Hypertension (HTN) is highly prevalent in the general adult population with recent estimates from the National Health and Nutrition Examination Survey (NHANES) of 29% in the United States.1, 2 The relationship between increasing levels of blood pressure (BP) and increasing risk for cardiovascular disease events and stroke is well established.3 However, while 64% of treated HTN patients have a BP <140/<90 mmHg, overall control rates for HTN in the adult population remain at approximately 44%.2 The 20% discrepancy in control rates between treated patients and the overall adult population reflects the fact that approximately 30% of patients are unaware of their HTN and that a substantial proportion of aware patients remain untreated. Historically, efforts to improve the recognition, treatment, and control of HTN have appropriately focused on the outpatient setting. However, programs to extend screening for HTN outside the clinic into the community, schools, and even dentists' offices have been around for some time.49

The potential also exists to improve the recognition, treatment, and control of HTN by focusing on hospitalized patients. Hospitalization is common in the U.S. with almost 35 million acute hospitalizations and more than 45,000 inpatient surgical procedures in 2006.10 Inpatient populations have increased in age and comorbidity over the past 3 decades whereas lengths of stay and continuity of care between the inpatient and outpatient arenas have diminished.10, 11 Multiple prior studies examining BP in different settings have noted that average BP among hospitalized patients is not systematically higher than that of outpatients.1214 Thus, patients with persistently elevated BP in the inpatient setting without mitigating factors may have HTN that will persist after hospital discharge. However, little information is available regarding the actual prevalence of HTN in the inpatient population and care patterns for inpatient HTN. Therefore, we performed a systematic review of the English‐language medical literature in order to describe the epidemiology of HTN observed in the inpatient setting.

Methods

Our search strategy was designed to identify randomized‐controlled trials, meta‐analyses, and observational studies that: (1) reported estimates of the prevalence of HTN in the inpatient setting, and (2) used HTN diagnosis or treatment as a primary focus. We performed an extensive review of the peer‐reviewed, English language medical literature in MEDLINE using a predetermined search algorithm. Search terms included HTN[Mesh] or BP[Mesh]. These results were cross‐referenced with the following search terms: Inpatients[Mesh] or Hospitalization[title/abstract] or Hospitalized[title/abstract]. Articles were further narrowed using the following terms: Prevalence[Mesh] or Epidemiology[Mesh] or Treatment[title/abstract] or Management [title/abstract]. Limits employed included limiting to humans and to adults 19 years‐of‐age and older. Studies published prior to 1976 were excluded because 1976 was the first year that the Joint National Committee on Prevention, Detection, Evaluation, and Treatment of High BP published consensus guidelines for the diagnosis and management of HTN. We also excluded randomized, controlled trials that recorded measures of inpatient BP but whose focus was not HTN, because such trials would not answer the primary epidemiologic question of this review. We did include trials focused on subspecialty populations for which the diagnosis and inpatient management of HTN were key outcomes.

Next, the bibliographies of reviewed studies were investigated for additional relevant reports. Abstracts from the American Heart Association (AHA) were reviewed for the past 15 years for reports that were presented but not subsequently published and available in MEDLINE. We also searched for articles using the online Google search engine. One author (RNA) performed the preliminary MEDLINE search and abstract review with the assistance of a reference librarian (LC), and a second author (BME) also reviewed full‐text articles for potential inclusion. Ultimate decision for study inclusion was reached through discussion among authors. Finally, a list of potential articles was submitted to 2 experts in this field of study to determine whether other reports met our inclusion criteria for this systematic review but were overlooked.

Results

Search Results

The initial MEDLINE search algorithm yielded a total of 826 articles. After title and abstract review, 41 full‐text articles were obtained for detailed review, and 5 met criteria for inclusion. Three additional articles were discovered through searching the bibliographies of the included studies. No AHA abstracts addressed this subject area. Experts were not aware of any additional studies. One article was located using a Google search. In all, 9 articles were deemed suitable for inclusion in this review. Search results at each stage are depicted in Figure 1.

Figure 1

Hypertension (HTN) is highly prevalent in the general adult population with recent estimates from the National Health and Nutrition Examination Survey (NHANES) of 29% in the United States.1, 2 The relationship between increasing levels of blood pressure (BP) and increasing risk for cardiovascular disease events and stroke is well established.3 However, while 64% of treated HTN patients have a BP <140/<90 mmHg, overall control rates for HTN in the adult population remain at approximately 44%.2 The 20% discrepancy in control rates between treated patients and the overall adult population reflects the fact that approximately 30% of patients are unaware of their HTN and that a substantial proportion of aware patients remain untreated. Historically, efforts to improve the recognition, treatment, and control of HTN have appropriately focused on the outpatient setting. However, programs to extend screening for HTN outside the clinic into the community, schools, and even dentists' offices have been around for some time.49

The potential also exists to improve the recognition, treatment, and control of HTN by focusing on hospitalized patients. Hospitalization is common in the U.S. with almost 35 million acute hospitalizations and more than 45,000 inpatient surgical procedures in 2006.10 Inpatient populations have increased in age and comorbidity over the past 3 decades whereas lengths of stay and continuity of care between the inpatient and outpatient arenas have diminished.10, 11 Multiple prior studies examining BP in different settings have noted that average BP among hospitalized patients is not systematically higher than that of outpatients.1214 Thus, patients with persistently elevated BP in the inpatient setting without mitigating factors may have HTN that will persist after hospital discharge. However, little information is available regarding the actual prevalence of HTN in the inpatient population and care patterns for inpatient HTN. Therefore, we performed a systematic review of the English‐language medical literature in order to describe the epidemiology of HTN observed in the inpatient setting.

Methods

Our search strategy was designed to identify randomized‐controlled trials, meta‐analyses, and observational studies that: (1) reported estimates of the prevalence of HTN in the inpatient setting, and (2) used HTN diagnosis or treatment as a primary focus. We performed an extensive review of the peer‐reviewed, English language medical literature in MEDLINE using a predetermined search algorithm. Search terms included HTN[Mesh] or BP[Mesh]. These results were cross‐referenced with the following search terms: Inpatients[Mesh] or Hospitalization[title/abstract] or Hospitalized[title/abstract]. Articles were further narrowed using the following terms: Prevalence[Mesh] or Epidemiology[Mesh] or Treatment[title/abstract] or Management [title/abstract]. Limits employed included limiting to humans and to adults 19 years‐of‐age and older. Studies published prior to 1976 were excluded because 1976 was the first year that the Joint National Committee on Prevention, Detection, Evaluation, and Treatment of High BP published consensus guidelines for the diagnosis and management of HTN. We also excluded randomized, controlled trials that recorded measures of inpatient BP but whose focus was not HTN, because such trials would not answer the primary epidemiologic question of this review. We did include trials focused on subspecialty populations for which the diagnosis and inpatient management of HTN were key outcomes.

Next, the bibliographies of reviewed studies were investigated for additional relevant reports. Abstracts from the American Heart Association (AHA) were reviewed for the past 15 years for reports that were presented but not subsequently published and available in MEDLINE. We also searched for articles using the online Google search engine. One author (RNA) performed the preliminary MEDLINE search and abstract review with the assistance of a reference librarian (LC), and a second author (BME) also reviewed full‐text articles for potential inclusion. Ultimate decision for study inclusion was reached through discussion among authors. Finally, a list of potential articles was submitted to 2 experts in this field of study to determine whether other reports met our inclusion criteria for this systematic review but were overlooked.

Results

Search Results

The initial MEDLINE search algorithm yielded a total of 826 articles. After title and abstract review, 41 full‐text articles were obtained for detailed review, and 5 met criteria for inclusion. Three additional articles were discovered through searching the bibliographies of the included studies. No AHA abstracts addressed this subject area. Experts were not aware of any additional studies. One article was located using a Google search. In all, 9 articles were deemed suitable for inclusion in this review. Search results at each stage are depicted in Figure 1.

Figure 1

Hypertension (HTN) is highly prevalent in the general adult population with recent estimates from the National Health and Nutrition Examination Survey (NHANES) of 29% in the United States.1, 2 The relationship between increasing levels of blood pressure (BP) and increasing risk for cardiovascular disease events and stroke is well established.3 However, while 64% of treated HTN patients have a BP <140/<90 mmHg, overall control rates for HTN in the adult population remain at approximately 44%.2 The 20% discrepancy in control rates between treated patients and the overall adult population reflects the fact that approximately 30% of patients are unaware of their HTN and that a substantial proportion of aware patients remain untreated. Historically, efforts to improve the recognition, treatment, and control of HTN have appropriately focused on the outpatient setting. However, programs to extend screening for HTN outside the clinic into the community, schools, and even dentists' offices have been around for some time.49

The potential also exists to improve the recognition, treatment, and control of HTN by focusing on hospitalized patients. Hospitalization is common in the U.S. with almost 35 million acute hospitalizations and more than 45,000 inpatient surgical procedures in 2006.10 Inpatient populations have increased in age and comorbidity over the past 3 decades whereas lengths of stay and continuity of care between the inpatient and outpatient arenas have diminished.10, 11 Multiple prior studies examining BP in different settings have noted that average BP among hospitalized patients is not systematically higher than that of outpatients.1214 Thus, patients with persistently elevated BP in the inpatient setting without mitigating factors may have HTN that will persist after hospital discharge. However, little information is available regarding the actual prevalence of HTN in the inpatient population and care patterns for inpatient HTN. Therefore, we performed a systematic review of the English‐language medical literature in order to describe the epidemiology of HTN observed in the inpatient setting.

Methods

Our search strategy was designed to identify randomized‐controlled trials, meta‐analyses, and observational studies that: (1) reported estimates of the prevalence of HTN in the inpatient setting, and (2) used HTN diagnosis or treatment as a primary focus. We performed an extensive review of the peer‐reviewed, English language medical literature in MEDLINE using a predetermined search algorithm. Search terms included HTN[Mesh] or BP[Mesh]. These results were cross‐referenced with the following search terms: Inpatients[Mesh] or Hospitalization[title/abstract] or Hospitalized[title/abstract]. Articles were further narrowed using the following terms: Prevalence[Mesh] or Epidemiology[Mesh] or Treatment[title/abstract] or Management [title/abstract]. Limits employed included limiting to humans and to adults 19 years‐of‐age and older. Studies published prior to 1976 were excluded because 1976 was the first year that the Joint National Committee on Prevention, Detection, Evaluation, and Treatment of High BP published consensus guidelines for the diagnosis and management of HTN. We also excluded randomized, controlled trials that recorded measures of inpatient BP but whose focus was not HTN, because such trials would not answer the primary epidemiologic question of this review. We did include trials focused on subspecialty populations for which the diagnosis and inpatient management of HTN were key outcomes.

Next, the bibliographies of reviewed studies were investigated for additional relevant reports. Abstracts from the American Heart Association (AHA) were reviewed for the past 15 years for reports that were presented but not subsequently published and available in MEDLINE. We also searched for articles using the online Google search engine. One author (RNA) performed the preliminary MEDLINE search and abstract review with the assistance of a reference librarian (LC), and a second author (BME) also reviewed full‐text articles for potential inclusion. Ultimate decision for study inclusion was reached through discussion among authors. Finally, a list of potential articles was submitted to 2 experts in this field of study to determine whether other reports met our inclusion criteria for this systematic review but were overlooked.

Results

Search Results

The initial MEDLINE search algorithm yielded a total of 826 articles. After title and abstract review, 41 full‐text articles were obtained for detailed review, and 5 met criteria for inclusion. Three additional articles were discovered through searching the bibliographies of the included studies. No AHA abstracts addressed this subject area. Experts were not aware of any additional studies. One article was located using a Google search. In all, 9 articles were deemed suitable for inclusion in this review. Search results at each stage are depicted in Figure 1.

Figure 1

Although antiretroviral therapy for human immunodeficiency virus (HIV)‐infected patients reduces viral load dramatically and improves immune function, some patients experience a clinical deterioration within the first few months of therapy because of an exuberant and dysregulated immune responsethe immune reconstitution inflammatory syndrome (IRIS). The exaggerated immune response associated with this syndrome can be stimulated by either antigens from infectious agents (typically a mycobacterium or cryptococcus) or from autoantigens, giving rise to a heterogeneous range of clinical manifestations.1 IRIS may present as an inflammatory reaction that unmasks a previously untreated infection or as a paradoxical worsening of an infection that is being treated appropriately. Although most cases of IRIS are mild and self‐limited, some patients require aggressive treatment.1

Case Report

A 53‐year‐old man was evaluated for a 5‐day history of intermittent chest pain. He had been diagnosed with HIV/acquired immune deficiency syndrome (AIDS) 11 years ago but he had not been compliant with therapy. Seven years earlier he had been treated for 9 months with isoniazid for a positive tuberculin skin test. Three months before admission, he developed methicillin‐resistant Staphylococcus aureus skin abscesses and was found to have a CD4 count of 1/L and a HIV viral load of over 400,000 copies/mL. He finished a course of vancomycin, and was started on lopinavir, ritonavir, abacavir, lamivudine, and zidovudine. Five days before admission, he was evaluated in the emergency department for intermittent chest pain and described using cocaine. There was only J‐point elevation on the electrocardiogram (Figure 1A), serial cardiac enzymes were negative, and he was discharged home. However, despite discontinuation of cocaine use, his chest pain worsened, became pleuritic, and was associated with dyspnea, which prompted this admission. Physical examination was remarkable only for tachycardia, although the electrocardiogram now revealed diffuse ST segment (ST) elevation and PR segment (PR) depression, consistent with acute pericarditis (Figure 1B). Serial cardiac enzymes, viral studies, and bacterial, fungal, and mycobacterial blood cultures were negative. His CD4 count was 16/L, and the HIV viral load was 870 copies/mL.

Figure 1

The patient was treated with high‐dose ibuprofen and colchicine, but mild chest pain and electrocardiogram changes persisted, and he developed a friction rub. A chest computed tomography (CT) scan was negative for pulmonary embolism and revealed no significant intra‐thoracic pathology, except for a moderate pericardial effusion that was confirmed by transthoracic echocardiogram (Figure 2A). There was no echocardiographic evidence of tamponade. He underwent thoracoscopic pericardial and mediastinal lymph node biopsy, along with drainage of the pericardial effusion. Pericardial biopsy showed acute on chronic inflammation consistent with pericarditis (Figure 2B) and culture was positive for Mycobacterium Avium Complex (MAC). He was treated with clarithromycin, ethambutol, and prednisone, and his antiretroviral medications were continued. At 2, 6, and 12 months follow‐up, he was asymptomatic, the electrocardiogram had normalized (Figure 1C), and the echocardiogram showed no effusion or evidence of pericardial constriction.

Figure 2

Discussion

This case demonstrates a unique manifestation of the IRIS associated with MAC infection, which more typically presents as peripheral, pulmonary, or intra‐abdominal lymphadenopathy.2, 3 It usually responds to MAC therapy, although intra‐abdominal disease portends a poor prognosis.3, 4 This patient has two significant risk factors for the development of IRIS: low CD4 count at the time of antiretroviral therapy and rapid viral clearance.5, 6 While his CD4 count response is lower than expected for IRIS, previous studies have shown that functional immune recovery usually precedes quantitative CD4 count recovery, and that IRIS could happen at low CD4 count.1, 7 Finally, we believe that the use of corticosteroids accounted for his rapid clinical improvement and favorable long‐term outcome, consistent with previous experience of corticosteroid use in MAC‐associated IRIS.3, 4 To our knowledge, this is the first reported case of MAC‐associated IRIS presenting as isolated acute pericarditis and pericardial effusion. In conclusion, our case illustrates that IRIS can present as an abnormal immune response to an opportunistic infection in an unusual location. Clinicians must be aware that after starting antiretroviral therapy, new symptoms, including chest pain, might represent 1 of the IRISs, and that corticosteroids might be beneficial when inflammation is severe.

Although antiretroviral therapy for human immunodeficiency virus (HIV)‐infected patients reduces viral load dramatically and improves immune function, some patients experience a clinical deterioration within the first few months of therapy because of an exuberant and dysregulated immune responsethe immune reconstitution inflammatory syndrome (IRIS). The exaggerated immune response associated with this syndrome can be stimulated by either antigens from infectious agents (typically a mycobacterium or cryptococcus) or from autoantigens, giving rise to a heterogeneous range of clinical manifestations.1 IRIS may present as an inflammatory reaction that unmasks a previously untreated infection or as a paradoxical worsening of an infection that is being treated appropriately. Although most cases of IRIS are mild and self‐limited, some patients require aggressive treatment.1

Case Report

A 53‐year‐old man was evaluated for a 5‐day history of intermittent chest pain. He had been diagnosed with HIV/acquired immune deficiency syndrome (AIDS) 11 years ago but he had not been compliant with therapy. Seven years earlier he had been treated for 9 months with isoniazid for a positive tuberculin skin test. Three months before admission, he developed methicillin‐resistant Staphylococcus aureus skin abscesses and was found to have a CD4 count of 1/L and a HIV viral load of over 400,000 copies/mL. He finished a course of vancomycin, and was started on lopinavir, ritonavir, abacavir, lamivudine, and zidovudine. Five days before admission, he was evaluated in the emergency department for intermittent chest pain and described using cocaine. There was only J‐point elevation on the electrocardiogram (Figure 1A), serial cardiac enzymes were negative, and he was discharged home. However, despite discontinuation of cocaine use, his chest pain worsened, became pleuritic, and was associated with dyspnea, which prompted this admission. Physical examination was remarkable only for tachycardia, although the electrocardiogram now revealed diffuse ST segment (ST) elevation and PR segment (PR) depression, consistent with acute pericarditis (Figure 1B). Serial cardiac enzymes, viral studies, and bacterial, fungal, and mycobacterial blood cultures were negative. His CD4 count was 16/L, and the HIV viral load was 870 copies/mL.

Figure 1

The patient was treated with high‐dose ibuprofen and colchicine, but mild chest pain and electrocardiogram changes persisted, and he developed a friction rub. A chest computed tomography (CT) scan was negative for pulmonary embolism and revealed no significant intra‐thoracic pathology, except for a moderate pericardial effusion that was confirmed by transthoracic echocardiogram (Figure 2A). There was no echocardiographic evidence of tamponade. He underwent thoracoscopic pericardial and mediastinal lymph node biopsy, along with drainage of the pericardial effusion. Pericardial biopsy showed acute on chronic inflammation consistent with pericarditis (Figure 2B) and culture was positive for Mycobacterium Avium Complex (MAC). He was treated with clarithromycin, ethambutol, and prednisone, and his antiretroviral medications were continued. At 2, 6, and 12 months follow‐up, he was asymptomatic, the electrocardiogram had normalized (Figure 1C), and the echocardiogram showed no effusion or evidence of pericardial constriction.

Figure 2

Discussion

This case demonstrates a unique manifestation of the IRIS associated with MAC infection, which more typically presents as peripheral, pulmonary, or intra‐abdominal lymphadenopathy.2, 3 It usually responds to MAC therapy, although intra‐abdominal disease portends a poor prognosis.3, 4 This patient has two significant risk factors for the development of IRIS: low CD4 count at the time of antiretroviral therapy and rapid viral clearance.5, 6 While his CD4 count response is lower than expected for IRIS, previous studies have shown that functional immune recovery usually precedes quantitative CD4 count recovery, and that IRIS could happen at low CD4 count.1, 7 Finally, we believe that the use of corticosteroids accounted for his rapid clinical improvement and favorable long‐term outcome, consistent with previous experience of corticosteroid use in MAC‐associated IRIS.3, 4 To our knowledge, this is the first reported case of MAC‐associated IRIS presenting as isolated acute pericarditis and pericardial effusion. In conclusion, our case illustrates that IRIS can present as an abnormal immune response to an opportunistic infection in an unusual location. Clinicians must be aware that after starting antiretroviral therapy, new symptoms, including chest pain, might represent 1 of the IRISs, and that corticosteroids might be beneficial when inflammation is severe.

Although antiretroviral therapy for human immunodeficiency virus (HIV)‐infected patients reduces viral load dramatically and improves immune function, some patients experience a clinical deterioration within the first few months of therapy because of an exuberant and dysregulated immune responsethe immune reconstitution inflammatory syndrome (IRIS). The exaggerated immune response associated with this syndrome can be stimulated by either antigens from infectious agents (typically a mycobacterium or cryptococcus) or from autoantigens, giving rise to a heterogeneous range of clinical manifestations.1 IRIS may present as an inflammatory reaction that unmasks a previously untreated infection or as a paradoxical worsening of an infection that is being treated appropriately. Although most cases of IRIS are mild and self‐limited, some patients require aggressive treatment.1

Case Report

A 53‐year‐old man was evaluated for a 5‐day history of intermittent chest pain. He had been diagnosed with HIV/acquired immune deficiency syndrome (AIDS) 11 years ago but he had not been compliant with therapy. Seven years earlier he had been treated for 9 months with isoniazid for a positive tuberculin skin test. Three months before admission, he developed methicillin‐resistant Staphylococcus aureus skin abscesses and was found to have a CD4 count of 1/L and a HIV viral load of over 400,000 copies/mL. He finished a course of vancomycin, and was started on lopinavir, ritonavir, abacavir, lamivudine, and zidovudine. Five days before admission, he was evaluated in the emergency department for intermittent chest pain and described using cocaine. There was only J‐point elevation on the electrocardiogram (Figure 1A), serial cardiac enzymes were negative, and he was discharged home. However, despite discontinuation of cocaine use, his chest pain worsened, became pleuritic, and was associated with dyspnea, which prompted this admission. Physical examination was remarkable only for tachycardia, although the electrocardiogram now revealed diffuse ST segment (ST) elevation and PR segment (PR) depression, consistent with acute pericarditis (Figure 1B). Serial cardiac enzymes, viral studies, and bacterial, fungal, and mycobacterial blood cultures were negative. His CD4 count was 16/L, and the HIV viral load was 870 copies/mL.

Figure 1

The patient was treated with high‐dose ibuprofen and colchicine, but mild chest pain and electrocardiogram changes persisted, and he developed a friction rub. A chest computed tomography (CT) scan was negative for pulmonary embolism and revealed no significant intra‐thoracic pathology, except for a moderate pericardial effusion that was confirmed by transthoracic echocardiogram (Figure 2A). There was no echocardiographic evidence of tamponade. He underwent thoracoscopic pericardial and mediastinal lymph node biopsy, along with drainage of the pericardial effusion. Pericardial biopsy showed acute on chronic inflammation consistent with pericarditis (Figure 2B) and culture was positive for Mycobacterium Avium Complex (MAC). He was treated with clarithromycin, ethambutol, and prednisone, and his antiretroviral medications were continued. At 2, 6, and 12 months follow‐up, he was asymptomatic, the electrocardiogram had normalized (Figure 1C), and the echocardiogram showed no effusion or evidence of pericardial constriction.

Figure 2

Discussion

This case demonstrates a unique manifestation of the IRIS associated with MAC infection, which more typically presents as peripheral, pulmonary, or intra‐abdominal lymphadenopathy.2, 3 It usually responds to MAC therapy, although intra‐abdominal disease portends a poor prognosis.3, 4 This patient has two significant risk factors for the development of IRIS: low CD4 count at the time of antiretroviral therapy and rapid viral clearance.5, 6 While his CD4 count response is lower than expected for IRIS, previous studies have shown that functional immune recovery usually precedes quantitative CD4 count recovery, and that IRIS could happen at low CD4 count.1, 7 Finally, we believe that the use of corticosteroids accounted for his rapid clinical improvement and favorable long‐term outcome, consistent with previous experience of corticosteroid use in MAC‐associated IRIS.3, 4 To our knowledge, this is the first reported case of MAC‐associated IRIS presenting as isolated acute pericarditis and pericardial effusion. In conclusion, our case illustrates that IRIS can present as an abnormal immune response to an opportunistic infection in an unusual location. Clinicians must be aware that after starting antiretroviral therapy, new symptoms, including chest pain, might represent 1 of the IRISs, and that corticosteroids might be beneficial when inflammation is severe.

Heart failure (HF) carries a high rate of morbidity and mortality.1 In the past decades, the incidence of HF and HF‐related hospital admissions has risen continuously, posing a formidable healthcare and economic burden.24 Extensive evidence has shown that treatment of angiotensin converting enzyme inhibitors (ACEi) and angiotensin receptor blockers (ARBs) reduces morbidity and mortality and improves quality of life in patients with HF and left ventricular systolic dysfunction (LVSD).57 Consequently, ACEi/ARB utilization in HF and LVSD has become one of the practice guidelines8 and a nationally required quality performance measure by The Joint Commission (TJC, formally known as JCAHO) and Centers for Medicare & Medicaid Services (CMS).

Despite the well‐demonstrated salutary effects and clear guidelines, under‐utilization of ACEi/ARB for HF patients has repeatedly been demonstrated.911 There seems to be a lasting quality chasm between the lifesaving therapy and its utilization in our practice.12 This chasm is illustrated by a recent study of 54,453 U.S. patients who were hospitalized for HF and discharged alive, showing that use of proven therapies such as ACEi/ARBs remains far from sufficient (48% for the total HF patients and 52% for HF patients with prior myocardial infarction).11 In large academic hospital centers, the ACEi/ARB utilization for HF patients has averaged between 8388%.13

Strides have been made to bridge the chasm;1419 however, these efforts have been impeded by complex and multifaceted problems. One of these problems is the sheer number of HF patients. In the current economic environment, traditional methods of pouring in more resources are unsustainable. Yet, the majority of quality improvement methods tried thus far involve increasing manpower, intensifying the delivery of staff and patient education, applying multiprong intervening systems, and prolonging the duration of the patients' hospital stay.1422

Although most of these measures achieve their intended goals, ongoing cost is required and the sustainability remains doubtful. Health information technology (IT) is emerging as a promising tool for improving care quality and containing cost.23 The electronic medical record (EMR) system at Mayo Clinic Rochester is built upon an IT patient record platform of Last Word (formerly a product of IDX, now General Electric, Fairfield, Connecticut) and has the capability of receiving vast input from databases in each department in our institution. In recent years, Mayo Clinic also has developed an IT hospital rule (algorithm)‐based system (HRBS) for comprehensive, multidisciplinary patient monitoring and cost containment (detailed in ref. 24). Pharmaceutical Care (P‐Care) is 1 of the 6 subsystems under HRBS. P‐care has been used primarily by inpatient pharmacists to detect situations where there is a high probability of suboptimal medication prescribing and where intervention by a pharmacist may be beneficial.

The primary goal of this project was to improve ACEi/ARB adherence for inpatients in a manner that would be sustainable. We intended to incorporate the existing features of our EMR as well as modify and utilize the P‐Care system to create a model that would improve ACEi/ARB adherence and work well with work‐flows of inpatient pharmacists and patient‐care teams.

Methods

Implementation of the Intervention

The intervention Model included three components: a computer‐based daily screening program developed from the existing P‐Care rule,24 inpatient pharmacists, and inpatient care teams. The interventional algorithm is illustrated in Figure 1. The computer‐based screening program that retrieved patients' LVEF data from EMR was up and running by the first quarter of 2006. A major attribute of the existing IT systems at Mayo Clinic has been that, however enormous, the data (input daily from diverse sources within the institution) are entered in a discrete, searchable and extractable format, which is critical for the data utilization. In the second quarter of 2006, we began an intense Plan‐Do‐Study‐Act (PDSA) cycle through multidisciplinary teamwork. To monitor e‐flagging efficiency, we randomly selected five units, manually monitored the number of patients who failed ACEi/ARB adherence and compared the number with that generated by the screening program. We found that the capturing rate was 100%.

Figure 1

Several problems were encountered with the model's operating process during implementation. The flagged list generated by the screening program was examined first by a pharmacist who then prepared a written note, indicating the deficiency along with a concise version of the guidelines. This note was placed in the patients' chart. Alternatively, the pharmacist might notify the patient‐care team by phone or in person during the teams' on their rounds.

However, notes were sometimes lost or overlooked, and verbal communications were inconsistent. In addition, the pharmacists were sometimes unsure whether, under certain clinical conditions (eg, serum creatinine elevation amidst diuresis), a HF patient should receive ACEis/ARBs.

Occasionally, care teams objected to the calls and viewed visits by pharmacists as interruption of their work flow resulting in awkward, and sometimes ineffective communications. Thus, the model seemed to have generated sizable extra work for the pharmacists and there was a notable time‐lag between the generation of the flag‐list and the successful delivery of the message.

To solve these problems, with the advantage of a programmer on the team, we created an electronic message (e‐message) delivery function within our EMR. When a patient‐care physician accesses the patient's information in EMR, a prompt indicating e‐message would appear. This modification allowed pharmacists' verification and an e‐message to be semiautomatically delivered to the patient‐care team. If the problem (non‐compliance to ACEi/ARB guidelines) was not addressed within 24 hours after the e‐message delivery, a pharmacist would then contact the team by phone or face‐to‐face. Additionally, an inpatient nephrologist was made available to answer any clinical questions that the pharmacists might have. We found that with these modifications the vast majority of the flags were corrected within 24 hours and pharmacists' workload was markedly reduced. After several initial communications between pharmacists and the nephrologist, the input by the nephrologist became minimal as pharmacists grew more accustomed to the majority of case scenarios.

Through such PDSA cycles, the operating process improved progressively. By March 2007, the implementation was complete and the model ran smoothly to the satisfaction of the team and other stakeholders.

Results

Percentage ACEi/ARB Adherence With the Intervention

The mean percentages of ACEi/ARB adherence in the periods before, during, and after the model implantation were 88.4%, 88.8%, and 97.6% respectively. Significant differences were detected between the three periods by Pearson's chi‐square test (P < 0.001). Fisher's Exact Test was used for comparing the periods before and after (P < 0.001, Figure 2A) and during and after (P < 0.001). Figure 2B shows the quarterly sensors of the adherence rate. Notably, after the implementation, the compliance rate remained high and the variations lessened.

Figure 2

Discussion

The results of this study show that the computer‐based quality improvement tool was associated with improved adherence to the ACEi/ARB guidelines for patients with LVSD/HF. This was accomplished without the need for additional, ongoing expenses in a system fitting our EMR capabilities and work flow.

Specific studies on the improvement of ACEi/ARB utilization for LVSD patients are limited.16, 21 One randomized controlled trial evaluated an inpatient HF intervention without a post‐discharge care plan.21 The intervention included inpatient guidelines for the use of ACEi, echocardiogram, daily weights and a consultative service provided by a nurse care manager and cardiologist. The consultative service included patient education, treatment recommendations, and discharge planning. This intervention significantly improved ACEi use at discharge.

Another randomized controlled study of 98 patients showed that compared to routine care, those who received multidisciplinary care (inpatient and outpatient education and intense telephone and clinic follow‐up), ACEi usage was maximized and re‐hospitalization and HFrelated death was significantly reduced at three months.16 Although effective, such interventions require substantial ongoing cost and sustainability is again called into question. Our initiative is unique in that incorporating a computer‐based semiautomatic system into the care‐delivery process has enhanced care quality without incurring ongoing extra cost (we have neither hired extra personnel nor created a heavier work burden for pharmacists and patientcare teams, as the model has been diffused into their daily routine) thus maximizing its longterm sustainability.

Notwithstanding the positive aspects, this study has several limitations. First, it is not a randomized, controlled trial, and unidentified external factors may have had some influence. However, in the examination of all potential external effects, we could not identify any factor that would have the capacity to substantially and consistently influence the results. Second, prepost study design is less ideal than randomized, controlled trials on the study design hierarchy. However, given the unsatisfactory adherence rate, anticipated positive effects with the model, and the pressing need for improving the adherence, a randomized trial was not an option at that juncture. Third, we could not precisely compare the difference in the awareness of ACEi/ARB guidelines among different classes of trainees during the study period. We did have a one‐time online, non‐mandatory education program for all providers. However, new trainees rotated in and‐ out on a monthly basis. This factor is unlikely to have caused a sustained change. Fourth, we did not have the outcome data for patients in whom HF was their secondary admission diagnosis. These patients were equally flagged by the model, and their ACEi/ARB status, when flagged, was obliged to be corrected. We suspect that these patients most likely benefited even more by the model because they were likely in a compensated state of HF, and the care‐teams tended to be more focused on their primary issue, leaving room for overlooking LVSD‐related issues.

Finally, we report the outcomes in the first 21 months after the full implementation of the model. We still need to monitor the long‐term outcome, although a reasonable length of time has elapsed. There has been no sign of decay in its effectiveness and we have no compelling reason to anticipate a significant regression.

Under ideal conditions, the outcome should consistently be 100% based on the design. In reality the adherence had been oscillating with an average of 97%. We noted two main scenarios that had contributed to this outcome. First, some LVSD/HF patients were taken off ACEi/ARB temporarily before discharge because of worsening pre‐renal azotemia with diuresis. They were discharged off ACEi/ARB with a plan to resume it. These patients would not have been labeled as ACEi/ARB‐intolerant but were classified as those without meeting the guidelines. Second, some patients had their echocardiogram on the same day or within 24 hours of discharge. A fraction of them had LVEF < 40%, but ACEi/ARB had not been initiated before discharge.

The rising volume of patients with increasing age and co‐morbidities, combined with constraints in healthcare resources, compels us to explore high‐efficiency care‐delivery models. Although computerized technology is well understood and readily available, the challenge we face is how to fully utilize the technology. A recent study shows that the improvement of IT infrastructure and research on implementation are interdependent and both can be translated to better patient care.25 Our experience serves as another example demonstrating that, when carefully conceived and properly executed, computer‐based care‐delivery prompts can be highly efficient and effective, suitable for large hospital settings with a heavy patient load like ours.

Moreover, because of the availability of basic IT platforms, similar algorithm‐based model systems can foreseeably be adopted by hospitals of comparable size and structure and also be applied to other care‐delivery settings including out‐patient clinics, chronic dialysis units and various long‐term care facilities.

Developing efficient, IT‐based quality improvement tools that facilitate the application of evidence‐based care and improve quality without significant additional resources is imperative in today's economic climate. Strategies such as our e‐messaging intervention with ACEi and ARB demonstrate sustainable improvement, can be applied to other conditions, and should be vigorously pursued.

Heart failure (HF) carries a high rate of morbidity and mortality.1 In the past decades, the incidence of HF and HF‐related hospital admissions has risen continuously, posing a formidable healthcare and economic burden.24 Extensive evidence has shown that treatment of angiotensin converting enzyme inhibitors (ACEi) and angiotensin receptor blockers (ARBs) reduces morbidity and mortality and improves quality of life in patients with HF and left ventricular systolic dysfunction (LVSD).57 Consequently, ACEi/ARB utilization in HF and LVSD has become one of the practice guidelines8 and a nationally required quality performance measure by The Joint Commission (TJC, formally known as JCAHO) and Centers for Medicare & Medicaid Services (CMS).

Despite the well‐demonstrated salutary effects and clear guidelines, under‐utilization of ACEi/ARB for HF patients has repeatedly been demonstrated.911 There seems to be a lasting quality chasm between the lifesaving therapy and its utilization in our practice.12 This chasm is illustrated by a recent study of 54,453 U.S. patients who were hospitalized for HF and discharged alive, showing that use of proven therapies such as ACEi/ARBs remains far from sufficient (48% for the total HF patients and 52% for HF patients with prior myocardial infarction).11 In large academic hospital centers, the ACEi/ARB utilization for HF patients has averaged between 8388%.13

Strides have been made to bridge the chasm;1419 however, these efforts have been impeded by complex and multifaceted problems. One of these problems is the sheer number of HF patients. In the current economic environment, traditional methods of pouring in more resources are unsustainable. Yet, the majority of quality improvement methods tried thus far involve increasing manpower, intensifying the delivery of staff and patient education, applying multiprong intervening systems, and prolonging the duration of the patients' hospital stay.1422

Although most of these measures achieve their intended goals, ongoing cost is required and the sustainability remains doubtful. Health information technology (IT) is emerging as a promising tool for improving care quality and containing cost.23 The electronic medical record (EMR) system at Mayo Clinic Rochester is built upon an IT patient record platform of Last Word (formerly a product of IDX, now General Electric, Fairfield, Connecticut) and has the capability of receiving vast input from databases in each department in our institution. In recent years, Mayo Clinic also has developed an IT hospital rule (algorithm)‐based system (HRBS) for comprehensive, multidisciplinary patient monitoring and cost containment (detailed in ref. 24). Pharmaceutical Care (P‐Care) is 1 of the 6 subsystems under HRBS. P‐care has been used primarily by inpatient pharmacists to detect situations where there is a high probability of suboptimal medication prescribing and where intervention by a pharmacist may be beneficial.

The primary goal of this project was to improve ACEi/ARB adherence for inpatients in a manner that would be sustainable. We intended to incorporate the existing features of our EMR as well as modify and utilize the P‐Care system to create a model that would improve ACEi/ARB adherence and work well with work‐flows of inpatient pharmacists and patient‐care teams.

Methods

Implementation of the Intervention

The intervention Model included three components: a computer‐based daily screening program developed from the existing P‐Care rule,24 inpatient pharmacists, and inpatient care teams. The interventional algorithm is illustrated in Figure 1. The computer‐based screening program that retrieved patients' LVEF data from EMR was up and running by the first quarter of 2006. A major attribute of the existing IT systems at Mayo Clinic has been that, however enormous, the data (input daily from diverse sources within the institution) are entered in a discrete, searchable and extractable format, which is critical for the data utilization. In the second quarter of 2006, we began an intense Plan‐Do‐Study‐Act (PDSA) cycle through multidisciplinary teamwork. To monitor e‐flagging efficiency, we randomly selected five units, manually monitored the number of patients who failed ACEi/ARB adherence and compared the number with that generated by the screening program. We found that the capturing rate was 100%.

Figure 1

Several problems were encountered with the model's operating process during implementation. The flagged list generated by the screening program was examined first by a pharmacist who then prepared a written note, indicating the deficiency along with a concise version of the guidelines. This note was placed in the patients' chart. Alternatively, the pharmacist might notify the patient‐care team by phone or in person during the teams' on their rounds.

However, notes were sometimes lost or overlooked, and verbal communications were inconsistent. In addition, the pharmacists were sometimes unsure whether, under certain clinical conditions (eg, serum creatinine elevation amidst diuresis), a HF patient should receive ACEis/ARBs.

Occasionally, care teams objected to the calls and viewed visits by pharmacists as interruption of their work flow resulting in awkward, and sometimes ineffective communications. Thus, the model seemed to have generated sizable extra work for the pharmacists and there was a notable time‐lag between the generation of the flag‐list and the successful delivery of the message.

To solve these problems, with the advantage of a programmer on the team, we created an electronic message (e‐message) delivery function within our EMR. When a patient‐care physician accesses the patient's information in EMR, a prompt indicating e‐message would appear. This modification allowed pharmacists' verification and an e‐message to be semiautomatically delivered to the patient‐care team. If the problem (non‐compliance to ACEi/ARB guidelines) was not addressed within 24 hours after the e‐message delivery, a pharmacist would then contact the team by phone or face‐to‐face. Additionally, an inpatient nephrologist was made available to answer any clinical questions that the pharmacists might have. We found that with these modifications the vast majority of the flags were corrected within 24 hours and pharmacists' workload was markedly reduced. After several initial communications between pharmacists and the nephrologist, the input by the nephrologist became minimal as pharmacists grew more accustomed to the majority of case scenarios.

Through such PDSA cycles, the operating process improved progressively. By March 2007, the implementation was complete and the model ran smoothly to the satisfaction of the team and other stakeholders.

Results

Percentage ACEi/ARB Adherence With the Intervention

The mean percentages of ACEi/ARB adherence in the periods before, during, and after the model implantation were 88.4%, 88.8%, and 97.6% respectively. Significant differences were detected between the three periods by Pearson's chi‐square test (P < 0.001). Fisher's Exact Test was used for comparing the periods before and after (P < 0.001, Figure 2A) and during and after (P < 0.001). Figure 2B shows the quarterly sensors of the adherence rate. Notably, after the implementation, the compliance rate remained high and the variations lessened.

Figure 2

Discussion

The results of this study show that the computer‐based quality improvement tool was associated with improved adherence to the ACEi/ARB guidelines for patients with LVSD/HF. This was accomplished without the need for additional, ongoing expenses in a system fitting our EMR capabilities and work flow.

Specific studies on the improvement of ACEi/ARB utilization for LVSD patients are limited.16, 21 One randomized controlled trial evaluated an inpatient HF intervention without a post‐discharge care plan.21 The intervention included inpatient guidelines for the use of ACEi, echocardiogram, daily weights and a consultative service provided by a nurse care manager and cardiologist. The consultative service included patient education, treatment recommendations, and discharge planning. This intervention significantly improved ACEi use at discharge.

Another randomized controlled study of 98 patients showed that compared to routine care, those who received multidisciplinary care (inpatient and outpatient education and intense telephone and clinic follow‐up), ACEi usage was maximized and re‐hospitalization and HFrelated death was significantly reduced at three months.16 Although effective, such interventions require substantial ongoing cost and sustainability is again called into question. Our initiative is unique in that incorporating a computer‐based semiautomatic system into the care‐delivery process has enhanced care quality without incurring ongoing extra cost (we have neither hired extra personnel nor created a heavier work burden for pharmacists and patientcare teams, as the model has been diffused into their daily routine) thus maximizing its longterm sustainability.

Notwithstanding the positive aspects, this study has several limitations. First, it is not a randomized, controlled trial, and unidentified external factors may have had some influence. However, in the examination of all potential external effects, we could not identify any factor that would have the capacity to substantially and consistently influence the results. Second, prepost study design is less ideal than randomized, controlled trials on the study design hierarchy. However, given the unsatisfactory adherence rate, anticipated positive effects with the model, and the pressing need for improving the adherence, a randomized trial was not an option at that juncture. Third, we could not precisely compare the difference in the awareness of ACEi/ARB guidelines among different classes of trainees during the study period. We did have a one‐time online, non‐mandatory education program for all providers. However, new trainees rotated in and‐ out on a monthly basis. This factor is unlikely to have caused a sustained change. Fourth, we did not have the outcome data for patients in whom HF was their secondary admission diagnosis. These patients were equally flagged by the model, and their ACEi/ARB status, when flagged, was obliged to be corrected. We suspect that these patients most likely benefited even more by the model because they were likely in a compensated state of HF, and the care‐teams tended to be more focused on their primary issue, leaving room for overlooking LVSD‐related issues.

Finally, we report the outcomes in the first 21 months after the full implementation of the model. We still need to monitor the long‐term outcome, although a reasonable length of time has elapsed. There has been no sign of decay in its effectiveness and we have no compelling reason to anticipate a significant regression.

Under ideal conditions, the outcome should consistently be 100% based on the design. In reality the adherence had been oscillating with an average of 97%. We noted two main scenarios that had contributed to this outcome. First, some LVSD/HF patients were taken off ACEi/ARB temporarily before discharge because of worsening pre‐renal azotemia with diuresis. They were discharged off ACEi/ARB with a plan to resume it. These patients would not have been labeled as ACEi/ARB‐intolerant but were classified as those without meeting the guidelines. Second, some patients had their echocardiogram on the same day or within 24 hours of discharge. A fraction of them had LVEF < 40%, but ACEi/ARB had not been initiated before discharge.

The rising volume of patients with increasing age and co‐morbidities, combined with constraints in healthcare resources, compels us to explore high‐efficiency care‐delivery models. Although computerized technology is well understood and readily available, the challenge we face is how to fully utilize the technology. A recent study shows that the improvement of IT infrastructure and research on implementation are interdependent and both can be translated to better patient care.25 Our experience serves as another example demonstrating that, when carefully conceived and properly executed, computer‐based care‐delivery prompts can be highly efficient and effective, suitable for large hospital settings with a heavy patient load like ours.

Moreover, because of the availability of basic IT platforms, similar algorithm‐based model systems can foreseeably be adopted by hospitals of comparable size and structure and also be applied to other care‐delivery settings including out‐patient clinics, chronic dialysis units and various long‐term care facilities.

Developing efficient, IT‐based quality improvement tools that facilitate the application of evidence‐based care and improve quality without significant additional resources is imperative in today's economic climate. Strategies such as our e‐messaging intervention with ACEi and ARB demonstrate sustainable improvement, can be applied to other conditions, and should be vigorously pursued.

Heart failure (HF) carries a high rate of morbidity and mortality.1 In the past decades, the incidence of HF and HF‐related hospital admissions has risen continuously, posing a formidable healthcare and economic burden.24 Extensive evidence has shown that treatment of angiotensin converting enzyme inhibitors (ACEi) and angiotensin receptor blockers (ARBs) reduces morbidity and mortality and improves quality of life in patients with HF and left ventricular systolic dysfunction (LVSD).57 Consequently, ACEi/ARB utilization in HF and LVSD has become one of the practice guidelines8 and a nationally required quality performance measure by The Joint Commission (TJC, formally known as JCAHO) and Centers for Medicare & Medicaid Services (CMS).

Despite the well‐demonstrated salutary effects and clear guidelines, under‐utilization of ACEi/ARB for HF patients has repeatedly been demonstrated.911 There seems to be a lasting quality chasm between the lifesaving therapy and its utilization in our practice.12 This chasm is illustrated by a recent study of 54,453 U.S. patients who were hospitalized for HF and discharged alive, showing that use of proven therapies such as ACEi/ARBs remains far from sufficient (48% for the total HF patients and 52% for HF patients with prior myocardial infarction).11 In large academic hospital centers, the ACEi/ARB utilization for HF patients has averaged between 8388%.13

Strides have been made to bridge the chasm;1419 however, these efforts have been impeded by complex and multifaceted problems. One of these problems is the sheer number of HF patients. In the current economic environment, traditional methods of pouring in more resources are unsustainable. Yet, the majority of quality improvement methods tried thus far involve increasing manpower, intensifying the delivery of staff and patient education, applying multiprong intervening systems, and prolonging the duration of the patients' hospital stay.1422

Although most of these measures achieve their intended goals, ongoing cost is required and the sustainability remains doubtful. Health information technology (IT) is emerging as a promising tool for improving care quality and containing cost.23 The electronic medical record (EMR) system at Mayo Clinic Rochester is built upon an IT patient record platform of Last Word (formerly a product of IDX, now General Electric, Fairfield, Connecticut) and has the capability of receiving vast input from databases in each department in our institution. In recent years, Mayo Clinic also has developed an IT hospital rule (algorithm)‐based system (HRBS) for comprehensive, multidisciplinary patient monitoring and cost containment (detailed in ref. 24). Pharmaceutical Care (P‐Care) is 1 of the 6 subsystems under HRBS. P‐care has been used primarily by inpatient pharmacists to detect situations where there is a high probability of suboptimal medication prescribing and where intervention by a pharmacist may be beneficial.

The primary goal of this project was to improve ACEi/ARB adherence for inpatients in a manner that would be sustainable. We intended to incorporate the existing features of our EMR as well as modify and utilize the P‐Care system to create a model that would improve ACEi/ARB adherence and work well with work‐flows of inpatient pharmacists and patient‐care teams.

Methods

Implementation of the Intervention

The intervention Model included three components: a computer‐based daily screening program developed from the existing P‐Care rule,24 inpatient pharmacists, and inpatient care teams. The interventional algorithm is illustrated in Figure 1. The computer‐based screening program that retrieved patients' LVEF data from EMR was up and running by the first quarter of 2006. A major attribute of the existing IT systems at Mayo Clinic has been that, however enormous, the data (input daily from diverse sources within the institution) are entered in a discrete, searchable and extractable format, which is critical for the data utilization. In the second quarter of 2006, we began an intense Plan‐Do‐Study‐Act (PDSA) cycle through multidisciplinary teamwork. To monitor e‐flagging efficiency, we randomly selected five units, manually monitored the number of patients who failed ACEi/ARB adherence and compared the number with that generated by the screening program. We found that the capturing rate was 100%.

Figure 1

Several problems were encountered with the model's operating process during implementation. The flagged list generated by the screening program was examined first by a pharmacist who then prepared a written note, indicating the deficiency along with a concise version of the guidelines. This note was placed in the patients' chart. Alternatively, the pharmacist might notify the patient‐care team by phone or in person during the teams' on their rounds.

However, notes were sometimes lost or overlooked, and verbal communications were inconsistent. In addition, the pharmacists were sometimes unsure whether, under certain clinical conditions (eg, serum creatinine elevation amidst diuresis), a HF patient should receive ACEis/ARBs.

Occasionally, care teams objected to the calls and viewed visits by pharmacists as interruption of their work flow resulting in awkward, and sometimes ineffective communications. Thus, the model seemed to have generated sizable extra work for the pharmacists and there was a notable time‐lag between the generation of the flag‐list and the successful delivery of the message.

To solve these problems, with the advantage of a programmer on the team, we created an electronic message (e‐message) delivery function within our EMR. When a patient‐care physician accesses the patient's information in EMR, a prompt indicating e‐message would appear. This modification allowed pharmacists' verification and an e‐message to be semiautomatically delivered to the patient‐care team. If the problem (non‐compliance to ACEi/ARB guidelines) was not addressed within 24 hours after the e‐message delivery, a pharmacist would then contact the team by phone or face‐to‐face. Additionally, an inpatient nephrologist was made available to answer any clinical questions that the pharmacists might have. We found that with these modifications the vast majority of the flags were corrected within 24 hours and pharmacists' workload was markedly reduced. After several initial communications between pharmacists and the nephrologist, the input by the nephrologist became minimal as pharmacists grew more accustomed to the majority of case scenarios.

Through such PDSA cycles, the operating process improved progressively. By March 2007, the implementation was complete and the model ran smoothly to the satisfaction of the team and other stakeholders.

Results

Percentage ACEi/ARB Adherence With the Intervention

The mean percentages of ACEi/ARB adherence in the periods before, during, and after the model implantation were 88.4%, 88.8%, and 97.6% respectively. Significant differences were detected between the three periods by Pearson's chi‐square test (P < 0.001). Fisher's Exact Test was used for comparing the periods before and after (P < 0.001, Figure 2A) and during and after (P < 0.001). Figure 2B shows the quarterly sensors of the adherence rate. Notably, after the implementation, the compliance rate remained high and the variations lessened.

Figure 2

Discussion

The results of this study show that the computer‐based quality improvement tool was associated with improved adherence to the ACEi/ARB guidelines for patients with LVSD/HF. This was accomplished without the need for additional, ongoing expenses in a system fitting our EMR capabilities and work flow.

Specific studies on the improvement of ACEi/ARB utilization for LVSD patients are limited.16, 21 One randomized controlled trial evaluated an inpatient HF intervention without a post‐discharge care plan.21 The intervention included inpatient guidelines for the use of ACEi, echocardiogram, daily weights and a consultative service provided by a nurse care manager and cardiologist. The consultative service included patient education, treatment recommendations, and discharge planning. This intervention significantly improved ACEi use at discharge.

Another randomized controlled study of 98 patients showed that compared to routine care, those who received multidisciplinary care (inpatient and outpatient education and intense telephone and clinic follow‐up), ACEi usage was maximized and re‐hospitalization and HFrelated death was significantly reduced at three months.16 Although effective, such interventions require substantial ongoing cost and sustainability is again called into question. Our initiative is unique in that incorporating a computer‐based semiautomatic system into the care‐delivery process has enhanced care quality without incurring ongoing extra cost (we have neither hired extra personnel nor created a heavier work burden for pharmacists and patientcare teams, as the model has been diffused into their daily routine) thus maximizing its longterm sustainability.

Notwithstanding the positive aspects, this study has several limitations. First, it is not a randomized, controlled trial, and unidentified external factors may have had some influence. However, in the examination of all potential external effects, we could not identify any factor that would have the capacity to substantially and consistently influence the results. Second, prepost study design is less ideal than randomized, controlled trials on the study design hierarchy. However, given the unsatisfactory adherence rate, anticipated positive effects with the model, and the pressing need for improving the adherence, a randomized trial was not an option at that juncture. Third, we could not precisely compare the difference in the awareness of ACEi/ARB guidelines among different classes of trainees during the study period. We did have a one‐time online, non‐mandatory education program for all providers. However, new trainees rotated in and‐ out on a monthly basis. This factor is unlikely to have caused a sustained change. Fourth, we did not have the outcome data for patients in whom HF was their secondary admission diagnosis. These patients were equally flagged by the model, and their ACEi/ARB status, when flagged, was obliged to be corrected. We suspect that these patients most likely benefited even more by the model because they were likely in a compensated state of HF, and the care‐teams tended to be more focused on their primary issue, leaving room for overlooking LVSD‐related issues.

Finally, we report the outcomes in the first 21 months after the full implementation of the model. We still need to monitor the long‐term outcome, although a reasonable length of time has elapsed. There has been no sign of decay in its effectiveness and we have no compelling reason to anticipate a significant regression.

Under ideal conditions, the outcome should consistently be 100% based on the design. In reality the adherence had been oscillating with an average of 97%. We noted two main scenarios that had contributed to this outcome. First, some LVSD/HF patients were taken off ACEi/ARB temporarily before discharge because of worsening pre‐renal azotemia with diuresis. They were discharged off ACEi/ARB with a plan to resume it. These patients would not have been labeled as ACEi/ARB‐intolerant but were classified as those without meeting the guidelines. Second, some patients had their echocardiogram on the same day or within 24 hours of discharge. A fraction of them had LVEF < 40%, but ACEi/ARB had not been initiated before discharge.

The rising volume of patients with increasing age and co‐morbidities, combined with constraints in healthcare resources, compels us to explore high‐efficiency care‐delivery models. Although computerized technology is well understood and readily available, the challenge we face is how to fully utilize the technology. A recent study shows that the improvement of IT infrastructure and research on implementation are interdependent and both can be translated to better patient care.25 Our experience serves as another example demonstrating that, when carefully conceived and properly executed, computer‐based care‐delivery prompts can be highly efficient and effective, suitable for large hospital settings with a heavy patient load like ours.

Moreover, because of the availability of basic IT platforms, similar algorithm‐based model systems can foreseeably be adopted by hospitals of comparable size and structure and also be applied to other care‐delivery settings including out‐patient clinics, chronic dialysis units and various long‐term care facilities.

Developing efficient, IT‐based quality improvement tools that facilitate the application of evidence‐based care and improve quality without significant additional resources is imperative in today's economic climate. Strategies such as our e‐messaging intervention with ACEi and ARB demonstrate sustainable improvement, can be applied to other conditions, and should be vigorously pursued.

Pneumonia results in some 1.2 million hospital admissions each year in the United States, is the second leading cause of hospitalization among patients over 65, and accounts for more than $10 billion annually in hospital expenditures.1, 2 As a result of complex demographic and clinical forces, including an aging population, increasing prevalence of comorbidities, and changes in antimicrobial resistance patterns, between the periods 1988 to 1990 and 2000 to 2002 the number of patients hospitalized for pneumonia grew by 20%, and pneumonia was the leading infectious cause of death.3, 4

Given its public health significance, pneumonia has been the subject of intensive quality measurement and improvement efforts for well over a decade. Two of the largest initiatives are the Centers for Medicare & Medicaid Services (CMS) National Pneumonia Project and The Joint Commission ORYX program.5, 6 These efforts have largely entailed measuring hospital performance on pneumonia‐specific processes of care, such as whether blood oxygen levels were assessed, whether blood cultures were drawn before antibiotic treatment was initiated, the choice and timing of antibiotics, and smoking cessation counseling and vaccination at the time of discharge. While measuring processes of care (especially when they are based on sound evidence), can provide insights about quality, and can help guide hospital improvement efforts, these measures necessarily focus on a narrow spectrum of the overall care provided. Outcomes can complement process measures by directing attention to the results of care, which are influenced by both measured and unmeasured factors, and which may be more relevant from the patient's perspective.79

In 2008 CMS expanded its public reporting initiatives by adding risk‐standardized hospital mortality rates for pneumonia to the Hospital Compare website (http://www.hospitalcompare.hhs.gov/).10 Readmission rates were added in 2009. We sought to examine patterns of hospital and regional performance for patients with pneumonia as reflected in 30‐day risk‐standardized readmission and mortality rates. Our report complements the June 2010 annual release of data on the Hospital Compare website. CMS also reports 30‐day risk‐standardized mortality and readmission for acute myocardial infarction and heart failure; a description of the 2010 reporting results for those measures are described elsewhere.

Methods

Results

Hospital‐Specific Risk‐Standardized 30‐Day Mortality and Readmission Rates

Of the 1,118,583 patients included in the mortality analysis 129,444 (11.6%) died within 30 days of hospital admission. The median (Q1, Q3) hospital 30‐day risk‐standardized mortality rate was 11.1% (10.0%, 12.3%), and ranged from 6.7% to 20.9% (Table 1, Figure 1). Hospitals at the 10th percentile had 30‐day risk‐standardized mortality rates of 9.0% while for those at the 90th percentile of performance the rate was 13.5%. The odds of all‐cause mortality for a patient treated at a hospital that was one standard deviation above the national average was 1.68 times higher than that of a patient treated at a hospital that was one standard deviation below the national average.

Figure 1

Table 1.Risk‐Standardized Hospital 30‐Day Pneumonia Mortality and Readmission Rates

	Mortality	Readmission
Abbreviation: SD, standard deviation.
Patients (n)	1118583	1161817
Hospitals (n)	4788	4813
Patient age, years, median (Q1, Q3)	81 (74,86)	80 (74,86)
Nonwhite, %	11.1	11.1
Hospital case volume, median (Q1, Q3)	168 (77,323)	174 (79,334)
Risk‐standardized hospital rate, mean (SD)	11.2 (1.2)	18.3 (0.9)
Minimum	6.7	13.6
1st percentile	7.5	14.9
5th percentile	8.5	15.8
10th percentile	9.0	16.4
25th percentile	10.0	17.2
Median	11.1	18.2
75th percentile	12.3	19.2
90th percentile	13.5	20.4
95th percentile	14.4	21.1
99th percentile	16.1	22.8
Maximum	20.9	26.7
Model fit statistics
c‐Statistic	0.72	0.63
Intrahospital Correlation	0.07	0.03

For the 3‐year period 2006 to 2009, 222 (4.7%) hospitals were categorized as having a mortality rate that was better than the national average, 3968 (83.7%) were no different than the national average, 221 (4.6%) were worse and 332 (7.0%) did not meet the minimum case threshold.

Among the 1,161,817 patients included in the readmission analysis 212,638 (18.3%) were readmitted within 30 days of hospital discharge. The median (Q1,Q3) 30‐day risk‐standardized readmission rate was 18.2% (17.2%, 19.2%) and ranged from 13.6% to 26.7% (Table 1, Figure 2). Hospitals at the 10th percentile had 30‐day risk‐standardized readmission rates of 16.4% while for those at the 90th percentile of performance the rate was 20.4%. The odds of all‐cause readmission for a patient treated at a hospital that was one standard deviation above the national average was 1.40 times higher than the odds of all‐cause readmission if treated at a hospital that was one standard deviation below the national average.

Figure 2

For the 3‐year period 2006 to 2009, 64 (1.3%) hospitals were categorized as having a readmission rate that was better than the national average, 4203 (88.2%) were no different than the national average, 163 (3.4%) were worse and 333 (7.0%) had less than 25 cases and were therefore not categorized.

While risk‐standardized readmission rates were substantially higher than risk‐standardized mortality rates, mortality rates varied more. For example, the top 10% of hospitals had a relative mortality rate that was 33% lower than those in the bottom 10%, as compared with just a 20% relative difference for readmission rates. The coefficient of variation, a normalized measure of dispersion that unlike the standard deviation is independent of the population mean, was 10.7 for risk‐standardized mortality rates and 4.9 for readmission rates.

Regional Risk‐Standardized 30‐Day Mortality and Readmission Rates

Figures 3 and 4 show the distribution of 30‐day risk‐standardized mortality and readmission rates among hospital referral regions by quintile. Highest mortality regions were found across the entire country, including parts of Northern New England, the Mid and South Atlantic, East and the West South Central, East and West North Central, and the Mountain and Pacific regions of the West. The lowest mortality rates were observed in Southern New England, parts of the Mid and South Atlantic, East and West South Central, and parts of the Mountain and Pacific regions of the West (Figure 3).

Figure 3

Figure 4

Readmission rates were higher in the eastern portions of the US (including the Northeast, Mid and South Atlantic, East South Central) as well as the East North Central, and small parts of the West North Central portions of the Midwest and in Central California. The lowest readmission rates were observed in the West (Mountain and Pacific regions), parts of the Midwest (East and West North Central) and small pockets within the South and Northeast (Figure 4).

Discussion

In this 3‐year analysis of patient, hospital, and regional outcomes we observed that pneumonia in the elderly remains a highly morbid illness, with a 30‐day mortality rate of approximately 11.6%. More notably we observed that risk‐standardized mortality rates, and to a lesser extent readmission rates, vary significantly across hospitals and regions. Finally, we observed that readmission rates, but not mortality rates, show strong geographic concentration.

These findings suggest possible opportunities to save or extend the lives of a substantial number of Americans, and to reduce the burden of rehospitalization on patients and families, if low performing institutions were able to achieve the performance of those with better outcomes. Additionally, because readmission is so common (nearly 1 in 5 patients), efforts to reduce overall health care spending should focus on this large potential source of savings.19 In this regard, impending changes in payment to hospitals around readmissions will change incentives for hospitals and physicians that may ultimately lead to lower readmission rates.20

Previous analyses of the quality of hospital care for patients with pneumonia have focused on the percentage of eligible patients who received guideline‐recommended antibiotics within a specified time frame (4 or 8 hours), and vaccination prior to hospital discharge.21, 22 These studies have highlighted large differences across hospitals and states in the percentage receiving recommended care. In contrast, the focus of this study was to compare risk‐standardized outcomes of care at the nation's hospitals and across its regions. This effort was guided by the notion that the measurement of care outcomes is an important complement to process measurement because outcomes represent a more holistic assessment of care, that an outcomes focus offers hospitals greater autonomy in terms of what processes to improve, and that outcomes are ultimately more meaningful to patients than the technical aspects of how the outcomes were achieved. In contrast to these earlier process‐oriented efforts, the magnitude of the differences we observed in mortality and readmission rates across hospitals was not nearly as large.

A recent analysis of the outcomes of care for patients with heart failure and acute myocardial infarction also found significant variation in both hospital and regional mortality and readmission rates.23 The relative differences in risk‐standardized hospital mortality rates across the 10th to 90th percentiles of hospital performance was 25% for acute myocardial infarction, and 39% for heart failure. By contrast, we found that the difference in risk‐standardized hospital mortality rates across the 10th to 90th percentiles in pneumonia was an even greater 50% (13.5% vs. 9.0%). Similar to the findings in acute myocardial infarction and heart failure, we observed that risk‐standardized mortality rates varied more so than did readmission rates.

Our study has a number of limitations. First, the analysis was restricted to Medicare patients only, and our findings may not be generalizable to younger patients. Second, our risk‐adjustment methods relied on claims data, not clinical information abstracted from charts. Nevertheless, we assessed comorbidities using all physician and hospital claims from the year prior to the index admission. Additionally our mortality and readmission models were validated against those based on medical record data and the outputs of the 2 approaches were highly correlated.15, 24, 25 Our study was restricted to patients with a principal diagnosis of pneumonia, and we therefore did not include those whose principal diagnosis was sepsis or respiratory failure and who had a secondary diagnosis of pneumonia. While this decision was made to reduce the risk of misclassifying complications of care as the reason for admission, we acknowledge that this is likely to have limited our study to patients with less severe disease, and may have introduced bias related to differences in hospital coding practices regarding the use of sepsis and respiratory failure codes. While we excluded patients with 1 day length of stay from the mortality analysis to reduce the risk of including patients in the measure who did not actually have pneumonia, we did not exclude them from the readmission analysis because very short length of stay may be a risk factor for readmission. An additional limitation of our study is that our findings are primarily descriptive, and we did not attempt to explain the sources of the variation we observed. For example, we did not examine the extent to which these differences might be explained by differences in adherence to process measures across hospitals or regions. However, if the experience in acute myocardial infarction can serve as a guide, then it is unlikely that more than a small fraction of the observed variation in outcomes can be attributed to factors such as antibiotic timing or selection.26 Additionally, we cannot explain why readmission rates were more geographically distributed than mortality rates, however it is possible that this may be related to the supply of physicians or hospital beds.27 Finally, some have argued that mortality and readmission rates do not necessarily reflect the very quality they intend to measure.2830

The outcomes of patients with pneumonia appear to be significantly influenced by both the hospital and region where they receive care. Efforts to improve population level outcomes might be informed by studying the practices of hospitals and regions that consistently achieve high levels of performance.31

Pneumonia results in some 1.2 million hospital admissions each year in the United States, is the second leading cause of hospitalization among patients over 65, and accounts for more than $10 billion annually in hospital expenditures.1, 2 As a result of complex demographic and clinical forces, including an aging population, increasing prevalence of comorbidities, and changes in antimicrobial resistance patterns, between the periods 1988 to 1990 and 2000 to 2002 the number of patients hospitalized for pneumonia grew by 20%, and pneumonia was the leading infectious cause of death.3, 4

Given its public health significance, pneumonia has been the subject of intensive quality measurement and improvement efforts for well over a decade. Two of the largest initiatives are the Centers for Medicare & Medicaid Services (CMS) National Pneumonia Project and The Joint Commission ORYX program.5, 6 These efforts have largely entailed measuring hospital performance on pneumonia‐specific processes of care, such as whether blood oxygen levels were assessed, whether blood cultures were drawn before antibiotic treatment was initiated, the choice and timing of antibiotics, and smoking cessation counseling and vaccination at the time of discharge. While measuring processes of care (especially when they are based on sound evidence), can provide insights about quality, and can help guide hospital improvement efforts, these measures necessarily focus on a narrow spectrum of the overall care provided. Outcomes can complement process measures by directing attention to the results of care, which are influenced by both measured and unmeasured factors, and which may be more relevant from the patient's perspective.79

In 2008 CMS expanded its public reporting initiatives by adding risk‐standardized hospital mortality rates for pneumonia to the Hospital Compare website (http://www.hospitalcompare.hhs.gov/).10 Readmission rates were added in 2009. We sought to examine patterns of hospital and regional performance for patients with pneumonia as reflected in 30‐day risk‐standardized readmission and mortality rates. Our report complements the June 2010 annual release of data on the Hospital Compare website. CMS also reports 30‐day risk‐standardized mortality and readmission for acute myocardial infarction and heart failure; a description of the 2010 reporting results for those measures are described elsewhere.

Methods

Results

Hospital‐Specific Risk‐Standardized 30‐Day Mortality and Readmission Rates

Of the 1,118,583 patients included in the mortality analysis 129,444 (11.6%) died within 30 days of hospital admission. The median (Q1, Q3) hospital 30‐day risk‐standardized mortality rate was 11.1% (10.0%, 12.3%), and ranged from 6.7% to 20.9% (Table 1, Figure 1). Hospitals at the 10th percentile had 30‐day risk‐standardized mortality rates of 9.0% while for those at the 90th percentile of performance the rate was 13.5%. The odds of all‐cause mortality for a patient treated at a hospital that was one standard deviation above the national average was 1.68 times higher than that of a patient treated at a hospital that was one standard deviation below the national average.

Figure 1

Table 1.Risk‐Standardized Hospital 30‐Day Pneumonia Mortality and Readmission Rates

	Mortality	Readmission
Abbreviation: SD, standard deviation.
Patients (n)	1118583	1161817
Hospitals (n)	4788	4813
Patient age, years, median (Q1, Q3)	81 (74,86)	80 (74,86)
Nonwhite, %	11.1	11.1
Hospital case volume, median (Q1, Q3)	168 (77,323)	174 (79,334)
Risk‐standardized hospital rate, mean (SD)	11.2 (1.2)	18.3 (0.9)
Minimum	6.7	13.6
1st percentile	7.5	14.9
5th percentile	8.5	15.8
10th percentile	9.0	16.4
25th percentile	10.0	17.2
Median	11.1	18.2
75th percentile	12.3	19.2
90th percentile	13.5	20.4
95th percentile	14.4	21.1
99th percentile	16.1	22.8
Maximum	20.9	26.7
Model fit statistics
c‐Statistic	0.72	0.63
Intrahospital Correlation	0.07	0.03

For the 3‐year period 2006 to 2009, 222 (4.7%) hospitals were categorized as having a mortality rate that was better than the national average, 3968 (83.7%) were no different than the national average, 221 (4.6%) were worse and 332 (7.0%) did not meet the minimum case threshold.

Among the 1,161,817 patients included in the readmission analysis 212,638 (18.3%) were readmitted within 30 days of hospital discharge. The median (Q1,Q3) 30‐day risk‐standardized readmission rate was 18.2% (17.2%, 19.2%) and ranged from 13.6% to 26.7% (Table 1, Figure 2). Hospitals at the 10th percentile had 30‐day risk‐standardized readmission rates of 16.4% while for those at the 90th percentile of performance the rate was 20.4%. The odds of all‐cause readmission for a patient treated at a hospital that was one standard deviation above the national average was 1.40 times higher than the odds of all‐cause readmission if treated at a hospital that was one standard deviation below the national average.

Figure 2

For the 3‐year period 2006 to 2009, 64 (1.3%) hospitals were categorized as having a readmission rate that was better than the national average, 4203 (88.2%) were no different than the national average, 163 (3.4%) were worse and 333 (7.0%) had less than 25 cases and were therefore not categorized.

While risk‐standardized readmission rates were substantially higher than risk‐standardized mortality rates, mortality rates varied more. For example, the top 10% of hospitals had a relative mortality rate that was 33% lower than those in the bottom 10%, as compared with just a 20% relative difference for readmission rates. The coefficient of variation, a normalized measure of dispersion that unlike the standard deviation is independent of the population mean, was 10.7 for risk‐standardized mortality rates and 4.9 for readmission rates.

Regional Risk‐Standardized 30‐Day Mortality and Readmission Rates

Figures 3 and 4 show the distribution of 30‐day risk‐standardized mortality and readmission rates among hospital referral regions by quintile. Highest mortality regions were found across the entire country, including parts of Northern New England, the Mid and South Atlantic, East and the West South Central, East and West North Central, and the Mountain and Pacific regions of the West. The lowest mortality rates were observed in Southern New England, parts of the Mid and South Atlantic, East and West South Central, and parts of the Mountain and Pacific regions of the West (Figure 3).

Figure 3

Figure 4

Readmission rates were higher in the eastern portions of the US (including the Northeast, Mid and South Atlantic, East South Central) as well as the East North Central, and small parts of the West North Central portions of the Midwest and in Central California. The lowest readmission rates were observed in the West (Mountain and Pacific regions), parts of the Midwest (East and West North Central) and small pockets within the South and Northeast (Figure 4).

Discussion

In this 3‐year analysis of patient, hospital, and regional outcomes we observed that pneumonia in the elderly remains a highly morbid illness, with a 30‐day mortality rate of approximately 11.6%. More notably we observed that risk‐standardized mortality rates, and to a lesser extent readmission rates, vary significantly across hospitals and regions. Finally, we observed that readmission rates, but not mortality rates, show strong geographic concentration.

These findings suggest possible opportunities to save or extend the lives of a substantial number of Americans, and to reduce the burden of rehospitalization on patients and families, if low performing institutions were able to achieve the performance of those with better outcomes. Additionally, because readmission is so common (nearly 1 in 5 patients), efforts to reduce overall health care spending should focus on this large potential source of savings.19 In this regard, impending changes in payment to hospitals around readmissions will change incentives for hospitals and physicians that may ultimately lead to lower readmission rates.20

Previous analyses of the quality of hospital care for patients with pneumonia have focused on the percentage of eligible patients who received guideline‐recommended antibiotics within a specified time frame (4 or 8 hours), and vaccination prior to hospital discharge.21, 22 These studies have highlighted large differences across hospitals and states in the percentage receiving recommended care. In contrast, the focus of this study was to compare risk‐standardized outcomes of care at the nation's hospitals and across its regions. This effort was guided by the notion that the measurement of care outcomes is an important complement to process measurement because outcomes represent a more holistic assessment of care, that an outcomes focus offers hospitals greater autonomy in terms of what processes to improve, and that outcomes are ultimately more meaningful to patients than the technical aspects of how the outcomes were achieved. In contrast to these earlier process‐oriented efforts, the magnitude of the differences we observed in mortality and readmission rates across hospitals was not nearly as large.

A recent analysis of the outcomes of care for patients with heart failure and acute myocardial infarction also found significant variation in both hospital and regional mortality and readmission rates.23 The relative differences in risk‐standardized hospital mortality rates across the 10th to 90th percentiles of hospital performance was 25% for acute myocardial infarction, and 39% for heart failure. By contrast, we found that the difference in risk‐standardized hospital mortality rates across the 10th to 90th percentiles in pneumonia was an even greater 50% (13.5% vs. 9.0%). Similar to the findings in acute myocardial infarction and heart failure, we observed that risk‐standardized mortality rates varied more so than did readmission rates.

Our study has a number of limitations. First, the analysis was restricted to Medicare patients only, and our findings may not be generalizable to younger patients. Second, our risk‐adjustment methods relied on claims data, not clinical information abstracted from charts. Nevertheless, we assessed comorbidities using all physician and hospital claims from the year prior to the index admission. Additionally our mortality and readmission models were validated against those based on medical record data and the outputs of the 2 approaches were highly correlated.15, 24, 25 Our study was restricted to patients with a principal diagnosis of pneumonia, and we therefore did not include those whose principal diagnosis was sepsis or respiratory failure and who had a secondary diagnosis of pneumonia. While this decision was made to reduce the risk of misclassifying complications of care as the reason for admission, we acknowledge that this is likely to have limited our study to patients with less severe disease, and may have introduced bias related to differences in hospital coding practices regarding the use of sepsis and respiratory failure codes. While we excluded patients with 1 day length of stay from the mortality analysis to reduce the risk of including patients in the measure who did not actually have pneumonia, we did not exclude them from the readmission analysis because very short length of stay may be a risk factor for readmission. An additional limitation of our study is that our findings are primarily descriptive, and we did not attempt to explain the sources of the variation we observed. For example, we did not examine the extent to which these differences might be explained by differences in adherence to process measures across hospitals or regions. However, if the experience in acute myocardial infarction can serve as a guide, then it is unlikely that more than a small fraction of the observed variation in outcomes can be attributed to factors such as antibiotic timing or selection.26 Additionally, we cannot explain why readmission rates were more geographically distributed than mortality rates, however it is possible that this may be related to the supply of physicians or hospital beds.27 Finally, some have argued that mortality and readmission rates do not necessarily reflect the very quality they intend to measure.2830

The outcomes of patients with pneumonia appear to be significantly influenced by both the hospital and region where they receive care. Efforts to improve population level outcomes might be informed by studying the practices of hospitals and regions that consistently achieve high levels of performance.31

Pneumonia results in some 1.2 million hospital admissions each year in the United States, is the second leading cause of hospitalization among patients over 65, and accounts for more than $10 billion annually in hospital expenditures.1, 2 As a result of complex demographic and clinical forces, including an aging population, increasing prevalence of comorbidities, and changes in antimicrobial resistance patterns, between the periods 1988 to 1990 and 2000 to 2002 the number of patients hospitalized for pneumonia grew by 20%, and pneumonia was the leading infectious cause of death.3, 4

Given its public health significance, pneumonia has been the subject of intensive quality measurement and improvement efforts for well over a decade. Two of the largest initiatives are the Centers for Medicare & Medicaid Services (CMS) National Pneumonia Project and The Joint Commission ORYX program.5, 6 These efforts have largely entailed measuring hospital performance on pneumonia‐specific processes of care, such as whether blood oxygen levels were assessed, whether blood cultures were drawn before antibiotic treatment was initiated, the choice and timing of antibiotics, and smoking cessation counseling and vaccination at the time of discharge. While measuring processes of care (especially when they are based on sound evidence), can provide insights about quality, and can help guide hospital improvement efforts, these measures necessarily focus on a narrow spectrum of the overall care provided. Outcomes can complement process measures by directing attention to the results of care, which are influenced by both measured and unmeasured factors, and which may be more relevant from the patient's perspective.79

In 2008 CMS expanded its public reporting initiatives by adding risk‐standardized hospital mortality rates for pneumonia to the Hospital Compare website (http://www.hospitalcompare.hhs.gov/).10 Readmission rates were added in 2009. We sought to examine patterns of hospital and regional performance for patients with pneumonia as reflected in 30‐day risk‐standardized readmission and mortality rates. Our report complements the June 2010 annual release of data on the Hospital Compare website. CMS also reports 30‐day risk‐standardized mortality and readmission for acute myocardial infarction and heart failure; a description of the 2010 reporting results for those measures are described elsewhere.

Methods

Results

Hospital‐Specific Risk‐Standardized 30‐Day Mortality and Readmission Rates

Of the 1,118,583 patients included in the mortality analysis 129,444 (11.6%) died within 30 days of hospital admission. The median (Q1, Q3) hospital 30‐day risk‐standardized mortality rate was 11.1% (10.0%, 12.3%), and ranged from 6.7% to 20.9% (Table 1, Figure 1). Hospitals at the 10th percentile had 30‐day risk‐standardized mortality rates of 9.0% while for those at the 90th percentile of performance the rate was 13.5%. The odds of all‐cause mortality for a patient treated at a hospital that was one standard deviation above the national average was 1.68 times higher than that of a patient treated at a hospital that was one standard deviation below the national average.

Figure 1

Table 1.Risk‐Standardized Hospital 30‐Day Pneumonia Mortality and Readmission Rates

	Mortality	Readmission
Abbreviation: SD, standard deviation.
Patients (n)	1118583	1161817
Hospitals (n)	4788	4813
Patient age, years, median (Q1, Q3)	81 (74,86)	80 (74,86)
Nonwhite, %	11.1	11.1
Hospital case volume, median (Q1, Q3)	168 (77,323)	174 (79,334)
Risk‐standardized hospital rate, mean (SD)	11.2 (1.2)	18.3 (0.9)
Minimum	6.7	13.6
1st percentile	7.5	14.9
5th percentile	8.5	15.8
10th percentile	9.0	16.4
25th percentile	10.0	17.2
Median	11.1	18.2
75th percentile	12.3	19.2
90th percentile	13.5	20.4
95th percentile	14.4	21.1
99th percentile	16.1	22.8
Maximum	20.9	26.7
Model fit statistics
c‐Statistic	0.72	0.63
Intrahospital Correlation	0.07	0.03

For the 3‐year period 2006 to 2009, 222 (4.7%) hospitals were categorized as having a mortality rate that was better than the national average, 3968 (83.7%) were no different than the national average, 221 (4.6%) were worse and 332 (7.0%) did not meet the minimum case threshold.

Among the 1,161,817 patients included in the readmission analysis 212,638 (18.3%) were readmitted within 30 days of hospital discharge. The median (Q1,Q3) 30‐day risk‐standardized readmission rate was 18.2% (17.2%, 19.2%) and ranged from 13.6% to 26.7% (Table 1, Figure 2). Hospitals at the 10th percentile had 30‐day risk‐standardized readmission rates of 16.4% while for those at the 90th percentile of performance the rate was 20.4%. The odds of all‐cause readmission for a patient treated at a hospital that was one standard deviation above the national average was 1.40 times higher than the odds of all‐cause readmission if treated at a hospital that was one standard deviation below the national average.

Figure 2

For the 3‐year period 2006 to 2009, 64 (1.3%) hospitals were categorized as having a readmission rate that was better than the national average, 4203 (88.2%) were no different than the national average, 163 (3.4%) were worse and 333 (7.0%) had less than 25 cases and were therefore not categorized.

While risk‐standardized readmission rates were substantially higher than risk‐standardized mortality rates, mortality rates varied more. For example, the top 10% of hospitals had a relative mortality rate that was 33% lower than those in the bottom 10%, as compared with just a 20% relative difference for readmission rates. The coefficient of variation, a normalized measure of dispersion that unlike the standard deviation is independent of the population mean, was 10.7 for risk‐standardized mortality rates and 4.9 for readmission rates.

Regional Risk‐Standardized 30‐Day Mortality and Readmission Rates

Figures 3 and 4 show the distribution of 30‐day risk‐standardized mortality and readmission rates among hospital referral regions by quintile. Highest mortality regions were found across the entire country, including parts of Northern New England, the Mid and South Atlantic, East and the West South Central, East and West North Central, and the Mountain and Pacific regions of the West. The lowest mortality rates were observed in Southern New England, parts of the Mid and South Atlantic, East and West South Central, and parts of the Mountain and Pacific regions of the West (Figure 3).

Figure 3

Figure 4

Readmission rates were higher in the eastern portions of the US (including the Northeast, Mid and South Atlantic, East South Central) as well as the East North Central, and small parts of the West North Central portions of the Midwest and in Central California. The lowest readmission rates were observed in the West (Mountain and Pacific regions), parts of the Midwest (East and West North Central) and small pockets within the South and Northeast (Figure 4).

Discussion

In this 3‐year analysis of patient, hospital, and regional outcomes we observed that pneumonia in the elderly remains a highly morbid illness, with a 30‐day mortality rate of approximately 11.6%. More notably we observed that risk‐standardized mortality rates, and to a lesser extent readmission rates, vary significantly across hospitals and regions. Finally, we observed that readmission rates, but not mortality rates, show strong geographic concentration.

These findings suggest possible opportunities to save or extend the lives of a substantial number of Americans, and to reduce the burden of rehospitalization on patients and families, if low performing institutions were able to achieve the performance of those with better outcomes. Additionally, because readmission is so common (nearly 1 in 5 patients), efforts to reduce overall health care spending should focus on this large potential source of savings.19 In this regard, impending changes in payment to hospitals around readmissions will change incentives for hospitals and physicians that may ultimately lead to lower readmission rates.20

Previous analyses of the quality of hospital care for patients with pneumonia have focused on the percentage of eligible patients who received guideline‐recommended antibiotics within a specified time frame (4 or 8 hours), and vaccination prior to hospital discharge.21, 22 These studies have highlighted large differences across hospitals and states in the percentage receiving recommended care. In contrast, the focus of this study was to compare risk‐standardized outcomes of care at the nation's hospitals and across its regions. This effort was guided by the notion that the measurement of care outcomes is an important complement to process measurement because outcomes represent a more holistic assessment of care, that an outcomes focus offers hospitals greater autonomy in terms of what processes to improve, and that outcomes are ultimately more meaningful to patients than the technical aspects of how the outcomes were achieved. In contrast to these earlier process‐oriented efforts, the magnitude of the differences we observed in mortality and readmission rates across hospitals was not nearly as large.

A recent analysis of the outcomes of care for patients with heart failure and acute myocardial infarction also found significant variation in both hospital and regional mortality and readmission rates.23 The relative differences in risk‐standardized hospital mortality rates across the 10th to 90th percentiles of hospital performance was 25% for acute myocardial infarction, and 39% for heart failure. By contrast, we found that the difference in risk‐standardized hospital mortality rates across the 10th to 90th percentiles in pneumonia was an even greater 50% (13.5% vs. 9.0%). Similar to the findings in acute myocardial infarction and heart failure, we observed that risk‐standardized mortality rates varied more so than did readmission rates.

Our study has a number of limitations. First, the analysis was restricted to Medicare patients only, and our findings may not be generalizable to younger patients. Second, our risk‐adjustment methods relied on claims data, not clinical information abstracted from charts. Nevertheless, we assessed comorbidities using all physician and hospital claims from the year prior to the index admission. Additionally our mortality and readmission models were validated against those based on medical record data and the outputs of the 2 approaches were highly correlated.15, 24, 25 Our study was restricted to patients with a principal diagnosis of pneumonia, and we therefore did not include those whose principal diagnosis was sepsis or respiratory failure and who had a secondary diagnosis of pneumonia. While this decision was made to reduce the risk of misclassifying complications of care as the reason for admission, we acknowledge that this is likely to have limited our study to patients with less severe disease, and may have introduced bias related to differences in hospital coding practices regarding the use of sepsis and respiratory failure codes. While we excluded patients with 1 day length of stay from the mortality analysis to reduce the risk of including patients in the measure who did not actually have pneumonia, we did not exclude them from the readmission analysis because very short length of stay may be a risk factor for readmission. An additional limitation of our study is that our findings are primarily descriptive, and we did not attempt to explain the sources of the variation we observed. For example, we did not examine the extent to which these differences might be explained by differences in adherence to process measures across hospitals or regions. However, if the experience in acute myocardial infarction can serve as a guide, then it is unlikely that more than a small fraction of the observed variation in outcomes can be attributed to factors such as antibiotic timing or selection.26 Additionally, we cannot explain why readmission rates were more geographically distributed than mortality rates, however it is possible that this may be related to the supply of physicians or hospital beds.27 Finally, some have argued that mortality and readmission rates do not necessarily reflect the very quality they intend to measure.2830

The outcomes of patients with pneumonia appear to be significantly influenced by both the hospital and region where they receive care. Efforts to improve population level outcomes might be informed by studying the practices of hospitals and regions that consistently achieve high levels of performance.31

Technological advances drive medical providers to specialize through the need for proficiency around increasingly focused areas of expertise.1 But the benefits of specialization are attained only by balancing the advantages of increasing expertise and the costs of coordinating care that must be borne as specialization increases.2 Integrating experts into modern medical delivery systems requires attention to the coordinating mechanisms that govern team‐based care.3

Coordination, defined as the management of task interdependencies,4 is a central component and a useful measure of teamwork.5 Several studies demonstrate the patient‐level impact of coordination among providers.69 Gittell et al.8 demonstrated that orthopedic hospitals whose staff had better relational coordination (RC) measures had shorter lengths of stay and better post‐operative pain control for patients undergoing surgery. In medical intensive care units (ICUs), Wheelan et al.9 showed that staff members of units with lower mortality rates perceived their teams as functioning at higher stages of group development and perceived their team members as less dependent and more trusting.

Communication is the cornerstone of effective team coordination.10, 11 As such, practice model interventions that facilitate frequent communication of higher quality are associated with lower error rates10 and better teamwork.11 The use of hospitalists, for example, is shown to capitalize on this advantage by improving coordination through physician availability that facilitates communication and relational interactions among hospital‐based staff.12 While system‐level interventions such as this have received significant attention from experts in organizations, empirical studies that explore the contribution of team member characteristics to overall coordination are lacking.13

Inpatient comanagement services offer a unique model for studying teamwork. While the label is used to describe a variety of arrangements,1416 comanagement broadly describes a practice model wherein providers of various specialties deliver direct care to patients, in contrast to the traditional generalist‐consultant model in which specialists lend expertise.17 Many recent comanagement practices involve hospitalists in partnership with surgeons in the care of patients with concurrent medical and surgical needs,18 but similar arrangements between hospitalists and medical subspecialists are being adopted in some medical centers for the care of complex patients with conditions such as heart failure, cancer, stroke, and solid organ transplantations. Coordination among providers has not been studied in this context.

The goals of this study are: (1) to measure the input of individual providers to the overall coordination of care on a highly interdependent medical comanagement service, (2) to characterize high and low coordinators, and (3) to explore the relationship between coordination and patient outcomes. The main hypothesis is that the quality of team coordination is determined partly by the attributes of its members such that their individual contributions to the coordination of care affect the outcomes of vulnerable hospitalized patients.

Materials and Methods

Results

All 32 providers (100%) completed the Baseline Survey and participated in the Repeated Surveys of which 177/224 (79%) were completed. The median number of surveys that contributed to the calculation of individual RC and the mean RC by provider type are summarized in Table 1.

Of the 119 patients managed on the service, the mean age (standard deviation [SD]) was 55 (14) years and 48% were women. Of the 201 hospitalizations, there were 13 floor‐to‐ICU transfers and 5 in‐hospital deaths, however, we excluded from the analysis 1 death of a patient who was admitted under inpatient hospice status.

Discussion

By adapting Gittell's RC instrument to focus on individual providers, we found that their characteristic attributes such as preference for particular management styles, leadership quality, and patient ownership are associated with their externally perceived contributions to the overall coordination of care. In an unadjusted analysis, we also observed an intriguing trend towards more frequent major hospital complications when the worst coordinators of each physician type were on service.

Existing evidence22, 23 mostly summarized in a recent RAND Health report shows a weak association between clinical teamwork quality and patient mortality. While our data also support this association, it does so with limitations. Most importantly, the small sample size limited our ability to rigorously account for potential confounders that may have contributed to this apparent association. Further studies may better address whether or not bad outcomes are indeed associated with poor coordinators in highly interdependent clinical teams. In addition to confounding, the small sample size of providers makes the analysis vulnerable to type 1 errors. We addressed this issue by intensively surveying providers repeatedly to achieve a high resolution of the coordination and management structure measures from each comanaged team. The potential for omitted variables and reverse causality in that the coordination scores may be negatively influenced by particularly complex patients and bad outcomes remains a valid concern. We addressed this by confirming the stability of provider RC over time and excluding the RC data from rotations with a bad outcome, but the negative perception of an individual tied to past bad outcomes may persist beyond a particular rotation. Survey responses are subject to recall and hindsight biases, which we attempted to minimize by surveying respondents immediately after each team rotation. Finally, all of our findings may be not be generalizable to other comanagement settings. However, the important correlations between coordination and quality have been observed in other contexts.24, 25

In our study, in‐hospital deaths and ICU transfers are treated as consequences of uncoordinated care. This interpretation may be problematic for circumstances when death is inevitable no matter how well coordinated the care, or when transfer to a higher level of care is appropriate. The rationale for grouping the 2 events into 1 composite bad outcome is based on the assumption that both death and the escalation of care can be delayed to an extent, if not wholly prevented, with the coordinated utilization of a modern hospital's resources. The attribution of these events to poor coordinators may indicate the unraveling of coordination that normally must be maintained to help patients overcome decompensating events that are particularly common in the course of patients with severe liver diseases. Due to the exploratory nature of this analysis, additional studies are necessary to fully characterize the relationship between care coordination and care transfers.

An important implication of this study is that the communication skill and ethical disposition of each individual provider is relevant to the coordination that is sought in multi‐provider teams. Training medical professionals to be better team members may have direct impact on the patients they serve. Our finding about patient ownership suggests that commitment to patients in the framework of care is not merely tradition but a characteristic of competent physicians. Moreover, physicians' commitment to patients is a possible factor, not just in achieving patients' satisfaction, but in securing better outcomes. To that end, the teaching of this and other humanistic principles must remain a vital part of medical education at all levels of training.

Several implications about team leadership and hierarchy are apparent from the data. Findings around the perceived assignment of responsibilities show that high coordinator hepatologists acknowledge the advantages of overlapping task boundaries to prevent critical tasks from being missed and risking bad outcomes. High RC hepatologists in our study adopted a more participatory than supervisory role which presumably facilitated better coordination by transmitting organizational goals to other team members. The function of a comanaged team is likely to be enhanced by a fluid assignment of roles to better handle tasks with high uncertainty. Accordingly, comanagement models of care may not be appropriate in settings where tasks are not interdependent.26 Inherent hierarchy appears to be a feature of well coordinated teams. One possible interpretation of our data is that hospitalists who yield the leadership role to the hepatologist are perceived to be better coordinators and that those who insist on exerting more influence in team decisions are perceived to be poor coordinators.

Existing evidence around care coordination predicts that comanagement designs improve provider coordination through stage‐based and site‐based specialization.12 However, the mechanisms that mediate coordination and patient outcomes are not clear. Moreover, the mechanisms of coordinating multi‐disciplinary teams may be specific to each clinical setting. The role of individual provider characteristics on coordination deserves more attention. Similarly, the impact of organizational culture under which favorable provider characteristics thrive is unknown. Finally, a detailed exposition of patient ownership and the role patients play in affecting the coordination of healthcare resources needs further exploration.

Technological advances drive medical providers to specialize through the need for proficiency around increasingly focused areas of expertise.1 But the benefits of specialization are attained only by balancing the advantages of increasing expertise and the costs of coordinating care that must be borne as specialization increases.2 Integrating experts into modern medical delivery systems requires attention to the coordinating mechanisms that govern team‐based care.3

Coordination, defined as the management of task interdependencies,4 is a central component and a useful measure of teamwork.5 Several studies demonstrate the patient‐level impact of coordination among providers.69 Gittell et al.8 demonstrated that orthopedic hospitals whose staff had better relational coordination (RC) measures had shorter lengths of stay and better post‐operative pain control for patients undergoing surgery. In medical intensive care units (ICUs), Wheelan et al.9 showed that staff members of units with lower mortality rates perceived their teams as functioning at higher stages of group development and perceived their team members as less dependent and more trusting.

Communication is the cornerstone of effective team coordination.10, 11 As such, practice model interventions that facilitate frequent communication of higher quality are associated with lower error rates10 and better teamwork.11 The use of hospitalists, for example, is shown to capitalize on this advantage by improving coordination through physician availability that facilitates communication and relational interactions among hospital‐based staff.12 While system‐level interventions such as this have received significant attention from experts in organizations, empirical studies that explore the contribution of team member characteristics to overall coordination are lacking.13

Inpatient comanagement services offer a unique model for studying teamwork. While the label is used to describe a variety of arrangements,1416 comanagement broadly describes a practice model wherein providers of various specialties deliver direct care to patients, in contrast to the traditional generalist‐consultant model in which specialists lend expertise.17 Many recent comanagement practices involve hospitalists in partnership with surgeons in the care of patients with concurrent medical and surgical needs,18 but similar arrangements between hospitalists and medical subspecialists are being adopted in some medical centers for the care of complex patients with conditions such as heart failure, cancer, stroke, and solid organ transplantations. Coordination among providers has not been studied in this context.

The goals of this study are: (1) to measure the input of individual providers to the overall coordination of care on a highly interdependent medical comanagement service, (2) to characterize high and low coordinators, and (3) to explore the relationship between coordination and patient outcomes. The main hypothesis is that the quality of team coordination is determined partly by the attributes of its members such that their individual contributions to the coordination of care affect the outcomes of vulnerable hospitalized patients.

Materials and Methods

Results

All 32 providers (100%) completed the Baseline Survey and participated in the Repeated Surveys of which 177/224 (79%) were completed. The median number of surveys that contributed to the calculation of individual RC and the mean RC by provider type are summarized in Table 1.

Of the 119 patients managed on the service, the mean age (standard deviation [SD]) was 55 (14) years and 48% were women. Of the 201 hospitalizations, there were 13 floor‐to‐ICU transfers and 5 in‐hospital deaths, however, we excluded from the analysis 1 death of a patient who was admitted under inpatient hospice status.

Discussion

By adapting Gittell's RC instrument to focus on individual providers, we found that their characteristic attributes such as preference for particular management styles, leadership quality, and patient ownership are associated with their externally perceived contributions to the overall coordination of care. In an unadjusted analysis, we also observed an intriguing trend towards more frequent major hospital complications when the worst coordinators of each physician type were on service.

Existing evidence22, 23 mostly summarized in a recent RAND Health report shows a weak association between clinical teamwork quality and patient mortality. While our data also support this association, it does so with limitations. Most importantly, the small sample size limited our ability to rigorously account for potential confounders that may have contributed to this apparent association. Further studies may better address whether or not bad outcomes are indeed associated with poor coordinators in highly interdependent clinical teams. In addition to confounding, the small sample size of providers makes the analysis vulnerable to type 1 errors. We addressed this issue by intensively surveying providers repeatedly to achieve a high resolution of the coordination and management structure measures from each comanaged team. The potential for omitted variables and reverse causality in that the coordination scores may be negatively influenced by particularly complex patients and bad outcomes remains a valid concern. We addressed this by confirming the stability of provider RC over time and excluding the RC data from rotations with a bad outcome, but the negative perception of an individual tied to past bad outcomes may persist beyond a particular rotation. Survey responses are subject to recall and hindsight biases, which we attempted to minimize by surveying respondents immediately after each team rotation. Finally, all of our findings may be not be generalizable to other comanagement settings. However, the important correlations between coordination and quality have been observed in other contexts.24, 25

In our study, in‐hospital deaths and ICU transfers are treated as consequences of uncoordinated care. This interpretation may be problematic for circumstances when death is inevitable no matter how well coordinated the care, or when transfer to a higher level of care is appropriate. The rationale for grouping the 2 events into 1 composite bad outcome is based on the assumption that both death and the escalation of care can be delayed to an extent, if not wholly prevented, with the coordinated utilization of a modern hospital's resources. The attribution of these events to poor coordinators may indicate the unraveling of coordination that normally must be maintained to help patients overcome decompensating events that are particularly common in the course of patients with severe liver diseases. Due to the exploratory nature of this analysis, additional studies are necessary to fully characterize the relationship between care coordination and care transfers.

An important implication of this study is that the communication skill and ethical disposition of each individual provider is relevant to the coordination that is sought in multi‐provider teams. Training medical professionals to be better team members may have direct impact on the patients they serve. Our finding about patient ownership suggests that commitment to patients in the framework of care is not merely tradition but a characteristic of competent physicians. Moreover, physicians' commitment to patients is a possible factor, not just in achieving patients' satisfaction, but in securing better outcomes. To that end, the teaching of this and other humanistic principles must remain a vital part of medical education at all levels of training.

Several implications about team leadership and hierarchy are apparent from the data. Findings around the perceived assignment of responsibilities show that high coordinator hepatologists acknowledge the advantages of overlapping task boundaries to prevent critical tasks from being missed and risking bad outcomes. High RC hepatologists in our study adopted a more participatory than supervisory role which presumably facilitated better coordination by transmitting organizational goals to other team members. The function of a comanaged team is likely to be enhanced by a fluid assignment of roles to better handle tasks with high uncertainty. Accordingly, comanagement models of care may not be appropriate in settings where tasks are not interdependent.26 Inherent hierarchy appears to be a feature of well coordinated teams. One possible interpretation of our data is that hospitalists who yield the leadership role to the hepatologist are perceived to be better coordinators and that those who insist on exerting more influence in team decisions are perceived to be poor coordinators.

Existing evidence around care coordination predicts that comanagement designs improve provider coordination through stage‐based and site‐based specialization.12 However, the mechanisms that mediate coordination and patient outcomes are not clear. Moreover, the mechanisms of coordinating multi‐disciplinary teams may be specific to each clinical setting. The role of individual provider characteristics on coordination deserves more attention. Similarly, the impact of organizational culture under which favorable provider characteristics thrive is unknown. Finally, a detailed exposition of patient ownership and the role patients play in affecting the coordination of healthcare resources needs further exploration.

Technological advances drive medical providers to specialize through the need for proficiency around increasingly focused areas of expertise.1 But the benefits of specialization are attained only by balancing the advantages of increasing expertise and the costs of coordinating care that must be borne as specialization increases.2 Integrating experts into modern medical delivery systems requires attention to the coordinating mechanisms that govern team‐based care.3

Coordination, defined as the management of task interdependencies,4 is a central component and a useful measure of teamwork.5 Several studies demonstrate the patient‐level impact of coordination among providers.69 Gittell et al.8 demonstrated that orthopedic hospitals whose staff had better relational coordination (RC) measures had shorter lengths of stay and better post‐operative pain control for patients undergoing surgery. In medical intensive care units (ICUs), Wheelan et al.9 showed that staff members of units with lower mortality rates perceived their teams as functioning at higher stages of group development and perceived their team members as less dependent and more trusting.

Communication is the cornerstone of effective team coordination.10, 11 As such, practice model interventions that facilitate frequent communication of higher quality are associated with lower error rates10 and better teamwork.11 The use of hospitalists, for example, is shown to capitalize on this advantage by improving coordination through physician availability that facilitates communication and relational interactions among hospital‐based staff.12 While system‐level interventions such as this have received significant attention from experts in organizations, empirical studies that explore the contribution of team member characteristics to overall coordination are lacking.13

Inpatient comanagement services offer a unique model for studying teamwork. While the label is used to describe a variety of arrangements,1416 comanagement broadly describes a practice model wherein providers of various specialties deliver direct care to patients, in contrast to the traditional generalist‐consultant model in which specialists lend expertise.17 Many recent comanagement practices involve hospitalists in partnership with surgeons in the care of patients with concurrent medical and surgical needs,18 but similar arrangements between hospitalists and medical subspecialists are being adopted in some medical centers for the care of complex patients with conditions such as heart failure, cancer, stroke, and solid organ transplantations. Coordination among providers has not been studied in this context.

The goals of this study are: (1) to measure the input of individual providers to the overall coordination of care on a highly interdependent medical comanagement service, (2) to characterize high and low coordinators, and (3) to explore the relationship between coordination and patient outcomes. The main hypothesis is that the quality of team coordination is determined partly by the attributes of its members such that their individual contributions to the coordination of care affect the outcomes of vulnerable hospitalized patients.

Materials and Methods

Results

All 32 providers (100%) completed the Baseline Survey and participated in the Repeated Surveys of which 177/224 (79%) were completed. The median number of surveys that contributed to the calculation of individual RC and the mean RC by provider type are summarized in Table 1.

Of the 119 patients managed on the service, the mean age (standard deviation [SD]) was 55 (14) years and 48% were women. Of the 201 hospitalizations, there were 13 floor‐to‐ICU transfers and 5 in‐hospital deaths, however, we excluded from the analysis 1 death of a patient who was admitted under inpatient hospice status.

Discussion

By adapting Gittell's RC instrument to focus on individual providers, we found that their characteristic attributes such as preference for particular management styles, leadership quality, and patient ownership are associated with their externally perceived contributions to the overall coordination of care. In an unadjusted analysis, we also observed an intriguing trend towards more frequent major hospital complications when the worst coordinators of each physician type were on service.

Existing evidence22, 23 mostly summarized in a recent RAND Health report shows a weak association between clinical teamwork quality and patient mortality. While our data also support this association, it does so with limitations. Most importantly, the small sample size limited our ability to rigorously account for potential confounders that may have contributed to this apparent association. Further studies may better address whether or not bad outcomes are indeed associated with poor coordinators in highly interdependent clinical teams. In addition to confounding, the small sample size of providers makes the analysis vulnerable to type 1 errors. We addressed this issue by intensively surveying providers repeatedly to achieve a high resolution of the coordination and management structure measures from each comanaged team. The potential for omitted variables and reverse causality in that the coordination scores may be negatively influenced by particularly complex patients and bad outcomes remains a valid concern. We addressed this by confirming the stability of provider RC over time and excluding the RC data from rotations with a bad outcome, but the negative perception of an individual tied to past bad outcomes may persist beyond a particular rotation. Survey responses are subject to recall and hindsight biases, which we attempted to minimize by surveying respondents immediately after each team rotation. Finally, all of our findings may be not be generalizable to other comanagement settings. However, the important correlations between coordination and quality have been observed in other contexts.24, 25

In our study, in‐hospital deaths and ICU transfers are treated as consequences of uncoordinated care. This interpretation may be problematic for circumstances when death is inevitable no matter how well coordinated the care, or when transfer to a higher level of care is appropriate. The rationale for grouping the 2 events into 1 composite bad outcome is based on the assumption that both death and the escalation of care can be delayed to an extent, if not wholly prevented, with the coordinated utilization of a modern hospital's resources. The attribution of these events to poor coordinators may indicate the unraveling of coordination that normally must be maintained to help patients overcome decompensating events that are particularly common in the course of patients with severe liver diseases. Due to the exploratory nature of this analysis, additional studies are necessary to fully characterize the relationship between care coordination and care transfers.

An important implication of this study is that the communication skill and ethical disposition of each individual provider is relevant to the coordination that is sought in multi‐provider teams. Training medical professionals to be better team members may have direct impact on the patients they serve. Our finding about patient ownership suggests that commitment to patients in the framework of care is not merely tradition but a characteristic of competent physicians. Moreover, physicians' commitment to patients is a possible factor, not just in achieving patients' satisfaction, but in securing better outcomes. To that end, the teaching of this and other humanistic principles must remain a vital part of medical education at all levels of training.

Several implications about team leadership and hierarchy are apparent from the data. Findings around the perceived assignment of responsibilities show that high coordinator hepatologists acknowledge the advantages of overlapping task boundaries to prevent critical tasks from being missed and risking bad outcomes. High RC hepatologists in our study adopted a more participatory than supervisory role which presumably facilitated better coordination by transmitting organizational goals to other team members. The function of a comanaged team is likely to be enhanced by a fluid assignment of roles to better handle tasks with high uncertainty. Accordingly, comanagement models of care may not be appropriate in settings where tasks are not interdependent.26 Inherent hierarchy appears to be a feature of well coordinated teams. One possible interpretation of our data is that hospitalists who yield the leadership role to the hepatologist are perceived to be better coordinators and that those who insist on exerting more influence in team decisions are perceived to be poor coordinators.

Existing evidence around care coordination predicts that comanagement designs improve provider coordination through stage‐based and site‐based specialization.12 However, the mechanisms that mediate coordination and patient outcomes are not clear. Moreover, the mechanisms of coordinating multi‐disciplinary teams may be specific to each clinical setting. The role of individual provider characteristics on coordination deserves more attention. Similarly, the impact of organizational culture under which favorable provider characteristics thrive is unknown. Finally, a detailed exposition of patient ownership and the role patients play in affecting the coordination of healthcare resources needs further exploration.