,17 Outcomes of patients who received intra-articular corticosteroid injections before total hip arthroplasty (THA) were evaluated in a systematic review of 9 studies by Pereira and colleagues.¹⁷ These authors found that the infection rate (both superficial and deep surgical site infections [SSI]) after THA in patients who received local steroid injection into the hip before surgery was between 0% and 30%.¹⁷ However, similar studies assessing the impact that systemic steroids have on outcomes after THA are lacking. Patients who undergo THA for conditions associated with higher lifetime steroid usage have worse outcomes than those who do not. For instance, in patients undergoing THA for rheumatoid arthritis, the rates of both postoperative periprosthetic joint infection and hip dislocation are higher, when compared with osteoarthritis.^18,19 However, it is unclear how much of this difference in outcomes is due to the underlying disease, adverse effects of steroids, or both. Given the high prevalence of chronic systemic steroid use, it is essential to elucidate more clearly the impact that these medications have on perioperative outcomes after THA.  

Therefore, the purpose of this study was to characterize short-term perioperative outcomes, including complication and readmission rates in patients undergoing THA while taking chronic preoperative corticosteroids. We also sought to identify the most common reasons for hospital readmission in patients who did and did not use long-term steroids.

MATERIALS AND METHODS

STUDY DESIGN AND SETTING

This investigation was a retrospective cohort study that utilized the American College of Surgeons National Surgical Quality Improvement Program (ACS-NSQIP) registry.²⁰ The ACS-NSQIP is a prospectively collected, multi-institutional database that collects demographical information, operative variables, and both postoperative complications and hospital readmission data. Data is collected for up to 30 days after the index procedure, and patients are contacted by telephone if they are discharged before 30 days. Patient data is entered by specially trained surgical clinical reviewers and is routinely audited by the ACS-NSQIP, leading to more accurate data when compared with administrative research databases.^21,22 The ACS-NSQIP has been used in orthopedic surgery outcomes-based studies.^23-25

Continue to: All patients undergoing...

All patients undergoing THA between 2005 and 2015 were identified in the registry using primary Current Procedural Terminology code 27130. Patients were split into 2 groups based on whether or not they chronically used corticosteroids preoperatively for a medical condition. A patient was considered a chronic corticosteroid user if he/she used oral or parenteral corticosteroids within 30 days before the index procedure for >10 of the preceding 30 days. Those who received a 1-time steroid pulse or those who used topical or inhaled steroids were not considered as steroid users in this study.

BASELINE CHARACTERISTICS AND PERIOPERATIVE OUTCOMES

Baseline patient and operative characteristics, including patient age, gender, body mass index (BMI), functional status, American Society of Anesthesiologists (ASA) class, anesthesia type, operative duration, and medical comorbidities including hypertension, COPD, diabetes mellitus, and smoking history, were compared between both groups. Perioperative outcomes that were assessed in this study include death, renal, respiratory, and cardiac complications, deep vein thrombosis or pulmonary embolism, stroke, sepsis, return to the operating room, urinary tract infection (UTI), wound dehiscence, superficial and deep SSI, need for a blood transfusion within 72 hours of index surgical procedure, and hospital readmissions. Renal complications were defined as acute or progressive renal insufficiency; respiratory complications were defined as failure to wean from the ventilator, need for intubation after the index procedure, and the occurrence of pneumonia; and cardiac complications were defined as myocardial infarction or cardiac arrest requiring cardiopulmonary resuscitation. Patients were excluded if they had missing baseline or operative characteristic data, an unclean wound classification at the time of admission, or if their THA was considered emergent.

STATISTICAL ANALYSIS

A propensity score-matched comparison was performed to adjust for differences in baseline and operative characteristics between the 2 cohorts in this study. In the current study, the propensity score was defined as the conditional probability that a patient chronically used preoperative corticosteroids for a medical condition, as a function of age, BMI, gender, ASA class, functional status, medical comorbidities, anesthesia type, and operative duration. A 1:1 matching with tight calipers (0.0001), and nearest-neighbor matching was used to generate 2 equally-sized, propensity-matched cohorts based on steroid status.²⁶ Nearest-neighbor matching identifies patients in both cohorts with the closest propensity scores for inclusion in propensity-matched cohorts. This matching is continued until 1 group runs out of patients to match. Baseline patient and operative characteristics for the unadjusted and propensity-matched groups were compared using Pearson’s χ² analysis. Outcomes after THA by steroid status were also compared in both unadjusted and propensity-matched groups. Finally, all patients who were readmitted were identified, and the reason for readmission was determined using the International Classification of Disease Ninth (ICD-9) and Tenth (ICD-10) edition codes. Patients were classified as having an infectious readmission only if the ICD code clearly stated an infectious etiology. For instance, a patient with an intestinal infection due to Clostridium difficile (ICD-9 008.45) was counted as a gastrointestinal infection, whereas diarrhea without a distinctly specified etiology (ICD-9 787.91, ICD-10 R19.7) was counted as a gastrointestinal medical complication. Readmission data was only available in ACS-NSQIP from 2011 to 2015, constituting 92.5% of all patients included in this study. We used SPSS version 23 (IBM Corporation) for all statistical analyses, and defined a significant P value as <.05.

RESULTS

BASELINE PATIENTS AND OPERATIVE CHARACTERISTICS

In total, we identified 101,532 patients who underwent THA (Table 1). O these, 3714 (3.7%) chronically used corticosteroids preoperatively, whereas 97,818 (96.3%) did not. 

When the unadjusted cohorts were compared, patients using corticosteroids were more likely to be female, less likely to obese, more likely to have hypertension, diabetes mellitus, COPD, higher ASA class, undergone THA with general anesthesia, and have a dependent functional status (P < .001 for all comparisons). After propensity matching, 2 equally sized cohorts of 3618 patients each were generated based on steroid status and no differences in baseline and operative characteristics were identified between the 2 groups.

Continue to: CLINICAL OUTCOMES BY STEROID STATUS

CLINCIAL OUTCOMES BY STEROID STATUS

A comparison of unadjusted cohorts showed that patients who used preoperative steroids had an increased rate of any complication (7.89%) when compared with those who did not (4.87%) (Table 2). 

Similarly, those who used corticosteroids preoperatively had an increased rate of renal complications, respiratory complications, return to the operating room, sepsis, UTI, superficial and deep SSI, and perioperative blood transfusions. They also were more likely to have a 30-day hospital readmission (P < .05 for all comparisons).

When propensity-matched cohorts were compared, patients who used steroids preoperatively were found to have higher rates of any complication (odds Ratio [OR] 1.30, P = .003), sepsis (OR 2.07, P = .022), UTI (OR 1.61, P = .020), superficial SSI (OR 1.73, P = .038), and hospital readmission (OR 1.50, P < .001; Table 3).

REASONS FOR HOSPITAL READMISSION

In total, 3397 patients were readmitted to the hospital within thirty days. Of these, 226 used steroids preoperatively, and 3171 did not (Table 4). 

The most common reason for hospital readmission in patients who used preoperative corticosteroids was infectious complications (72 patients, 31.9% of all readmitted patients in this cohort), followed by medical complications (59 patients, 26.1%), and hip-related complications (48 patients, 21.2%). In those who did not use steroids preoperatively, the most common reason for hospital readmission was medical complications (932 patients, 29.4% of all readmitted patients in this cohort), followed by infectious complications (792 patients, 25.0%), and hip-related complications (763 patients, 24.1%).

Continue to: DISCUSSION

Nearly 3% of individuals >80 years in the US population chronically use corticosteroids for a medical condition,⁸ and this rate is likely higher in specific subsets of patients, such as those with rheumatoid arthritis.⁹ While some studies have assessed the impact of intra-articular corticosteroid hip injections on perioperative outcomes in THA,¹⁷ similar studies assessing systemic corticosteroid usage are lacking. The purpose of this study was to characterize short-term perioperative outcomes in patients undergoing THA who chronically use systemic steroids when compared with those who do not. We found that the prevalence of preoperative chronic steroid use in this cohort of THA patients was 3.7%. We also identified increased rates of infectious complications, including sepsis, UTI, and superficial SSI, in patients who used preoperative corticosteroids. Furthermore, we found an increased rate of hospital readmissions in corticosteroid users and identified the most common reason for hospital readmission as infectious complications in this cohort.

The primary finding of this study was an increase in postoperative infections in patients who use preoperative steroids chronically for medical conditions. Immunosuppression has previously been identified as a risk factor for developing periprosthetic joint infections. Tannenbaum and colleagues²⁷ performed a retrospective study of 19 patients who underwent either a kidney or liver transplant and were maintained on an induction regimen of either prednisone and azathioprine or cyclosporine. These 19 patients also underwent either a THA or total knee arthroplasty, and 5 of these patients (26.3%) developed a periprosthetic joint infection after an average of 3.4 years following the arthroplasty procedure. In another study of 37 renal transplant and dialysis patients who underwent a total of 45 THA procedures, there were 3 instances of superficial SSI and 2 instances of deep SSI.²⁸ However, reported infection rates in transplant patients undergoing THA vary significantly, and studies have been unable to assess the true impact that chronic immunosuppression has on perioperative infection rates.²⁹ In this study, patients who used preoperative corticosteroids chronically were at increased risk of perioperative infections, including sepsis, UTI, and superficial SSI.

Deep vein thrombosis is another postoperative complication that has been associated with chronic steroid use.³⁰ In a case-control study of 38,765 patients who developed a venous thromboembolism and 387,650 control patients who did not, Johannesdottir and colleagues³⁰ found an increased thromboembolic risk in current users of systemic glucocorticoids, but not former users, as well as an increased risk as the dose of glucocorticoids increased. We were not able to identify a similar increase in DVT/PE in chronic corticosteroid users, perhaps due to our sample size, or because we could not do subgroup analyses based on the type or dosage of steroid that a patient was taking. Future studies that identify the highest risk patients among those using systemic corticosteroids are important because parenteral corticosteroids are being increasingly used in THA to alleviate postoperative pain as an opioid-sparing measure.^31,32

Finally, we also found that patients who use chronic, systemic corticosteroids are at an increased risk for hospital readmission, when compared with those patients who are not using steroids and are most likely to be readmitted for an infectious complication. Schairer and colleagues³³ assessed readmission rates after THA and found 30- and 90-day readmission rate of 4% and 7%, respectively. These authors also found that medical complications accounted for approximately 25% of readmissions, and hip-related complications (eg, dislocation, SSI) accounted for >50%. In our study, we found a 30-day readmission rate in non-steroid users of 3.53% and a rate of 6.52% in chronic steroid users. More than 30% of patients using a steroid were readmitted for infectious complications. As THA is becoming increasingly reimbursed under a bundled payments model by Medicare and Medicaid,^34-36 reducing short-term readmissions is imperative. Therefore, discharge counseling that emphasizes how to recognize both the signs and symptoms of infection as well as how to prevent infections, such as reducing SSIs through appropriate wound care, may be warranted in higher risk chronic steroid users.

This study has a number of limitations that are inherent to ACS-NSQIP. First, we lacked specific information on a patient’s steroid history, including which corticosteroid they were using, dosage, frequency, and the indication for corticosteroid therapy. Therefore, we were unable to establish a dose-dependent relationship between steroid exposure and postoperative complications after THA. Second, we were able to assess only 30-day rates of complications and readmissions, and therefore, we were unable to identify intermediate- and long-term effects of systemic corticosteroid use on THA. Finally, we could not determine orthopedic- or hip-specific postoperative outcomes, such as functional scores and range of motion.

Continue to: CONCLUSION

In conclusion, this study quantified the increased risk for perioperative complications and hospital readmissions in patients who chronically use corticosteroids and are undergoing THA, when compared with those who do not use corticosteroids. These results suggest that patients who are on long-term steroids are at an increased risk for complications, primarily infectious complications. This finding has important implications for patient counseling, preoperative risk stratification, and suggests that higher risk patients, such as chronic steroid users, may benefit from improved discharge care to decrease complication rates.

ABSTRACT

Systemic corticosteroids are used to treat a number of medical conditions; however, they are associated with numerous adverse effects. The impact of preoperative chronic corticosteroid use on postoperative outcomes following total hip arthroplasty (THA) is unclear. The purpose of this study was to assess the independent effect of chronic systemic preoperative steroid use on short-term perioperative complications and readmissions after THA.

All patients undergoing primary THA in the American College of Surgeons National Surgical Quality Improvement Program registry from 2005 to -–2015 were identified. Patients were considered chronic steroid users if they used any dosage of oral or parenteral steroids for >10 of the preceding 30 days before THA. Two equally sized propensity-matched groups based on preoperative steroid use were generated to account for differences in operative and baseline characteristics between the groups. Thirty-day complications and hospital readmissions rates were compared using bivariate analysis.

Of 101,532 THA patients who underwent primary THA, 3714 (3.7%) were identified as chronic corticosteroid users. Comparison of propensity-matched cohorts identified an increased rate of any complication (odds ratio [OR] 1.30, P = .003), sepsis (OR 2.07, P = .022), urinary tract infection (OR 1.61, P = .020), superficial surgical site infection (OR 1.73, P = .038), and hospital readmission (OR 1.50, P < .001) in patients who used systemic steroids preoperatively. Readmissions in preoperative steroid users were most commonly for infectious reasons.

Patients prescribed chronic corticosteroids are at a significantly increased risk of both 30-day periopative complications and hospital readmissions. This finding has important implications for pre- and postoperative patient counseling as well as preoperative risk stratification.

Continue to: Corticosteroids are powerful...

Corticosteroids are powerful anti-inflammatory steroid hormones that have many indications in the treatment of medical diseases, including advanced or poorly controlled asthma, chronic obstructive pulmonary disease (COPD), inflammatory bowel disease, allergic conditions, among other indications.^1-4 In orthopedics and rheumatology, systemic steroids are, at times, used in patients with rheumatoid arthritis, systemic lupus erythematosus, and vasculitides.^5-7 Overman and colleagues,⁸ using data from the National Health and Nutrition Examination Survey between 1999 and 2008 identified both a 1.2% prevalence of chronic corticosteroid usage in the United States across all age groups and a positive correlation between steroid use prevalence and increasing age. In that study, nearly two-thirds of survey respondents reported using corticosteroids chronically for >90 days. Another observational study in the United Kingdom found that long-term steroid prescriptions increased between 1989 to 2008 and that 13.6% of patients with rheumatoid arthritis and 66.5% of patients with polymyalgia rheumatica or giant cell arteritis used long-term steroids.⁹

Enterally- or parenterally-administered corticosteroids have numerous systemic effects that are of particular relevance to orthopedic surgeons. Corticosteroids induce osteoporosis by preferentially inducing osteoclastic activity while inhibiting the differentiation of osteoblasts, ultimately leading to decreased bone quality and mass.¹⁰ As a consequence, patients who have previously used corticosteroids are more than twice as likely to have a hip fracture.¹¹ Steroids also increase the risk of both osteonecrosis and myopathy, among other musculoskeletal effects.¹² In addition to orthopedic complications, steroids have broad inhibitory effects on both acquired and innate immunity, which significantly increases the risk of infections.¹³ This increased risk of infection is dose-dependent¹⁴ and synergistic with other immunosuppressive drugs.¹⁵

Patients with hip pain may receive localized corticosteroid hip joint injections during the nonoperative management of various hip pathologies, including arthritis, bursitis, and labral tears.^16,17 Outcomes of patients who received intra-articular corticosteroid injections before total hip arthroplasty (THA) were evaluated in a systematic review of 9 studies by Pereira and colleagues.¹⁷ These authors found that the infection rate (both superficial and deep surgical site infections [SSI]) after THA in patients who received local steroid injection into the hip before surgery was between 0% and 30%.¹⁷ However, similar studies assessing the impact that systemic steroids have on outcomes after THA are lacking. Patients who undergo THA for conditions associated with higher lifetime steroid usage have worse outcomes than those who do not. For instance, in patients undergoing THA for rheumatoid arthritis, the rates of both postoperative periprosthetic joint infection and hip dislocation are higher, when compared with osteoarthritis.^18,19 However, it is unclear how much of this difference in outcomes is due to the underlying disease, adverse effects of steroids, or both. Given the high prevalence of chronic systemic steroid use, it is essential to elucidate more clearly the impact that these medications have on perioperative outcomes after THA.  

Therefore, the purpose of this study was to characterize short-term perioperative outcomes, including complication and readmission rates in patients undergoing THA while taking chronic preoperative corticosteroids. We also sought to identify the most common reasons for hospital readmission in patients who did and did not use long-term steroids.

MATERIALS AND METHODS

STUDY DESIGN AND SETTING

This investigation was a retrospective cohort study that utilized the American College of Surgeons National Surgical Quality Improvement Program (ACS-NSQIP) registry.²⁰ The ACS-NSQIP is a prospectively collected, multi-institutional database that collects demographical information, operative variables, and both postoperative complications and hospital readmission data. Data is collected for up to 30 days after the index procedure, and patients are contacted by telephone if they are discharged before 30 days. Patient data is entered by specially trained surgical clinical reviewers and is routinely audited by the ACS-NSQIP, leading to more accurate data when compared with administrative research databases.^21,22 The ACS-NSQIP has been used in orthopedic surgery outcomes-based studies.^23-25

Continue to: All patients undergoing...

All patients undergoing THA between 2005 and 2015 were identified in the registry using primary Current Procedural Terminology code 27130. Patients were split into 2 groups based on whether or not they chronically used corticosteroids preoperatively for a medical condition. A patient was considered a chronic corticosteroid user if he/she used oral or parenteral corticosteroids within 30 days before the index procedure for >10 of the preceding 30 days. Those who received a 1-time steroid pulse or those who used topical or inhaled steroids were not considered as steroid users in this study.

BASELINE CHARACTERISTICS AND PERIOPERATIVE OUTCOMES

Baseline patient and operative characteristics, including patient age, gender, body mass index (BMI), functional status, American Society of Anesthesiologists (ASA) class, anesthesia type, operative duration, and medical comorbidities including hypertension, COPD, diabetes mellitus, and smoking history, were compared between both groups. Perioperative outcomes that were assessed in this study include death, renal, respiratory, and cardiac complications, deep vein thrombosis or pulmonary embolism, stroke, sepsis, return to the operating room, urinary tract infection (UTI), wound dehiscence, superficial and deep SSI, need for a blood transfusion within 72 hours of index surgical procedure, and hospital readmissions. Renal complications were defined as acute or progressive renal insufficiency; respiratory complications were defined as failure to wean from the ventilator, need for intubation after the index procedure, and the occurrence of pneumonia; and cardiac complications were defined as myocardial infarction or cardiac arrest requiring cardiopulmonary resuscitation. Patients were excluded if they had missing baseline or operative characteristic data, an unclean wound classification at the time of admission, or if their THA was considered emergent.

STATISTICAL ANALYSIS

A propensity score-matched comparison was performed to adjust for differences in baseline and operative characteristics between the 2 cohorts in this study. In the current study, the propensity score was defined as the conditional probability that a patient chronically used preoperative corticosteroids for a medical condition, as a function of age, BMI, gender, ASA class, functional status, medical comorbidities, anesthesia type, and operative duration. A 1:1 matching with tight calipers (0.0001), and nearest-neighbor matching was used to generate 2 equally-sized, propensity-matched cohorts based on steroid status.²⁶ Nearest-neighbor matching identifies patients in both cohorts with the closest propensity scores for inclusion in propensity-matched cohorts. This matching is continued until 1 group runs out of patients to match. Baseline patient and operative characteristics for the unadjusted and propensity-matched groups were compared using Pearson’s χ² analysis. Outcomes after THA by steroid status were also compared in both unadjusted and propensity-matched groups. Finally, all patients who were readmitted were identified, and the reason for readmission was determined using the International Classification of Disease Ninth (ICD-9) and Tenth (ICD-10) edition codes. Patients were classified as having an infectious readmission only if the ICD code clearly stated an infectious etiology. For instance, a patient with an intestinal infection due to Clostridium difficile (ICD-9 008.45) was counted as a gastrointestinal infection, whereas diarrhea without a distinctly specified etiology (ICD-9 787.91, ICD-10 R19.7) was counted as a gastrointestinal medical complication. Readmission data was only available in ACS-NSQIP from 2011 to 2015, constituting 92.5% of all patients included in this study. We used SPSS version 23 (IBM Corporation) for all statistical analyses, and defined a significant P value as <.05.

RESULTS

BASELINE PATIENTS AND OPERATIVE CHARACTERISTICS

In total, we identified 101,532 patients who underwent THA (Table 1). O these, 3714 (3.7%) chronically used corticosteroids preoperatively, whereas 97,818 (96.3%) did not. 

When the unadjusted cohorts were compared, patients using corticosteroids were more likely to be female, less likely to obese, more likely to have hypertension, diabetes mellitus, COPD, higher ASA class, undergone THA with general anesthesia, and have a dependent functional status (P < .001 for all comparisons). After propensity matching, 2 equally sized cohorts of 3618 patients each were generated based on steroid status and no differences in baseline and operative characteristics were identified between the 2 groups.

Continue to: CLINICAL OUTCOMES BY STEROID STATUS

CLINCIAL OUTCOMES BY STEROID STATUS

A comparison of unadjusted cohorts showed that patients who used preoperative steroids had an increased rate of any complication (7.89%) when compared with those who did not (4.87%) (Table 2). 

Similarly, those who used corticosteroids preoperatively had an increased rate of renal complications, respiratory complications, return to the operating room, sepsis, UTI, superficial and deep SSI, and perioperative blood transfusions. They also were more likely to have a 30-day hospital readmission (P < .05 for all comparisons).

When propensity-matched cohorts were compared, patients who used steroids preoperatively were found to have higher rates of any complication (odds Ratio [OR] 1.30, P = .003), sepsis (OR 2.07, P = .022), UTI (OR 1.61, P = .020), superficial SSI (OR 1.73, P = .038), and hospital readmission (OR 1.50, P < .001; Table 3).

REASONS FOR HOSPITAL READMISSION

In total, 3397 patients were readmitted to the hospital within thirty days. Of these, 226 used steroids preoperatively, and 3171 did not (Table 4). 

The most common reason for hospital readmission in patients who used preoperative corticosteroids was infectious complications (72 patients, 31.9% of all readmitted patients in this cohort), followed by medical complications (59 patients, 26.1%), and hip-related complications (48 patients, 21.2%). In those who did not use steroids preoperatively, the most common reason for hospital readmission was medical complications (932 patients, 29.4% of all readmitted patients in this cohort), followed by infectious complications (792 patients, 25.0%), and hip-related complications (763 patients, 24.1%).

Continue to: DISCUSSION

Nearly 3% of individuals >80 years in the US population chronically use corticosteroids for a medical condition,⁸ and this rate is likely higher in specific subsets of patients, such as those with rheumatoid arthritis.⁹ While some studies have assessed the impact of intra-articular corticosteroid hip injections on perioperative outcomes in THA,¹⁷ similar studies assessing systemic corticosteroid usage are lacking. The purpose of this study was to characterize short-term perioperative outcomes in patients undergoing THA who chronically use systemic steroids when compared with those who do not. We found that the prevalence of preoperative chronic steroid use in this cohort of THA patients was 3.7%. We also identified increased rates of infectious complications, including sepsis, UTI, and superficial SSI, in patients who used preoperative corticosteroids. Furthermore, we found an increased rate of hospital readmissions in corticosteroid users and identified the most common reason for hospital readmission as infectious complications in this cohort.

The primary finding of this study was an increase in postoperative infections in patients who use preoperative steroids chronically for medical conditions. Immunosuppression has previously been identified as a risk factor for developing periprosthetic joint infections. Tannenbaum and colleagues²⁷ performed a retrospective study of 19 patients who underwent either a kidney or liver transplant and were maintained on an induction regimen of either prednisone and azathioprine or cyclosporine. These 19 patients also underwent either a THA or total knee arthroplasty, and 5 of these patients (26.3%) developed a periprosthetic joint infection after an average of 3.4 years following the arthroplasty procedure. In another study of 37 renal transplant and dialysis patients who underwent a total of 45 THA procedures, there were 3 instances of superficial SSI and 2 instances of deep SSI.²⁸ However, reported infection rates in transplant patients undergoing THA vary significantly, and studies have been unable to assess the true impact that chronic immunosuppression has on perioperative infection rates.²⁹ In this study, patients who used preoperative corticosteroids chronically were at increased risk of perioperative infections, including sepsis, UTI, and superficial SSI.

Deep vein thrombosis is another postoperative complication that has been associated with chronic steroid use.³⁰ In a case-control study of 38,765 patients who developed a venous thromboembolism and 387,650 control patients who did not, Johannesdottir and colleagues³⁰ found an increased thromboembolic risk in current users of systemic glucocorticoids, but not former users, as well as an increased risk as the dose of glucocorticoids increased. We were not able to identify a similar increase in DVT/PE in chronic corticosteroid users, perhaps due to our sample size, or because we could not do subgroup analyses based on the type or dosage of steroid that a patient was taking. Future studies that identify the highest risk patients among those using systemic corticosteroids are important because parenteral corticosteroids are being increasingly used in THA to alleviate postoperative pain as an opioid-sparing measure.^31,32

Finally, we also found that patients who use chronic, systemic corticosteroids are at an increased risk for hospital readmission, when compared with those patients who are not using steroids and are most likely to be readmitted for an infectious complication. Schairer and colleagues³³ assessed readmission rates after THA and found 30- and 90-day readmission rate of 4% and 7%, respectively. These authors also found that medical complications accounted for approximately 25% of readmissions, and hip-related complications (eg, dislocation, SSI) accounted for >50%. In our study, we found a 30-day readmission rate in non-steroid users of 3.53% and a rate of 6.52% in chronic steroid users. More than 30% of patients using a steroid were readmitted for infectious complications. As THA is becoming increasingly reimbursed under a bundled payments model by Medicare and Medicaid,^34-36 reducing short-term readmissions is imperative. Therefore, discharge counseling that emphasizes how to recognize both the signs and symptoms of infection as well as how to prevent infections, such as reducing SSIs through appropriate wound care, may be warranted in higher risk chronic steroid users.

This study has a number of limitations that are inherent to ACS-NSQIP. First, we lacked specific information on a patient’s steroid history, including which corticosteroid they were using, dosage, frequency, and the indication for corticosteroid therapy. Therefore, we were unable to establish a dose-dependent relationship between steroid exposure and postoperative complications after THA. Second, we were able to assess only 30-day rates of complications and readmissions, and therefore, we were unable to identify intermediate- and long-term effects of systemic corticosteroid use on THA. Finally, we could not determine orthopedic- or hip-specific postoperative outcomes, such as functional scores and range of motion.

Continue to: CONCLUSION

In conclusion, this study quantified the increased risk for perioperative complications and hospital readmissions in patients who chronically use corticosteroids and are undergoing THA, when compared with those who do not use corticosteroids. These results suggest that patients who are on long-term steroids are at an increased risk for complications, primarily infectious complications. This finding has important implications for patient counseling, preoperative risk stratification, and suggests that higher risk patients, such as chronic steroid users, may benefit from improved discharge care to decrease complication rates.

There will be 53 million older adults in the US by 2020.¹ Increasing age often brings medical comorbidities and prescriptions for multiple medications. An increasing number of prescribed medications combined with age-related changes in the ability to metabolize drugs makes older adults highly vulnerable to adverse drug events (ADEs).² In addition, older adults often have difficulty self-managing their medications and adhering to prescribed regimens.³ As a result, ADEs can lead to poor health outcomes, including hospitalizations, in older adults.

Medication errors and ADEs are particularly common during transitions from hospital to home and can lead to unnecessary readmissions,a major cause of wasteful health care spending in the US.^4,5 More than $25 billion are estimated to be spent annually on hospital readmissions, with Medicare picking up the bill for $17 billion of the total.^6,7 Researchers have found that the majority of ADEs following hospital discharge are either entirely preventable or at least ameliorable (ie, the negative impact or harm resulting from the ADE could have been reduced).⁸

To address these issues, we undertook a clinical demonstration project that implemented a new transitional care program to improve the quality of care for older veterans transitioning from the Audie L. Murphy Veterans Memorial Hospital of the South Texas Veterans Health Care System (STVHCS) in San Antonio to home. The Geriatrics Medication Education at Discharge project (GMED) falls under the auspices of the San Antonio Geriatrics Research Education and Clinical Center (GRECC). Clinical demonstration projects are mandated for US Department of Veterans Affairs (VA) GRECCs to create and promote innovative models of care for older veterans. Dissemination of successful clinical demonstration projects to other VA sites is strongly encouraged. The GMED program was modeled after the Boston GRECC Pharmacological Intervention in Late Life (PILL) program.⁹ The PILL program, which focuses on serving older veterans with cognitive impairment, demonstrated that a postdischarge pharmacist telephone visit for medication reconciliation leads to a reduction in readmission within 60 days of discharge.⁹ The goals of the GMED program were to reduce polypharmacy, inappropriate prescribing and 30-day readmissions.

Methods

The project was conducted when a full-time clinical pharmacy specialist (CPS) was available (May-September 2013 and April 2014-March 2015). This project was approved as nonresearch/quality improvement by the University of Texas Health Science Center Institutional Review Board, which serves the STVHCS. Consent was not required.

Eligibility

Patients were identified via a daily hospital database query of all adults aged ≥ 65 years admitted to the hospital through Inpatient Medicine, Neurology, or Cardiology services within the prior 24 hours. Patients meeting any of the following criteria based on review of the Computerized Patient Record System (CPRS) by the team geriatrician and CPS were considered eligible: (1) aged ≥ 70 years prescribed ≥ 12 outpatient medications; (2) aged ≥ 65 years with a medical history of dementia; (3) aged ≥ 65 years prescribed outpatient medications meeting Beers criteria¹⁰; (4) age ≥ 65 years with ≥ 2 hospital admissions (including the current, index admission) within the past calendar year; or (5) aged ≥ 65 years with ≥ 3 emergency department visits within the past calendar year. For the first polypharmacy criterion, patients aged ≥ 70 years were selected instead of aged ≥ 65 years so as not to exceed the capacity of 1 CPS. Twelve or more medications were used as a cutoff for polypharmacy based on prior quality improvement information gathered from our VA geriatrics clinic examining the average number of medications taken by older veterans in the outpatient setting.

Related: Reducing COPD Readmission Rates: Using a COPD Care Service During Care Transitions

Patients were excluded if they were expected to be discharged to any facility where the patient and/or the caregiver were not primarily responsible for medication administration after discharge. Patients who met eligibility criteria but were not seen by the transitional program pharmacist (due to staff capacity) were included in this analysis as a convenience comparison group of patients who received usual care. Patients were not randomized. All communication occurred in English, but this project did not exclude patients with limited English proficiency.

A program support assistant conducted the daily query of the hospital database. The pharmacist conducted the chart review to determine eligibility and delivered the intervention. Eligible patients were selected at random for the intervention with the intention of providing the intervention to as many veterans as possible.

The GMED Intervention

The GMED program included 2 phases, which were both conducted by a CPS with oversight from a senior CPS with geriatric pharmacology expertise and an internist/geriatrician. 

The first phase of the transitional care program included an individual, face-to-face meeting between the CPS and the patient during the hospitalization. If a veteran was not present in the room at the time of an attempted visit, the pharmacist made 2 additional attempts (3 total) to include the patient in the transitional care program during the hospitalization. 

The second component of the transitional care program included a telephone visit within 2 to 3 days of discharge, conducted by the same CPS who performed the face-to-face visit. The purpose of the telephone visit was to perform medication reconciliation, identify and rectify medication errors, provide further patient education, and assist in facilitating appropriate follow-up by the patient’s primary care provider (PCP), if required. At a minimum, veterans were asked a series of questions pertaining to their concerns about medication regimens, receipt of newly prescribed medications at discharge, additional education regarding medications after the CPS encounter during hospitalization, and whether the veteran required assistance with the medication regimen in the home setting. Follow-up questions were asked as needed to clarify and identify potential medication problems. All information from this telephone encounter was communicated to the PCP through CPRS documentation and by telephone as needed.

Related: Initiative to Minimize Pharmaceutical Risk in Older Veterans (IMPROVE) Polypharmacy Clinic

Data Collection

A standardized questionnaire was used prospectively for patients in the transitional care program group to assess patient education, primary residence, presence of a caregiver, fall history, medication adherence, and cognitive status (using Mini-Cog).¹³ Additional information (patient age, number of outpatient medications prior to and following the admission, presence of Beers criteria outpatient medications prior to and following the admission, new outpatient prescriptions, and changes to existing prescriptions as a result of the hospitalization) was gathered prospectively from patient interviews or from chart review.

For patients included in the comparison group, a retrospective administrative chart review was conducted to collect information such as age, sex, ethnic group, admission within 1 year prior to index admission, frailty, and Charlson Comorbidity Index (CCI) score, a method of categorizing comorbidities of patients based on the diagnosis codes found in administrative data.¹⁴ Each comorbidity category has an associated weight (from 1 to 6), based on the adjusted risk of mortality or resource use, and the sum of all the weights results in a single comorbidity score for a patient (0 indicates no comorbidities; higher scores predict greater risk of mortality or increased resource use).

We used the index developed from 17 disease categories. The range for CCI was 0 to 25. Frailty was defined as the presence of any of the following frailty-related diagnoses: anemia; fall, head injury, other injury; coagulopathy; electrolyte disturbance; or gait disorder. These diagnoses are either primary frailty characteristics within the frailty phenotype or have been shown in prior studies to be associated with the frailty phenotype.^15-18 While more widely accepted frailty definitions exist,these other definitions require direct examination of the patient and could not be used in this project because we did not directly interact with the comparison group.^16,19 The frailty definition used has been previously identified as a predictor of health care utilization and 30-day readmission in a veteran population.²⁰ Whether or not the CPS detected a postdischarge medication error was recorded. All CPS recommendations were documented.

An index admission was defined as a hospital admission that occurred during the project period. Thirty-day readmission was defined as a hospital admission that occurred within 30 days of the discharge date of an index admission. Each index admission was considered individually for readmission (yes vs no) even if it occurred in the same patient over the project period. A 30-day readmission was not considered an index admission. An admission that occurred after a 30-day readmission was considered a subsequent index admission. Patients who died in the hospital were not included in this analysis, as they would not have participated in the entire intervention.

Statistical Analysis

We compared characteristics between patients who received GMED and patients who never received GMED (comparison group). Generalized estimating equations (GEE) were used to determine whether the rate of 30-day readmission (yes vs no) in the transitional care program group differed from that of the comparison group. In our GEE analysis, we assumed a binomial distribution and the logit link to model the log-odds of readmission as a linear function of transitional care program status (yes vs no) and other covariates, including age, frailty, hospital admission within 1 year prior to the index admission, and CCI score as covariates. Thirty-day readmission status associated with each index admission was coded as 1 for a readmission within 30 days of the discharge date of the index admission, or 0 for no readmission within 30 days.

Transitional care program status was determined whether or not the individual received the transitional care program for each index admission. This analysis allowed us to model repeated measures of index admissions as a function of the project period and whether the patient was seen by the GMED CPS during the index admission. The patient identifier was used as a cluster variable in the GEE analysis. Inverse propensity scores of receiving GMED at the index admission were adjusted as weights in the GEE analysis to minimize confounding and, hence, to strengthen the causal interpretation of the effect of the transitional care program. If there was ≥ 1 index admission, the GMED status (yes vs no) at the initial index admission was used as the dependent variable to calculate propensity scores. The propensity scores of transitional care program status were derived from the logistic regression analysis that modeled the log-odds of receiving the transitional care program at the index admission as a linear function of age, CCI, frailty, and prior hospitalization during the 1-year period prior to the index admission.

Related: Development and Implementation of a Geriatric Walking Clinic

Results

The GMED CPS saw 435 patients during the project period; 47 (10.8%) died prior to 30 days and were excluded, leaving 388 patients who received the transitional care program included in this evaluation. 

Data from the CPS-patient interviews and chart reviews were available for 378 of the 388 patients (Table 2). Patients were primarily male, non-Hispanic white, with a high school education. More than half (65%) the patients were admitted for a new diagnosis or clinical condition. 

The 30-day readmission rate was 15.6% for the transitional care program group and 21.9% for the comparison group. Three hundred seventy-one patients received the transitional care program only once, 16 patients received the transitional care program twice (ie, they had 2 index admissions during the study period and received the intervention both times), and 1 patient received the transitional care program 3 times.

In an unadjusted GEE model, the odds ratio (OR) for readmission in the transitional care program group was 0.74 (95% CI, 0.54-1.0, P = .06) compared with the usual care group (Table 3). 

Thirty-five percent of patients had ≥ 1 CPS-recommended change in their treatment at the time of the inpatient admission (Table 4). 

Discussion

We developed a transitional care program for hospitalized older veterans to improve the transition from hospital to home. After adjusting for clinical factors, GMED was associated with 26% lower odds of readmission within 30 days of discharge compared with that of the control group. The GMED CPS made changes to the medical regimen both during the inpatient admission as well as after discharge to correct medication errors and educate patients.

In addition, GMED led to a reduction in the number of prescribed medications, which impacts inappropriate polypharmacy—a significant problem in older adults, which contributes to ADEs.²¹ Our intervention was patient centered, as all decisions and education regarding medication management were tailored to each patient, taking into account medical and psychosocial factors.

Studies of similar programs have shown that a pharmacist-based program can improve outcomes in patients transitioning from hospital to home. A meta-analysis of 19 studies that evaluated the effectiveness of pharmacy-led medication reconciliation interventions at the time of a care transition showed that compared with usual care a pharmacist intervention led to reduced medication discrepancies.²² In this meta-analysis, medication discrepancies of higher clinical impact were more easily identified through pharmacy-led interventions than with usual care, suggesting improved safety. Although not all studies have shown a clear reduction in readmission rates or other health care utilization, the addition of clinical pharmacist services in the care of inpatients has generally resulted in improved care with no evidence of harm.²³

Based on these findings and collaboration with another GRECC, we designed our program to focus on older adults with polypharmacy, cognitive impairment, high-risk medication usage, and/or a history of high health care use.⁹ Our findings add to the growing body of evidence that a CPS-led transitional care program results in reduced polypharmacy and reduced unnecessary hospital readmissions. Further, our findings have demonstrated the effectiveness of this type of program in a practical, clinical setting with veteran patients.

At the time of project inception, we believed that the majority of our interventions would occur postdischarge. We were somewhat surprised that a major component of GMED was suggested interventions by our pharmacist at the time of admission. We believe that because the CPS made suggestions during admission, we prevented postdischarge ADEs. A frequent intervention corrected the medication reconciliation on file at admission. This finding also was seen in another study by Gleason and colleagues, which examined medication errors at admission for 651 adult medicine inpatients.²⁴ This study found that more than one-third of patients had medication reconciliation errors. Further, older age (≥ 65 years) was associated with increased odds of medication errors in this study.

Of note, a survey of hospital-based pharmacists indicated medication reconciliation is the most important role of the pharmacist in improving care transitions.²⁵ The pharmacists stated that detection of errors at the time of admission is very important. The pharmacists further reported that additional education and counseling for patients with poor understanding of their medications was also important. Our findings support these findings and the use of a pharmacist as part of the medical team to improve medication reconciliation and education.

Limitations

A limitation of GMED is that we monitored only admissions to our hospital; therefore, we did not account for any hospitalizations that may have occurred outside the STVHCS. Another limitation is that this was not a randomized controlled trial, and we used a convenience sample of patients who met our criteria for eligibility but were not seen due to time constraints. This introduces potential bias such that patients admitted and discharged on nights or weekends when the CPS was not available were not included in the transitional care program group, and these patients may fundamentally differ from those admitted and discharged Monday through Friday.

However, Khanna and colleagues found that night or weekend admission was not associated with 30-day readmission or other worse outcomes (such as length of stay, 30-day emergency department visit, or intensive care unit transfer) in 857 general medicine admissions at a tertiary care hospital.²⁶ Every effort was made to include as many eligible patients as possible in the transitional program group, and we were able to demonstrate that the patients in the 2 groups were similar. Frailty and prior hospital admission were more prevalent, although not significantly so, in the transitional program group, suggesting that any selection bias would have actually attenuated—not enhanced—the observed effect of the transitional program. Although the transitional program group patients were slightly younger by 0.3 years, they were similar in frailty status and CCI score.

Conclusion

The GMED program was associated with reduced 30-day hospital readmission, discontinuation of unnecessary medications, and corrected medication errors and discrepancies. We propose that a CPS-based transitional care program can improve the quality of care for older patients being discharged to home.

Acknowledgments

Supported by funding from the Veterans Health Administration T21 Non-Institutional Long-Term Care Initiative and VA Office of Rural Health and the San Antonio Geriatrics Research, Education, and Clinical Center. The sponsor did not have any role in the design, methods, data collection, or analysis, and preparation.

Author Contributions

R. Rottman-Sagebiel developed the transitional program concept and design and executed the program implementation, interpretation of data, and preparation of the manuscript. S. Pastewait, N. Cupples, A. Conde, M. Moris, and E. Gonzalez assisted with program design and implementation. S. Cope assisted with interpretation of data and preparation of the manuscript. H. Braden assisted with interpretation of data. D. MacCarthy assisted with data management and statistical analysis. C. Wang and S. Espinoza developed the program concept and design, performed statistical analysis and interpretation of data, and helped prepare the manuscript.

Advances in Geriatrics

Advances in Geriatrics features articles focused on quality improvement/quality assurance initiatives, pilot studies, best practices, research, patient education, and patient-centered care written by health care providers associated with Veteran Health Administration Geriatric Research Education and Clinical Centers. Interested authors can submit articles at editorialmanager.com/fedprac or send a brief 2 to 3 sentence abstract to [email protected] for feedback and publication recommendations.

There will be 53 million older adults in the US by 2020.¹ Increasing age often brings medical comorbidities and prescriptions for multiple medications. An increasing number of prescribed medications combined with age-related changes in the ability to metabolize drugs makes older adults highly vulnerable to adverse drug events (ADEs).² In addition, older adults often have difficulty self-managing their medications and adhering to prescribed regimens.³ As a result, ADEs can lead to poor health outcomes, including hospitalizations, in older adults.

Medication errors and ADEs are particularly common during transitions from hospital to home and can lead to unnecessary readmissions,a major cause of wasteful health care spending in the US.^4,5 More than $25 billion are estimated to be spent annually on hospital readmissions, with Medicare picking up the bill for $17 billion of the total.^6,7 Researchers have found that the majority of ADEs following hospital discharge are either entirely preventable or at least ameliorable (ie, the negative impact or harm resulting from the ADE could have been reduced).⁸

To address these issues, we undertook a clinical demonstration project that implemented a new transitional care program to improve the quality of care for older veterans transitioning from the Audie L. Murphy Veterans Memorial Hospital of the South Texas Veterans Health Care System (STVHCS) in San Antonio to home. The Geriatrics Medication Education at Discharge project (GMED) falls under the auspices of the San Antonio Geriatrics Research Education and Clinical Center (GRECC). Clinical demonstration projects are mandated for US Department of Veterans Affairs (VA) GRECCs to create and promote innovative models of care for older veterans. Dissemination of successful clinical demonstration projects to other VA sites is strongly encouraged. The GMED program was modeled after the Boston GRECC Pharmacological Intervention in Late Life (PILL) program.⁹ The PILL program, which focuses on serving older veterans with cognitive impairment, demonstrated that a postdischarge pharmacist telephone visit for medication reconciliation leads to a reduction in readmission within 60 days of discharge.⁹ The goals of the GMED program were to reduce polypharmacy, inappropriate prescribing and 30-day readmissions.

Methods

The project was conducted when a full-time clinical pharmacy specialist (CPS) was available (May-September 2013 and April 2014-March 2015). This project was approved as nonresearch/quality improvement by the University of Texas Health Science Center Institutional Review Board, which serves the STVHCS. Consent was not required.

Eligibility

Patients were identified via a daily hospital database query of all adults aged ≥ 65 years admitted to the hospital through Inpatient Medicine, Neurology, or Cardiology services within the prior 24 hours. Patients meeting any of the following criteria based on review of the Computerized Patient Record System (CPRS) by the team geriatrician and CPS were considered eligible: (1) aged ≥ 70 years prescribed ≥ 12 outpatient medications; (2) aged ≥ 65 years with a medical history of dementia; (3) aged ≥ 65 years prescribed outpatient medications meeting Beers criteria¹⁰; (4) age ≥ 65 years with ≥ 2 hospital admissions (including the current, index admission) within the past calendar year; or (5) aged ≥ 65 years with ≥ 3 emergency department visits within the past calendar year. For the first polypharmacy criterion, patients aged ≥ 70 years were selected instead of aged ≥ 65 years so as not to exceed the capacity of 1 CPS. Twelve or more medications were used as a cutoff for polypharmacy based on prior quality improvement information gathered from our VA geriatrics clinic examining the average number of medications taken by older veterans in the outpatient setting.

Related: Reducing COPD Readmission Rates: Using a COPD Care Service During Care Transitions

Patients were excluded if they were expected to be discharged to any facility where the patient and/or the caregiver were not primarily responsible for medication administration after discharge. Patients who met eligibility criteria but were not seen by the transitional program pharmacist (due to staff capacity) were included in this analysis as a convenience comparison group of patients who received usual care. Patients were not randomized. All communication occurred in English, but this project did not exclude patients with limited English proficiency.

A program support assistant conducted the daily query of the hospital database. The pharmacist conducted the chart review to determine eligibility and delivered the intervention. Eligible patients were selected at random for the intervention with the intention of providing the intervention to as many veterans as possible.

The GMED Intervention

The GMED program included 2 phases, which were both conducted by a CPS with oversight from a senior CPS with geriatric pharmacology expertise and an internist/geriatrician. 

The first phase of the transitional care program included an individual, face-to-face meeting between the CPS and the patient during the hospitalization. If a veteran was not present in the room at the time of an attempted visit, the pharmacist made 2 additional attempts (3 total) to include the patient in the transitional care program during the hospitalization. 

The second component of the transitional care program included a telephone visit within 2 to 3 days of discharge, conducted by the same CPS who performed the face-to-face visit. The purpose of the telephone visit was to perform medication reconciliation, identify and rectify medication errors, provide further patient education, and assist in facilitating appropriate follow-up by the patient’s primary care provider (PCP), if required. At a minimum, veterans were asked a series of questions pertaining to their concerns about medication regimens, receipt of newly prescribed medications at discharge, additional education regarding medications after the CPS encounter during hospitalization, and whether the veteran required assistance with the medication regimen in the home setting. Follow-up questions were asked as needed to clarify and identify potential medication problems. All information from this telephone encounter was communicated to the PCP through CPRS documentation and by telephone as needed.

Related: Initiative to Minimize Pharmaceutical Risk in Older Veterans (IMPROVE) Polypharmacy Clinic

Data Collection

A standardized questionnaire was used prospectively for patients in the transitional care program group to assess patient education, primary residence, presence of a caregiver, fall history, medication adherence, and cognitive status (using Mini-Cog).¹³ Additional information (patient age, number of outpatient medications prior to and following the admission, presence of Beers criteria outpatient medications prior to and following the admission, new outpatient prescriptions, and changes to existing prescriptions as a result of the hospitalization) was gathered prospectively from patient interviews or from chart review.

For patients included in the comparison group, a retrospective administrative chart review was conducted to collect information such as age, sex, ethnic group, admission within 1 year prior to index admission, frailty, and Charlson Comorbidity Index (CCI) score, a method of categorizing comorbidities of patients based on the diagnosis codes found in administrative data.¹⁴ Each comorbidity category has an associated weight (from 1 to 6), based on the adjusted risk of mortality or resource use, and the sum of all the weights results in a single comorbidity score for a patient (0 indicates no comorbidities; higher scores predict greater risk of mortality or increased resource use).

We used the index developed from 17 disease categories. The range for CCI was 0 to 25. Frailty was defined as the presence of any of the following frailty-related diagnoses: anemia; fall, head injury, other injury; coagulopathy; electrolyte disturbance; or gait disorder. These diagnoses are either primary frailty characteristics within the frailty phenotype or have been shown in prior studies to be associated with the frailty phenotype.^15-18 While more widely accepted frailty definitions exist,these other definitions require direct examination of the patient and could not be used in this project because we did not directly interact with the comparison group.^16,19 The frailty definition used has been previously identified as a predictor of health care utilization and 30-day readmission in a veteran population.²⁰ Whether or not the CPS detected a postdischarge medication error was recorded. All CPS recommendations were documented.

An index admission was defined as a hospital admission that occurred during the project period. Thirty-day readmission was defined as a hospital admission that occurred within 30 days of the discharge date of an index admission. Each index admission was considered individually for readmission (yes vs no) even if it occurred in the same patient over the project period. A 30-day readmission was not considered an index admission. An admission that occurred after a 30-day readmission was considered a subsequent index admission. Patients who died in the hospital were not included in this analysis, as they would not have participated in the entire intervention.

Statistical Analysis

We compared characteristics between patients who received GMED and patients who never received GMED (comparison group). Generalized estimating equations (GEE) were used to determine whether the rate of 30-day readmission (yes vs no) in the transitional care program group differed from that of the comparison group. In our GEE analysis, we assumed a binomial distribution and the logit link to model the log-odds of readmission as a linear function of transitional care program status (yes vs no) and other covariates, including age, frailty, hospital admission within 1 year prior to the index admission, and CCI score as covariates. Thirty-day readmission status associated with each index admission was coded as 1 for a readmission within 30 days of the discharge date of the index admission, or 0 for no readmission within 30 days.

Transitional care program status was determined whether or not the individual received the transitional care program for each index admission. This analysis allowed us to model repeated measures of index admissions as a function of the project period and whether the patient was seen by the GMED CPS during the index admission. The patient identifier was used as a cluster variable in the GEE analysis. Inverse propensity scores of receiving GMED at the index admission were adjusted as weights in the GEE analysis to minimize confounding and, hence, to strengthen the causal interpretation of the effect of the transitional care program. If there was ≥ 1 index admission, the GMED status (yes vs no) at the initial index admission was used as the dependent variable to calculate propensity scores. The propensity scores of transitional care program status were derived from the logistic regression analysis that modeled the log-odds of receiving the transitional care program at the index admission as a linear function of age, CCI, frailty, and prior hospitalization during the 1-year period prior to the index admission.

Related: Development and Implementation of a Geriatric Walking Clinic

Results

The GMED CPS saw 435 patients during the project period; 47 (10.8%) died prior to 30 days and were excluded, leaving 388 patients who received the transitional care program included in this evaluation. 

Data from the CPS-patient interviews and chart reviews were available for 378 of the 388 patients (Table 2). Patients were primarily male, non-Hispanic white, with a high school education. More than half (65%) the patients were admitted for a new diagnosis or clinical condition. 

The 30-day readmission rate was 15.6% for the transitional care program group and 21.9% for the comparison group. Three hundred seventy-one patients received the transitional care program only once, 16 patients received the transitional care program twice (ie, they had 2 index admissions during the study period and received the intervention both times), and 1 patient received the transitional care program 3 times.

In an unadjusted GEE model, the odds ratio (OR) for readmission in the transitional care program group was 0.74 (95% CI, 0.54-1.0, P = .06) compared with the usual care group (Table 3). 

Thirty-five percent of patients had ≥ 1 CPS-recommended change in their treatment at the time of the inpatient admission (Table 4). 

Discussion

We developed a transitional care program for hospitalized older veterans to improve the transition from hospital to home. After adjusting for clinical factors, GMED was associated with 26% lower odds of readmission within 30 days of discharge compared with that of the control group. The GMED CPS made changes to the medical regimen both during the inpatient admission as well as after discharge to correct medication errors and educate patients.

In addition, GMED led to a reduction in the number of prescribed medications, which impacts inappropriate polypharmacy—a significant problem in older adults, which contributes to ADEs.²¹ Our intervention was patient centered, as all decisions and education regarding medication management were tailored to each patient, taking into account medical and psychosocial factors.

Studies of similar programs have shown that a pharmacist-based program can improve outcomes in patients transitioning from hospital to home. A meta-analysis of 19 studies that evaluated the effectiveness of pharmacy-led medication reconciliation interventions at the time of a care transition showed that compared with usual care a pharmacist intervention led to reduced medication discrepancies.²² In this meta-analysis, medication discrepancies of higher clinical impact were more easily identified through pharmacy-led interventions than with usual care, suggesting improved safety. Although not all studies have shown a clear reduction in readmission rates or other health care utilization, the addition of clinical pharmacist services in the care of inpatients has generally resulted in improved care with no evidence of harm.²³

Based on these findings and collaboration with another GRECC, we designed our program to focus on older adults with polypharmacy, cognitive impairment, high-risk medication usage, and/or a history of high health care use.⁹ Our findings add to the growing body of evidence that a CPS-led transitional care program results in reduced polypharmacy and reduced unnecessary hospital readmissions. Further, our findings have demonstrated the effectiveness of this type of program in a practical, clinical setting with veteran patients.

At the time of project inception, we believed that the majority of our interventions would occur postdischarge. We were somewhat surprised that a major component of GMED was suggested interventions by our pharmacist at the time of admission. We believe that because the CPS made suggestions during admission, we prevented postdischarge ADEs. A frequent intervention corrected the medication reconciliation on file at admission. This finding also was seen in another study by Gleason and colleagues, which examined medication errors at admission for 651 adult medicine inpatients.²⁴ This study found that more than one-third of patients had medication reconciliation errors. Further, older age (≥ 65 years) was associated with increased odds of medication errors in this study.

Of note, a survey of hospital-based pharmacists indicated medication reconciliation is the most important role of the pharmacist in improving care transitions.²⁵ The pharmacists stated that detection of errors at the time of admission is very important. The pharmacists further reported that additional education and counseling for patients with poor understanding of their medications was also important. Our findings support these findings and the use of a pharmacist as part of the medical team to improve medication reconciliation and education.

Limitations

A limitation of GMED is that we monitored only admissions to our hospital; therefore, we did not account for any hospitalizations that may have occurred outside the STVHCS. Another limitation is that this was not a randomized controlled trial, and we used a convenience sample of patients who met our criteria for eligibility but were not seen due to time constraints. This introduces potential bias such that patients admitted and discharged on nights or weekends when the CPS was not available were not included in the transitional care program group, and these patients may fundamentally differ from those admitted and discharged Monday through Friday.

However, Khanna and colleagues found that night or weekend admission was not associated with 30-day readmission or other worse outcomes (such as length of stay, 30-day emergency department visit, or intensive care unit transfer) in 857 general medicine admissions at a tertiary care hospital.²⁶ Every effort was made to include as many eligible patients as possible in the transitional program group, and we were able to demonstrate that the patients in the 2 groups were similar. Frailty and prior hospital admission were more prevalent, although not significantly so, in the transitional program group, suggesting that any selection bias would have actually attenuated—not enhanced—the observed effect of the transitional program. Although the transitional program group patients were slightly younger by 0.3 years, they were similar in frailty status and CCI score.

Conclusion

The GMED program was associated with reduced 30-day hospital readmission, discontinuation of unnecessary medications, and corrected medication errors and discrepancies. We propose that a CPS-based transitional care program can improve the quality of care for older patients being discharged to home.

Acknowledgments

Supported by funding from the Veterans Health Administration T21 Non-Institutional Long-Term Care Initiative and VA Office of Rural Health and the San Antonio Geriatrics Research, Education, and Clinical Center. The sponsor did not have any role in the design, methods, data collection, or analysis, and preparation.

Author Contributions

R. Rottman-Sagebiel developed the transitional program concept and design and executed the program implementation, interpretation of data, and preparation of the manuscript. S. Pastewait, N. Cupples, A. Conde, M. Moris, and E. Gonzalez assisted with program design and implementation. S. Cope assisted with interpretation of data and preparation of the manuscript. H. Braden assisted with interpretation of data. D. MacCarthy assisted with data management and statistical analysis. C. Wang and S. Espinoza developed the program concept and design, performed statistical analysis and interpretation of data, and helped prepare the manuscript.

Advances in Geriatrics

Advances in Geriatrics features articles focused on quality improvement/quality assurance initiatives, pilot studies, best practices, research, patient education, and patient-centered care written by health care providers associated with Veteran Health Administration Geriatric Research Education and Clinical Centers. Interested authors can submit articles at editorialmanager.com/fedprac or send a brief 2 to 3 sentence abstract to [email protected] for feedback and publication recommendations.

There will be 53 million older adults in the US by 2020.¹ Increasing age often brings medical comorbidities and prescriptions for multiple medications. An increasing number of prescribed medications combined with age-related changes in the ability to metabolize drugs makes older adults highly vulnerable to adverse drug events (ADEs).² In addition, older adults often have difficulty self-managing their medications and adhering to prescribed regimens.³ As a result, ADEs can lead to poor health outcomes, including hospitalizations, in older adults.

Medication errors and ADEs are particularly common during transitions from hospital to home and can lead to unnecessary readmissions,a major cause of wasteful health care spending in the US.^4,5 More than $25 billion are estimated to be spent annually on hospital readmissions, with Medicare picking up the bill for $17 billion of the total.^6,7 Researchers have found that the majority of ADEs following hospital discharge are either entirely preventable or at least ameliorable (ie, the negative impact or harm resulting from the ADE could have been reduced).⁸

To address these issues, we undertook a clinical demonstration project that implemented a new transitional care program to improve the quality of care for older veterans transitioning from the Audie L. Murphy Veterans Memorial Hospital of the South Texas Veterans Health Care System (STVHCS) in San Antonio to home. The Geriatrics Medication Education at Discharge project (GMED) falls under the auspices of the San Antonio Geriatrics Research Education and Clinical Center (GRECC). Clinical demonstration projects are mandated for US Department of Veterans Affairs (VA) GRECCs to create and promote innovative models of care for older veterans. Dissemination of successful clinical demonstration projects to other VA sites is strongly encouraged. The GMED program was modeled after the Boston GRECC Pharmacological Intervention in Late Life (PILL) program.⁹ The PILL program, which focuses on serving older veterans with cognitive impairment, demonstrated that a postdischarge pharmacist telephone visit for medication reconciliation leads to a reduction in readmission within 60 days of discharge.⁹ The goals of the GMED program were to reduce polypharmacy, inappropriate prescribing and 30-day readmissions.

Methods

The project was conducted when a full-time clinical pharmacy specialist (CPS) was available (May-September 2013 and April 2014-March 2015). This project was approved as nonresearch/quality improvement by the University of Texas Health Science Center Institutional Review Board, which serves the STVHCS. Consent was not required.

Eligibility

Patients were identified via a daily hospital database query of all adults aged ≥ 65 years admitted to the hospital through Inpatient Medicine, Neurology, or Cardiology services within the prior 24 hours. Patients meeting any of the following criteria based on review of the Computerized Patient Record System (CPRS) by the team geriatrician and CPS were considered eligible: (1) aged ≥ 70 years prescribed ≥ 12 outpatient medications; (2) aged ≥ 65 years with a medical history of dementia; (3) aged ≥ 65 years prescribed outpatient medications meeting Beers criteria¹⁰; (4) age ≥ 65 years with ≥ 2 hospital admissions (including the current, index admission) within the past calendar year; or (5) aged ≥ 65 years with ≥ 3 emergency department visits within the past calendar year. For the first polypharmacy criterion, patients aged ≥ 70 years were selected instead of aged ≥ 65 years so as not to exceed the capacity of 1 CPS. Twelve or more medications were used as a cutoff for polypharmacy based on prior quality improvement information gathered from our VA geriatrics clinic examining the average number of medications taken by older veterans in the outpatient setting.

Related: Reducing COPD Readmission Rates: Using a COPD Care Service During Care Transitions

Patients were excluded if they were expected to be discharged to any facility where the patient and/or the caregiver were not primarily responsible for medication administration after discharge. Patients who met eligibility criteria but were not seen by the transitional program pharmacist (due to staff capacity) were included in this analysis as a convenience comparison group of patients who received usual care. Patients were not randomized. All communication occurred in English, but this project did not exclude patients with limited English proficiency.

A program support assistant conducted the daily query of the hospital database. The pharmacist conducted the chart review to determine eligibility and delivered the intervention. Eligible patients were selected at random for the intervention with the intention of providing the intervention to as many veterans as possible.

The GMED Intervention

The GMED program included 2 phases, which were both conducted by a CPS with oversight from a senior CPS with geriatric pharmacology expertise and an internist/geriatrician. 

The first phase of the transitional care program included an individual, face-to-face meeting between the CPS and the patient during the hospitalization. If a veteran was not present in the room at the time of an attempted visit, the pharmacist made 2 additional attempts (3 total) to include the patient in the transitional care program during the hospitalization. 

The second component of the transitional care program included a telephone visit within 2 to 3 days of discharge, conducted by the same CPS who performed the face-to-face visit. The purpose of the telephone visit was to perform medication reconciliation, identify and rectify medication errors, provide further patient education, and assist in facilitating appropriate follow-up by the patient’s primary care provider (PCP), if required. At a minimum, veterans were asked a series of questions pertaining to their concerns about medication regimens, receipt of newly prescribed medications at discharge, additional education regarding medications after the CPS encounter during hospitalization, and whether the veteran required assistance with the medication regimen in the home setting. Follow-up questions were asked as needed to clarify and identify potential medication problems. All information from this telephone encounter was communicated to the PCP through CPRS documentation and by telephone as needed.

Related: Initiative to Minimize Pharmaceutical Risk in Older Veterans (IMPROVE) Polypharmacy Clinic

Data Collection

A standardized questionnaire was used prospectively for patients in the transitional care program group to assess patient education, primary residence, presence of a caregiver, fall history, medication adherence, and cognitive status (using Mini-Cog).¹³ Additional information (patient age, number of outpatient medications prior to and following the admission, presence of Beers criteria outpatient medications prior to and following the admission, new outpatient prescriptions, and changes to existing prescriptions as a result of the hospitalization) was gathered prospectively from patient interviews or from chart review.

For patients included in the comparison group, a retrospective administrative chart review was conducted to collect information such as age, sex, ethnic group, admission within 1 year prior to index admission, frailty, and Charlson Comorbidity Index (CCI) score, a method of categorizing comorbidities of patients based on the diagnosis codes found in administrative data.¹⁴ Each comorbidity category has an associated weight (from 1 to 6), based on the adjusted risk of mortality or resource use, and the sum of all the weights results in a single comorbidity score for a patient (0 indicates no comorbidities; higher scores predict greater risk of mortality or increased resource use).

We used the index developed from 17 disease categories. The range for CCI was 0 to 25. Frailty was defined as the presence of any of the following frailty-related diagnoses: anemia; fall, head injury, other injury; coagulopathy; electrolyte disturbance; or gait disorder. These diagnoses are either primary frailty characteristics within the frailty phenotype or have been shown in prior studies to be associated with the frailty phenotype.^15-18 While more widely accepted frailty definitions exist,these other definitions require direct examination of the patient and could not be used in this project because we did not directly interact with the comparison group.^16,19 The frailty definition used has been previously identified as a predictor of health care utilization and 30-day readmission in a veteran population.²⁰ Whether or not the CPS detected a postdischarge medication error was recorded. All CPS recommendations were documented.

An index admission was defined as a hospital admission that occurred during the project period. Thirty-day readmission was defined as a hospital admission that occurred within 30 days of the discharge date of an index admission. Each index admission was considered individually for readmission (yes vs no) even if it occurred in the same patient over the project period. A 30-day readmission was not considered an index admission. An admission that occurred after a 30-day readmission was considered a subsequent index admission. Patients who died in the hospital were not included in this analysis, as they would not have participated in the entire intervention.

Statistical Analysis

We compared characteristics between patients who received GMED and patients who never received GMED (comparison group). Generalized estimating equations (GEE) were used to determine whether the rate of 30-day readmission (yes vs no) in the transitional care program group differed from that of the comparison group. In our GEE analysis, we assumed a binomial distribution and the logit link to model the log-odds of readmission as a linear function of transitional care program status (yes vs no) and other covariates, including age, frailty, hospital admission within 1 year prior to the index admission, and CCI score as covariates. Thirty-day readmission status associated with each index admission was coded as 1 for a readmission within 30 days of the discharge date of the index admission, or 0 for no readmission within 30 days.

Transitional care program status was determined whether or not the individual received the transitional care program for each index admission. This analysis allowed us to model repeated measures of index admissions as a function of the project period and whether the patient was seen by the GMED CPS during the index admission. The patient identifier was used as a cluster variable in the GEE analysis. Inverse propensity scores of receiving GMED at the index admission were adjusted as weights in the GEE analysis to minimize confounding and, hence, to strengthen the causal interpretation of the effect of the transitional care program. If there was ≥ 1 index admission, the GMED status (yes vs no) at the initial index admission was used as the dependent variable to calculate propensity scores. The propensity scores of transitional care program status were derived from the logistic regression analysis that modeled the log-odds of receiving the transitional care program at the index admission as a linear function of age, CCI, frailty, and prior hospitalization during the 1-year period prior to the index admission.

Related: Development and Implementation of a Geriatric Walking Clinic

Results

The GMED CPS saw 435 patients during the project period; 47 (10.8%) died prior to 30 days and were excluded, leaving 388 patients who received the transitional care program included in this evaluation. 

Data from the CPS-patient interviews and chart reviews were available for 378 of the 388 patients (Table 2). Patients were primarily male, non-Hispanic white, with a high school education. More than half (65%) the patients were admitted for a new diagnosis or clinical condition. 

The 30-day readmission rate was 15.6% for the transitional care program group and 21.9% for the comparison group. Three hundred seventy-one patients received the transitional care program only once, 16 patients received the transitional care program twice (ie, they had 2 index admissions during the study period and received the intervention both times), and 1 patient received the transitional care program 3 times.

In an unadjusted GEE model, the odds ratio (OR) for readmission in the transitional care program group was 0.74 (95% CI, 0.54-1.0, P = .06) compared with the usual care group (Table 3).