Original Research

To the Editor:

Oral immunosuppressive therapies are prescribed for moderate to severe atopic dermatitis. Patients often consult YouTube to make informed decisions about these therapies. In the United States, most health-related online searches are initiated through a search engine, which frequently leads to social media sites such as YouTube. Recent studies have examined the reasons why users turn to the Internet for health-related information, indicating that users typically seek specific information regarding health concerns.^1,2 Furthermore, social media platforms such as YouTube are a popular means of sharing health information with the public.^3-5 Currently, YouTube has more than 1 billion registered users, and 30 million health-related videos are watched each day.⁶ Almost one-third of US consumers use YouTube, Facebook, and Twitter to obtain medical information.⁷ YouTube is a versatile tool because of its video-discovery mechanisms such as a keyword-based search engine, video-recommendation system, highlight feature for videos on home pages, and the capacity to embed YouTube videos on various web pages.⁸ Searchers use videos that are short, fast paced, emotion evoking, from credible sources, recently uploaded, and relevant to the searcher for aiding in health decisions.⁹ Furthermore, studies have demonstrated YouTube’s capacity to support a change in attitude and increase users’ knowledge. In fact, YouTube had higher impact on recall, attitudes, and behaviors when compared with written materials on other social media platforms, such as Facebook and Twitter.⁹ We conducted a cross-sectional study to examine the quality of YouTube videos on oral therapies for atopic dermatitis, such as cyclosporine, methotrexate, azathioprine, and mycophenolate mofetil.

On April 23, 2020, we performed 8 searches using a private browser with default filters on YouTube (Figure). Injectables were not included in the analysis, as the YouTube experience on dupilumab previously has been investigated.¹⁰ The top 40 videos from each search were screened by 3 researchers. Duplicates, non–English-language videos, and videos that did not discuss atopic dermatitis or oral therapies were excluded, resulting in 73 videos included in this analysis. Testimonials generated by patients made up 39 of 73 (53.4%) videos. Health care professionals created 23 of 73 (31.5%) videos, and educators with financial interest created 11 of 73 (15.1%) videos. The dates of production for the videos spanned from 2008 to 2020.

Importantly, this analysis is limited by its lack of quality-assessment tools for video-based resources such as JAMA score and DISCERN score.^14,15 However, these metrics have limited ability to evaluate audiovisual elements, indicating the need for novel tools to score their validity.

To the Editor:

Oral immunosuppressive therapies are prescribed for moderate to severe atopic dermatitis. Patients often consult YouTube to make informed decisions about these therapies. In the United States, most health-related online searches are initiated through a search engine, which frequently leads to social media sites such as YouTube. Recent studies have examined the reasons why users turn to the Internet for health-related information, indicating that users typically seek specific information regarding health concerns.^1,2 Furthermore, social media platforms such as YouTube are a popular means of sharing health information with the public.^3-5 Currently, YouTube has more than 1 billion registered users, and 30 million health-related videos are watched each day.⁶ Almost one-third of US consumers use YouTube, Facebook, and Twitter to obtain medical information.⁷ YouTube is a versatile tool because of its video-discovery mechanisms such as a keyword-based search engine, video-recommendation system, highlight feature for videos on home pages, and the capacity to embed YouTube videos on various web pages.⁸ Searchers use videos that are short, fast paced, emotion evoking, from credible sources, recently uploaded, and relevant to the searcher for aiding in health decisions.⁹ Furthermore, studies have demonstrated YouTube’s capacity to support a change in attitude and increase users’ knowledge. In fact, YouTube had higher impact on recall, attitudes, and behaviors when compared with written materials on other social media platforms, such as Facebook and Twitter.⁹ We conducted a cross-sectional study to examine the quality of YouTube videos on oral therapies for atopic dermatitis, such as cyclosporine, methotrexate, azathioprine, and mycophenolate mofetil.

On April 23, 2020, we performed 8 searches using a private browser with default filters on YouTube (Figure). Injectables were not included in the analysis, as the YouTube experience on dupilumab previously has been investigated.¹⁰ The top 40 videos from each search were screened by 3 researchers. Duplicates, non–English-language videos, and videos that did not discuss atopic dermatitis or oral therapies were excluded, resulting in 73 videos included in this analysis. Testimonials generated by patients made up 39 of 73 (53.4%) videos. Health care professionals created 23 of 73 (31.5%) videos, and educators with financial interest created 11 of 73 (15.1%) videos. The dates of production for the videos spanned from 2008 to 2020.

Importantly, this analysis is limited by its lack of quality-assessment tools for video-based resources such as JAMA score and DISCERN score.^14,15 However, these metrics have limited ability to evaluate audiovisual elements, indicating the need for novel tools to score their validity.

To the Editor:

Oral immunosuppressive therapies are prescribed for moderate to severe atopic dermatitis. Patients often consult YouTube to make informed decisions about these therapies. In the United States, most health-related online searches are initiated through a search engine, which frequently leads to social media sites such as YouTube. Recent studies have examined the reasons why users turn to the Internet for health-related information, indicating that users typically seek specific information regarding health concerns.^1,2 Furthermore, social media platforms such as YouTube are a popular means of sharing health information with the public.^3-5 Currently, YouTube has more than 1 billion registered users, and 30 million health-related videos are watched each day.⁶ Almost one-third of US consumers use YouTube, Facebook, and Twitter to obtain medical information.⁷ YouTube is a versatile tool because of its video-discovery mechanisms such as a keyword-based search engine, video-recommendation system, highlight feature for videos on home pages, and the capacity to embed YouTube videos on various web pages.⁸ Searchers use videos that are short, fast paced, emotion evoking, from credible sources, recently uploaded, and relevant to the searcher for aiding in health decisions.⁹ Furthermore, studies have demonstrated YouTube’s capacity to support a change in attitude and increase users’ knowledge. In fact, YouTube had higher impact on recall, attitudes, and behaviors when compared with written materials on other social media platforms, such as Facebook and Twitter.⁹ We conducted a cross-sectional study to examine the quality of YouTube videos on oral therapies for atopic dermatitis, such as cyclosporine, methotrexate, azathioprine, and mycophenolate mofetil.

On April 23, 2020, we performed 8 searches using a private browser with default filters on YouTube (Figure). Injectables were not included in the analysis, as the YouTube experience on dupilumab previously has been investigated.¹⁰ The top 40 videos from each search were screened by 3 researchers. Duplicates, non–English-language videos, and videos that did not discuss atopic dermatitis or oral therapies were excluded, resulting in 73 videos included in this analysis. Testimonials generated by patients made up 39 of 73 (53.4%) videos. Health care professionals created 23 of 73 (31.5%) videos, and educators with financial interest created 11 of 73 (15.1%) videos. The dates of production for the videos spanned from 2008 to 2020.

Importantly, this analysis is limited by its lack of quality-assessment tools for video-based resources such as JAMA score and DISCERN score.^14,15 However, these metrics have limited ability to evaluate audiovisual elements, indicating the need for novel tools to score their validity.

To the Editor:

Atopic dermatitis (eczema) affects approximately 20% of children worldwide.¹ In atopic dermatitis management, patient education is crucial for optimal outcomes.² The COVID-19 pandemic has impacted patient-physician interactions. To ensure safety of patients and physicians, visits may have been canceled, postponed, or conducted virtually, leaving less time for discussion and questions.³ As a consequence, patients may seek information about atopic dermatitis from alternative sources, including YouTube videos. We performed a cross-sectional study to analyze YouTube videos about topical treatments for atopic dermatitis.

During the week of July 16, 2020, we performed 4 private browser YouTube searches with default filters using the following terms: eczema topicals, eczema topical treatments, atopic dermatitis topicals, and atopic dermatitis topical treatments. For video selection, we defined topical treatments as topical corticosteroids, topical calcineurin inhibitors, crisaborole, emollients, wet wraps, and any prospective treatment topically administered. For each of the 4 searches, 2 researchers (A.M. and A.T.) independently examined the top 75 videos, yielding a total of 300 videos. Of them, 98 videos were duplicates, 19 videos were not about atopic dermatitis, and 91 videos were not about topical treatments, leaving a total of 92 videos for analysis (Figure 1).

Although less popular, professional source videos had a significantly higher DISCERN overall quality rating score (question 16) compared to those categorized as nonprofessional source (3.92 vs 1.53; P<.001). In contrast, nonprofessional source videos scored significantly higher on the quality-of-life question (question 13) compared to professional source videos (3.90 vs 2.56; P<.001)(eTable). (Three professional source videos were removed from YouTube before DISCERN scores could be assigned.)

To the Editor:

Atopic dermatitis (eczema) affects approximately 20% of children worldwide.¹ In atopic dermatitis management, patient education is crucial for optimal outcomes.² The COVID-19 pandemic has impacted patient-physician interactions. To ensure safety of patients and physicians, visits may have been canceled, postponed, or conducted virtually, leaving less time for discussion and questions.³ As a consequence, patients may seek information about atopic dermatitis from alternative sources, including YouTube videos. We performed a cross-sectional study to analyze YouTube videos about topical treatments for atopic dermatitis.

During the week of July 16, 2020, we performed 4 private browser YouTube searches with default filters using the following terms: eczema topicals, eczema topical treatments, atopic dermatitis topicals, and atopic dermatitis topical treatments. For video selection, we defined topical treatments as topical corticosteroids, topical calcineurin inhibitors, crisaborole, emollients, wet wraps, and any prospective treatment topically administered. For each of the 4 searches, 2 researchers (A.M. and A.T.) independently examined the top 75 videos, yielding a total of 300 videos. Of them, 98 videos were duplicates, 19 videos were not about atopic dermatitis, and 91 videos were not about topical treatments, leaving a total of 92 videos for analysis (Figure 1).

Although less popular, professional source videos had a significantly higher DISCERN overall quality rating score (question 16) compared to those categorized as nonprofessional source (3.92 vs 1.53; P<.001). In contrast, nonprofessional source videos scored significantly higher on the quality-of-life question (question 13) compared to professional source videos (3.90 vs 2.56; P<.001)(eTable). (Three professional source videos were removed from YouTube before DISCERN scores could be assigned.)

To the Editor:

Atopic dermatitis (eczema) affects approximately 20% of children worldwide.¹ In atopic dermatitis management, patient education is crucial for optimal outcomes.² The COVID-19 pandemic has impacted patient-physician interactions. To ensure safety of patients and physicians, visits may have been canceled, postponed, or conducted virtually, leaving less time for discussion and questions.³ As a consequence, patients may seek information about atopic dermatitis from alternative sources, including YouTube videos. We performed a cross-sectional study to analyze YouTube videos about topical treatments for atopic dermatitis.

During the week of July 16, 2020, we performed 4 private browser YouTube searches with default filters using the following terms: eczema topicals, eczema topical treatments, atopic dermatitis topicals, and atopic dermatitis topical treatments. For video selection, we defined topical treatments as topical corticosteroids, topical calcineurin inhibitors, crisaborole, emollients, wet wraps, and any prospective treatment topically administered. For each of the 4 searches, 2 researchers (A.M. and A.T.) independently examined the top 75 videos, yielding a total of 300 videos. Of them, 98 videos were duplicates, 19 videos were not about atopic dermatitis, and 91 videos were not about topical treatments, leaving a total of 92 videos for analysis (Figure 1).

Although less popular, professional source videos had a significantly higher DISCERN overall quality rating score (question 16) compared to those categorized as nonprofessional source (3.92 vs 1.53; P<.001). In contrast, nonprofessional source videos scored significantly higher on the quality-of-life question (question 13) compared to professional source videos (3.90 vs 2.56; P<.001)(eTable). (Three professional source videos were removed from YouTube before DISCERN scores could be assigned.)

Hepatocellular cancer (HCC) is the most common type of hepatic cancers, accounting for 65% of all hepatic cancers.¹ Among all cancers, HCC is one of the fastest growing causes of death in the United States, and the rate of new HCC cases are on the rise over several decades.² There are many risk factors leading to HCC, including alcohol use, obesity, and smoking. Infection with hepatitis C virus (HCV) poses a significant risk.¹

The pathogenesis of HCV-induced carcinogenesis is mediated by a unique host-induced immunologic response. Viral replication induces production of inflammatory factors, such as tumor necrosis factor (TNF-α), interferon (IFN), and oxidative stress on hepatocytes, resulting in cell injury, death, and regeneration. Repetitive cycles of cellular death and regeneration induce fibrosis, which may lead to cirrhosis.³ Hence, early treatment of HCV infection and achieving sustained virologic response (SVR) may lead to decreased incidence and mortality associated with HCC.

Treatment of HCV infection has become more effective with the development of direct-acting antivirals (DAAs) leading to SVR in > 90% of patients compared with 40 to 50% with IFN-based treatment.^4,5 DAAs have been proved safe and highly effective in eradicating HCV infection even in patients with advanced liver disease with decompensated cirrhosis.⁶ Although achieving SVR indicates a complete cure from chronic HCV infection, several studies have shown subsequent risk of developing HCC persists even after successful HCV treatment.^7-9 Some studies show that using DAAs to achieve SVR in patients with HCV infection leads to a decreased relative risk of HCC development compared with patients who do not receive treatment.^10-12 But data on HCC risk following DAA-induced SVR vs IFN-induced SVR are somewhat conflicting.

Much of the information regarding the association between SVR and HCC has been gleaned from large data banks without accounting for individual patient characteristics that can be obtained through full chart review. Due to small sample sizes in many chart review studies, the impact that SVR from DAA therapy has on the progression and severity of HCC is not entirely clear. The aim of our study is to evaluate the effect of HCV treatment and SVR status on overall survival (OS) in patients with HCC. Second, we aim to compare survival benefits, if any exist, among the 2 major HCV treatment modalities (IFN vs DAA).

Methods

We performed a retrospective review of patients at Memphis Veterans Affairs Medical Center (VAMC) in Tennessee to determine whether treatment for HCV infection in general, and achieving SVR in particular, makes a difference in progression, recurrence, or OS among patients with HCV infection who develop HCC. We identified 111 patients with a diagnosis of both HCV and new or recurrent HCC lesions from November 2008 to March 2019 (Table 1). We divided these patients based on their HCV treatment status, SVR status, and treatment types (IFN vs DAA).

The inclusion criteria were patients aged > 18 years treated at the Memphis VAMC who have HCV infection and developed HCC. Exclusion criteria were patients who developed HCC from other causes such as alcoholic steatohepatitis, hepatitis B virus infection, hemochromatosis, patients without HCV infection, and patients who were not established at the Memphis VAMC. This protocol was approved by the Memphis VAMC Institutional Review Board.

Statistical Analysis

OS was measured from the date of HCC diagnosis to the date of death or last follow-up. Progression-free survival (PFS) was defined from the date of HCC treatment initiation to the date of first HCC recurrence. We compared survival data for the SVR and non-SVR subgroups, the HCV treatment vs non-HCV treatment subgroups, and the IFN therapy vs DAA therapy subgroups, using the Kaplan-Meier method. The differences between subgroups were assessed using a log-rank test. Multivariate analysis using Cox proportional hazards regression model was used to identify factors that had significant impact on OS. Those factors included age; race; alcohol, tobacco, and illicit drug use; SVR status; HCV treatment status; IFN-based regimen vs DAA; MELD, and Child-Pugh scores. The results were expressed as hazard ratios (HRs) and 95% CI. Calculations were made using Statistical Analysis SAS and IBM SPSS software.

Results

The study included 111 patients. The mean age was 65.7 years; all were male and half of were Black patients. The gender imbalance was due to the predominantly male patient population at Memphis VAMC. Among 111 patients with HCV infection and HCC, 68 patients were treated for HCV infection and had significantly improved OS and PFS compared with the nontreatment group. The median 5-year OS was 44.6 months (95% CI, 966-3202) in the treated HCV infection group compared with 15.1 months in the untreated HCV infection group with a Wilcoxon P = .0005 (Figure 1). Similarly, patients treated for HCV infection had a significantly better 5-year PFS of 15.3 months (95% CI, 294-726) compared with the nontreatment group 9.5 months (95% CI, 205-405) with a Wilcoxon P = .04 (Figure 2).

Among 68 patients treated for HCV infection, 51 achieved SVR, and 34 achieved SVR after the diagnosis of HCC. Patients who achieved SVR had an improved 5-year OS when compared with patients who did not achieve SVR (median 65.8 months [95% CI, 1222-NA] vs 15.7 months [95% CI, 242-853], Wilcoxon P < .001) (Figure 3). Similarly, patients with SVR had improved 5-year PFS when compared with the non-SVR group (median 20.5 months [95% CI, 431-914] vs 8.9 months [95% CI, 191-340], Wilcoxon P = .007 (Figure 4). Achievement of SVR after HCC diagnosis suggests a significantly improved OS (HR 0.37) compared with achievement prior to HCC diagnosis (HR, 0.65; 95% CI, 0.23-1.82, P = .41)

Discussion

The survival benefit of HCV eradication and achieving SVR status has been well established in patients with HCC.¹³ In a retrospective cohort study of 250 patients with HCV infection who had received curative treatment for HCC, multivariate analysis demonstrated that achieving SVR is an independent predictor of OS.¹⁴ The 3-year and 5-year OS rates were 97% and 94% for the SVR group, and 91% and 60% for the non‐SVR group, respectively (P < .001). Similarly, according to Sou and colleagues, of 122 patients with HCV-related HCC, patients with SVR had longer OS than patients with no SVR (P = .04).¹⁵ One of the hypotheses that could explain the survival benefit in patients who achieved SVR is the effect of achieving SVR in reducing persistent liver inflammation and associated liver mortality, and therefore lowering risks of complication in patients with HCC.¹⁶ In our study, multivariate analysis shows that achieving SVR is associated with significant improved OS (HR, 0.33). In contrast, patients with HCC who have not achieved SVR are associated with worse survival (HR, 3.24). This finding supports early treatment of HCV to obtain SVR in HCV-related patients with HCC, even after development of HCC.

Among 68 patients treated for HCV infection, 45 patients were treated after HCC diagnosis, and 34 patients achieved SVR after HCC diagnosis. The average time between HCV infection treatment after HCC diagnosis was 6 months. Our data suggested that achievement of SVR after HCC diagnosis suggests an improved OS (HR, 0.37) compared with achievement prior to HCC diagnosis (HR, 0.65; 95% CI,0.23-1.82; P = .41). This lack of statistical significance is likely due to small sample size of patients achieving SVR prior to HCC diagnosis. Our results are consistent with the findings regarding the efficacy and timing of DAA treatment in patients with active HCC. According to Singal and colleagues, achieving SVR after DAA therapy may result in improved liver function and facilitate additional HCC-directed therapy, which potentially improves survival.^17-19

Nagaoki and colleagues found that there was no significant difference in OS in patients with HCC between the DAA and IFN groups. According to the study, the 3-year and 5-year OS rates were 96% and 96% for DAA patients and 93% and 73% for IFN patients, respectively (P = .16).¹⁴ This finding is consistent with the results of our study. HCV treatment type (IFN vs DAA) was not found to be associated with either OS or PFS time, regardless of time period.

Limitations

Our findings were limited in that our study population is too small to conduct further subset analysis that would allow statistical significance of those subsets, such as the suggested benefit of SVR in patients who presented with HCC after antiviral therapy. Another limitation is the all-male population, likely a result of the older veteran population at the Memphis VAMC. The mean age at diagnosis was 65 years, which is slightly higher than the general population. Compared to the SEER database, HCC is most frequently diagnosed among people aged 55 to 64 years.²³ The age difference was likely due to our aging veteran population.

Further studies are needed to determine the significance of SVR on HCC recurrence and treatment. Immunotherapy is now first-line treatment for patients with local advanced HCC. All the immunotherapy studies excluded patients with active HCV infection. Hence, we need more data on HCV treatment timing among patients scheduled to start treatment with immunotherapy.

Conclusions

In a population of older veterans, treatment of HCV infection leads to OS benefit among patients with HCC. In addition, patients with HCV infection who achieve SVR have an OS benefit over patients unable to achieve SVR. The type of treatment, DAA vs IFN-based regimen, did not show significant survival benefit.

Hepatocellular cancer (HCC) is the most common type of hepatic cancers, accounting for 65% of all hepatic cancers.¹ Among all cancers, HCC is one of the fastest growing causes of death in the United States, and the rate of new HCC cases are on the rise over several decades.² There are many risk factors leading to HCC, including alcohol use, obesity, and smoking. Infection with hepatitis C virus (HCV) poses a significant risk.¹

The pathogenesis of HCV-induced carcinogenesis is mediated by a unique host-induced immunologic response. Viral replication induces production of inflammatory factors, such as tumor necrosis factor (TNF-α), interferon (IFN), and oxidative stress on hepatocytes, resulting in cell injury, death, and regeneration. Repetitive cycles of cellular death and regeneration induce fibrosis, which may lead to cirrhosis.³ Hence, early treatment of HCV infection and achieving sustained virologic response (SVR) may lead to decreased incidence and mortality associated with HCC.

Treatment of HCV infection has become more effective with the development of direct-acting antivirals (DAAs) leading to SVR in > 90% of patients compared with 40 to 50% with IFN-based treatment.^4,5 DAAs have been proved safe and highly effective in eradicating HCV infection even in patients with advanced liver disease with decompensated cirrhosis.⁶ Although achieving SVR indicates a complete cure from chronic HCV infection, several studies have shown subsequent risk of developing HCC persists even after successful HCV treatment.^7-9 Some studies show that using DAAs to achieve SVR in patients with HCV infection leads to a decreased relative risk of HCC development compared with patients who do not receive treatment.^10-12 But data on HCC risk following DAA-induced SVR vs IFN-induced SVR are somewhat conflicting.

Much of the information regarding the association between SVR and HCC has been gleaned from large data banks without accounting for individual patient characteristics that can be obtained through full chart review. Due to small sample sizes in many chart review studies, the impact that SVR from DAA therapy has on the progression and severity of HCC is not entirely clear. The aim of our study is to evaluate the effect of HCV treatment and SVR status on overall survival (OS) in patients with HCC. Second, we aim to compare survival benefits, if any exist, among the 2 major HCV treatment modalities (IFN vs DAA).

Methods

We performed a retrospective review of patients at Memphis Veterans Affairs Medical Center (VAMC) in Tennessee to determine whether treatment for HCV infection in general, and achieving SVR in particular, makes a difference in progression, recurrence, or OS among patients with HCV infection who develop HCC. We identified 111 patients with a diagnosis of both HCV and new or recurrent HCC lesions from November 2008 to March 2019 (Table 1). We divided these patients based on their HCV treatment status, SVR status, and treatment types (IFN vs DAA).

The inclusion criteria were patients aged > 18 years treated at the Memphis VAMC who have HCV infection and developed HCC. Exclusion criteria were patients who developed HCC from other causes such as alcoholic steatohepatitis, hepatitis B virus infection, hemochromatosis, patients without HCV infection, and patients who were not established at the Memphis VAMC. This protocol was approved by the Memphis VAMC Institutional Review Board.

Statistical Analysis

OS was measured from the date of HCC diagnosis to the date of death or last follow-up. Progression-free survival (PFS) was defined from the date of HCC treatment initiation to the date of first HCC recurrence. We compared survival data for the SVR and non-SVR subgroups, the HCV treatment vs non-HCV treatment subgroups, and the IFN therapy vs DAA therapy subgroups, using the Kaplan-Meier method. The differences between subgroups were assessed using a log-rank test. Multivariate analysis using Cox proportional hazards regression model was used to identify factors that had significant impact on OS. Those factors included age; race; alcohol, tobacco, and illicit drug use; SVR status; HCV treatment status; IFN-based regimen vs DAA; MELD, and Child-Pugh scores. The results were expressed as hazard ratios (HRs) and 95% CI. Calculations were made using Statistical Analysis SAS and IBM SPSS software.

Results

The study included 111 patients. The mean age was 65.7 years; all were male and half of were Black patients. The gender imbalance was due to the predominantly male patient population at Memphis VAMC. Among 111 patients with HCV infection and HCC, 68 patients were treated for HCV infection and had significantly improved OS and PFS compared with the nontreatment group. The median 5-year OS was 44.6 months (95% CI, 966-3202) in the treated HCV infection group compared with 15.1 months in the untreated HCV infection group with a Wilcoxon P = .0005 (Figure 1). Similarly, patients treated for HCV infection had a significantly better 5-year PFS of 15.3 months (95% CI, 294-726) compared with the nontreatment group 9.5 months (95% CI, 205-405) with a Wilcoxon P = .04 (Figure 2).

Among 68 patients treated for HCV infection, 51 achieved SVR, and 34 achieved SVR after the diagnosis of HCC. Patients who achieved SVR had an improved 5-year OS when compared with patients who did not achieve SVR (median 65.8 months [95% CI, 1222-NA] vs 15.7 months [95% CI, 242-853], Wilcoxon P < .001) (Figure 3). Similarly, patients with SVR had improved 5-year PFS when compared with the non-SVR group (median 20.5 months [95% CI, 431-914] vs 8.9 months [95% CI, 191-340], Wilcoxon P = .007 (Figure 4). Achievement of SVR after HCC diagnosis suggests a significantly improved OS (HR 0.37) compared with achievement prior to HCC diagnosis (HR, 0.65; 95% CI, 0.23-1.82, P = .41)

Discussion

The survival benefit of HCV eradication and achieving SVR status has been well established in patients with HCC.¹³ In a retrospective cohort study of 250 patients with HCV infection who had received curative treatment for HCC, multivariate analysis demonstrated that achieving SVR is an independent predictor of OS.¹⁴ The 3-year and 5-year OS rates were 97% and 94% for the SVR group, and 91% and 60% for the non‐SVR group, respectively (P < .001). Similarly, according to Sou and colleagues, of 122 patients with HCV-related HCC, patients with SVR had longer OS than patients with no SVR (P = .04).¹⁵ One of the hypotheses that could explain the survival benefit in patients who achieved SVR is the effect of achieving SVR in reducing persistent liver inflammation and associated liver mortality, and therefore lowering risks of complication in patients with HCC.¹⁶ In our study, multivariate analysis shows that achieving SVR is associated with significant improved OS (HR, 0.33). In contrast, patients with HCC who have not achieved SVR are associated with worse survival (HR, 3.24). This finding supports early treatment of HCV to obtain SVR in HCV-related patients with HCC, even after development of HCC.

Among 68 patients treated for HCV infection, 45 patients were treated after HCC diagnosis, and 34 patients achieved SVR after HCC diagnosis. The average time between HCV infection treatment after HCC diagnosis was 6 months. Our data suggested that achievement of SVR after HCC diagnosis suggests an improved OS (HR, 0.37) compared with achievement prior to HCC diagnosis (HR, 0.65; 95% CI,0.23-1.82; P = .41). This lack of statistical significance is likely due to small sample size of patients achieving SVR prior to HCC diagnosis. Our results are consistent with the findings regarding the efficacy and timing of DAA treatment in patients with active HCC. According to Singal and colleagues, achieving SVR after DAA therapy may result in improved liver function and facilitate additional HCC-directed therapy, which potentially improves survival.^17-19

Nagaoki and colleagues found that there was no significant difference in OS in patients with HCC between the DAA and IFN groups. According to the study, the 3-year and 5-year OS rates were 96% and 96% for DAA patients and 93% and 73% for IFN patients, respectively (P = .16).¹⁴ This finding is consistent with the results of our study. HCV treatment type (IFN vs DAA) was not found to be associated with either OS or PFS time, regardless of time period.

Limitations

Our findings were limited in that our study population is too small to conduct further subset analysis that would allow statistical significance of those subsets, such as the suggested benefit of SVR in patients who presented with HCC after antiviral therapy. Another limitation is the all-male population, likely a result of the older veteran population at the Memphis VAMC. The mean age at diagnosis was 65 years, which is slightly higher than the general population. Compared to the SEER database, HCC is most frequently diagnosed among people aged 55 to 64 years.²³ The age difference was likely due to our aging veteran population.

Further studies are needed to determine the significance of SVR on HCC recurrence and treatment. Immunotherapy is now first-line treatment for patients with local advanced HCC. All the immunotherapy studies excluded patients with active HCV infection. Hence, we need more data on HCV treatment timing among patients scheduled to start treatment with immunotherapy.

Conclusions

In a population of older veterans, treatment of HCV infection leads to OS benefit among patients with HCC. In addition, patients with HCV infection who achieve SVR have an OS benefit over patients unable to achieve SVR. The type of treatment, DAA vs IFN-based regimen, did not show significant survival benefit.

Hepatocellular cancer (HCC) is the most common type of hepatic cancers, accounting for 65% of all hepatic cancers.¹ Among all cancers, HCC is one of the fastest growing causes of death in the United States, and the rate of new HCC cases are on the rise over several decades.² There are many risk factors leading to HCC, including alcohol use, obesity, and smoking. Infection with hepatitis C virus (HCV) poses a significant risk.¹

The pathogenesis of HCV-induced carcinogenesis is mediated by a unique host-induced immunologic response. Viral replication induces production of inflammatory factors, such as tumor necrosis factor (TNF-α), interferon (IFN), and oxidative stress on hepatocytes, resulting in cell injury, death, and regeneration. Repetitive cycles of cellular death and regeneration induce fibrosis, which may lead to cirrhosis.³ Hence, early treatment of HCV infection and achieving sustained virologic response (SVR) may lead to decreased incidence and mortality associated with HCC.

Treatment of HCV infection has become more effective with the development of direct-acting antivirals (DAAs) leading to SVR in > 90% of patients compared with 40 to 50% with IFN-based treatment.^4,5 DAAs have been proved safe and highly effective in eradicating HCV infection even in patients with advanced liver disease with decompensated cirrhosis.⁶ Although achieving SVR indicates a complete cure from chronic HCV infection, several studies have shown subsequent risk of developing HCC persists even after successful HCV treatment.^7-9 Some studies show that using DAAs to achieve SVR in patients with HCV infection leads to a decreased relative risk of HCC development compared with patients who do not receive treatment.^10-12 But data on HCC risk following DAA-induced SVR vs IFN-induced SVR are somewhat conflicting.

Much of the information regarding the association between SVR and HCC has been gleaned from large data banks without accounting for individual patient characteristics that can be obtained through full chart review. Due to small sample sizes in many chart review studies, the impact that SVR from DAA therapy has on the progression and severity of HCC is not entirely clear. The aim of our study is to evaluate the effect of HCV treatment and SVR status on overall survival (OS) in patients with HCC. Second, we aim to compare survival benefits, if any exist, among the 2 major HCV treatment modalities (IFN vs DAA).

Methods

We performed a retrospective review of patients at Memphis Veterans Affairs Medical Center (VAMC) in Tennessee to determine whether treatment for HCV infection in general, and achieving SVR in particular, makes a difference in progression, recurrence, or OS among patients with HCV infection who develop HCC. We identified 111 patients with a diagnosis of both HCV and new or recurrent HCC lesions from November 2008 to March 2019 (Table 1). We divided these patients based on their HCV treatment status, SVR status, and treatment types (IFN vs DAA).

The inclusion criteria were patients aged > 18 years treated at the Memphis VAMC who have HCV infection and developed HCC. Exclusion criteria were patients who developed HCC from other causes such as alcoholic steatohepatitis, hepatitis B virus infection, hemochromatosis, patients without HCV infection, and patients who were not established at the Memphis VAMC. This protocol was approved by the Memphis VAMC Institutional Review Board.

Statistical Analysis

OS was measured from the date of HCC diagnosis to the date of death or last follow-up. Progression-free survival (PFS) was defined from the date of HCC treatment initiation to the date of first HCC recurrence. We compared survival data for the SVR and non-SVR subgroups, the HCV treatment vs non-HCV treatment subgroups, and the IFN therapy vs DAA therapy subgroups, using the Kaplan-Meier method. The differences between subgroups were assessed using a log-rank test. Multivariate analysis using Cox proportional hazards regression model was used to identify factors that had significant impact on OS. Those factors included age; race; alcohol, tobacco, and illicit drug use; SVR status; HCV treatment status; IFN-based regimen vs DAA; MELD, and Child-Pugh scores. The results were expressed as hazard ratios (HRs) and 95% CI. Calculations were made using Statistical Analysis SAS and IBM SPSS software.

Results

The study included 111 patients. The mean age was 65.7 years; all were male and half of were Black patients. The gender imbalance was due to the predominantly male patient population at Memphis VAMC. Among 111 patients with HCV infection and HCC, 68 patients were treated for HCV infection and had significantly improved OS and PFS compared with the nontreatment group. The median 5-year OS was 44.6 months (95% CI, 966-3202) in the treated HCV infection group compared with 15.1 months in the untreated HCV infection group with a Wilcoxon P = .0005 (Figure 1). Similarly, patients treated for HCV infection had a significantly better 5-year PFS of 15.3 months (95% CI, 294-726) compared with the nontreatment group 9.5 months (95% CI, 205-405) with a Wilcoxon P = .04 (Figure 2).

Among 68 patients treated for HCV infection, 51 achieved SVR, and 34 achieved SVR after the diagnosis of HCC. Patients who achieved SVR had an improved 5-year OS when compared with patients who did not achieve SVR (median 65.8 months [95% CI, 1222-NA] vs 15.7 months [95% CI, 242-853], Wilcoxon P < .001) (Figure 3). Similarly, patients with SVR had improved 5-year PFS when compared with the non-SVR group (median 20.5 months [95% CI, 431-914] vs 8.9 months [95% CI, 191-340], Wilcoxon P = .007 (Figure 4). Achievement of SVR after HCC diagnosis suggests a significantly improved OS (HR 0.37) compared with achievement prior to HCC diagnosis (HR, 0.65; 95% CI, 0.23-1.82, P = .41)

Discussion

The survival benefit of HCV eradication and achieving SVR status has been well established in patients with HCC.¹³ In a retrospective cohort study of 250 patients with HCV infection who had received curative treatment for HCC, multivariate analysis demonstrated that achieving SVR is an independent predictor of OS.¹⁴ The 3-year and 5-year OS rates were 97% and 94% for the SVR group, and 91% and 60% for the non‐SVR group, respectively (P < .001). Similarly, according to Sou and colleagues, of 122 patients with HCV-related HCC, patients with SVR had longer OS than patients with no SVR (P = .04).¹⁵ One of the hypotheses that could explain the survival benefit in patients who achieved SVR is the effect of achieving SVR in reducing persistent liver inflammation and associated liver mortality, and therefore lowering risks of complication in patients with HCC.¹⁶ In our study, multivariate analysis shows that achieving SVR is associated with significant improved OS (HR, 0.33). In contrast, patients with HCC who have not achieved SVR are associated with worse survival (HR, 3.24). This finding supports early treatment of HCV to obtain SVR in HCV-related patients with HCC, even after development of HCC.

Among 68 patients treated for HCV infection, 45 patients were treated after HCC diagnosis, and 34 patients achieved SVR after HCC diagnosis. The average time between HCV infection treatment after HCC diagnosis was 6 months. Our data suggested that achievement of SVR after HCC diagnosis suggests an improved OS (HR, 0.37) compared with achievement prior to HCC diagnosis (HR, 0.65; 95% CI,0.23-1.82; P = .41). This lack of statistical significance is likely due to small sample size of patients achieving SVR prior to HCC diagnosis. Our results are consistent with the findings regarding the efficacy and timing of DAA treatment in patients with active HCC. According to Singal and colleagues, achieving SVR after DAA therapy may result in improved liver function and facilitate additional HCC-directed therapy, which potentially improves survival.^17-19

Nagaoki and colleagues found that there was no significant difference in OS in patients with HCC between the DAA and IFN groups. According to the study, the 3-year and 5-year OS rates were 96% and 96% for DAA patients and 93% and 73% for IFN patients, respectively (P = .16).¹⁴ This finding is consistent with the results of our study. HCV treatment type (IFN vs DAA) was not found to be associated with either OS or PFS time, regardless of time period.

Limitations

Our findings were limited in that our study population is too small to conduct further subset analysis that would allow statistical significance of those subsets, such as the suggested benefit of SVR in patients who presented with HCC after antiviral therapy. Another limitation is the all-male population, likely a result of the older veteran population at the Memphis VAMC. The mean age at diagnosis was 65 years, which is slightly higher than the general population. Compared to the SEER database, HCC is most frequently diagnosed among people aged 55 to 64 years.²³ The age difference was likely due to our aging veteran population.

Further studies are needed to determine the significance of SVR on HCC recurrence and treatment. Immunotherapy is now first-line treatment for patients with local advanced HCC. All the immunotherapy studies excluded patients with active HCV infection. Hence, we need more data on HCV treatment timing among patients scheduled to start treatment with immunotherapy.

Conclusions

In a population of older veterans, treatment of HCV infection leads to OS benefit among patients with HCC. In addition, patients with HCV infection who achieve SVR have an OS benefit over patients unable to achieve SVR. The type of treatment, DAA vs IFN-based regimen, did not show significant survival benefit.

Due to the high cost of newer chemotherapy agents, institutions search for strategies to minimize drug cost and drug waste. Programmed death-1 (PD-1) inhibitors, nivolumab and pembrolizumab, are commonly used in the treatment of solid tumors; however, the agents cost thousands of dollars per dose. Nivolumab and pembrolizumab were initially approved using weight-based dosing, but package labeling for both agents now includes fixed dosing.^1,2 A combination of these 2 dosing strategies could be used by institutions depending on individual patient’s weight to maximize cost savings, minimize drug waste, and maintain safety and efficacy of PD-1 inhibitors. Irrespective of dosing strategy, the development of immune-related adverse events (IrAEs) has been demonstrated with PD-1 inhibitors as a result of the mechanism of action.

PD-1 expression suppresses T cell activity to prevent the development of autoimmunity; however, this is also a mechanism in which tumor cells can evade the host immune system.^3-5 Binding of PD-1 and programmed death-ligand 1 (PD-L1) suppresses T cell activity, whereas the inhibition of PD-1 and PD-L1 results in T cell activation.^4,5 Increased T cell activity elicits the anticancer effect, but also contributes to the development of IrAEs^.4,5 Hypothyroidism is one of the most common IrAEs, with a reported incidence of 9% with nivolumab therapy and 8.5% with pembrolizumab.^1,2

Data from the US Department of Veterans Affairs (VA) medical centers is stored in the centralized Corporate Data Warehouse (CDW). VA researchers can obtain approval to use CDW data, which allows for large scale retrospective review of veterans who have received care at VA medical centers (VAMCs). This study aimed to describe the PD-1 inhibitor dosing used within VAMCs and identify actual and potential cost savings. Due to the frequency of immunemediated hypothyroidism and objective data that can be obtained from CDW reports, the study estimated the incidence of immune-mediated hypothyroidism within the veteran population as a safety outcome.

Background

The US Food and Drug Administration (FDA) initially approved dosing for IV nivolumab at 3 mg/kg of patient body weight every 2 weeks and for IV pembrolizumab 2 mg/kg of patient body weight every 3 weeks.^1,2 Subsequent pharmacokinetic studies found that these agents have similar exposure and efficacy with fixed doses of nivolumab 240 mg IV every 2 weeks and pembrolizumab 200 mg IV every 3 weeks; in 2016, FDA labeling shifted from weight-based dosing to fixed dosing for most solid tumor indications.^6-9 Depending on patient weight, a combination of weightbased and fixed dosing could be used by institutions to maximize cost-savings opportunities, minimize drug waste, and maintain clinical efficacy with PD-1 inhibitors. For example, a patient initiating nivolumab who weighs 80 kg would receive 240 mg for both weight-based (3 mg/kg x 80 kg = 240 mg) and fixed dosing; therefore, no cost-savings opportunities would be available. However, for a patient who weighs ≤ 73.3 kg, it would be more costeffective to use weight-based dosing vs the fixed dose. Since nivolumab is available in 40- mg, 100-mg, and 240-mg vials with similar unit prices, a combination of vial sizes could be used to minimize drug waste. Alternatively, for a patient who weighs ≥ 86.7 kg, it would be more cost-effective to administer the fixed, 240 mg dose when compared with the weightbased dose. Pembrolizumab is available only in a 100-mg vial; therefore, weight-based dosing may result in drug waste.

IrAEs can be seen with PD-1 inhibitors due to increased T cell activity, which is independent of dosing strategy and can affect any organ system. However, immune-mediated hypothyroidism has been commonly seen with PD-1 inhibitors. For patients with immunemediated hypothyroidism, levothyroxine can be considered for asymptomatic patients with thyroid- stimulating hormone (TSH) > 10 uIU/mL with normal thyroxine (T4), or patients with clinical primary hypothyroidism (TSH > 10 uIU/mL with low free T4 and clinical symptoms). Additionally, since hypothyroidism usually follows immunotherapy induced thyrotoxicosis, thyroid function tests should be monitored and levothyroxine initiated if TSH is > 10 uIU/mL for these patients.^10,11

Hypothyroidism also can be graded according to the National Cancer Institute (NCI) Common Terminology Criteria for Adverse Events. Hypothyroidism is considered grade 1 when hypothyroidism is demonstrated through clinical or diagnostic observations only and the patient is asymptomatic and no intervention needed. Grade 2 occurs when the patient is symptomatic and limits instrumental activities of daily living (ADLs), prompting thyroid replacement therapy. In grade 3, patients experience severe symptoms that restrict self-care ADLs, and hospitalization is indicated. Grade 4 has life-threatening consequences, and urgent intervention is indicated. Grade 5 results in the death of the patient.¹²

Electronic health records (EHRs) of veterans who receive care at a VAMC are stored in CDW and available through the VA Informatics and Computing Infrastructure (VINCI), which provides access to data while ensuring veterans’ privacy and data security. This feature of the VA EHR allows for analysis of data across the VA health care system, and larger data sets can be used for retrospective chart reviews.

Using reports from CDW, the primary objective of this study was to describe the dosing strategy used for PD-1 inhibitors, and the primary safety outcome was to determine the incidence of immune-mediated hypothyroidism. The secondary objective was to estimate potential cost-savings opportunities using a combination of PD-1 inhibitor dosing strategies.

Methods

This was a retrospective study including data stored in CDW. The study was approved by the Durham VA Health Care System Institutional Review Board and VINCI/Data Access request tracker. Data were limited to nivolumab and pembrolizumab because they received earlier FDA approval, had multiple solid tumor indications, and 2 FDA-approved dosing strategies. The incidence of IrAEs was limited to hypothyroidism, which could be objectively verified with laboratory monitoring of thyroid function tests, including TSH, free or total T4, and triiodothyronine (T3), all of which were available in CDW data. Additionally, most patients with hypothyroidism initiate treatment with levothyroxine. Prescription refill history could also be retrieved using CDW reports.

Hypothyroidism was defined as T4 below lower limit of normal (LLN), TSH above upper limit of normal (ULN), or any increase in levothyroxine dosage. Patients were excluded if they received PD-1 inhibitor for an indication other than solid tumor treatment, such as hematologic malignancy, or if dosing did not follow weight-based or fixed-dosing strategies, such as nivolumab 1 mg/kg when used in combination with ipilimumab, or pembrolizumab 10 mg/kg. The primary endpoint was the percentage of orders for each dosing strategy, and the primary safety outcome was the incidence of immune-mediated hypothyroidism. Secondary endpoints included estimated cost savings and cost-savings opportunities through nivolumab dose rounding and incidence of levothyroxine initiation or dose change. Descriptive statistics were used for the primary and secondary endpoints.

A report in CDW identified patients who received a dose of nivolumab or pembrolizumab between January 1, 2015 and July 1, 2017 at any VAMC. The CDW report obtained weight at time of PD-1 inhibitor therapy initiation, dose of PD-1 inhibitor given, administration date of PD-1 inhibitor, and VA site. Depending on PD-1 inhibitor administered, weight in kg was multiplied by 3 mg/kg or 2 mg/kg to obtain patient’s anticipated weight-based nivolumab and pembrolizumab dose, respectively. The calculated weight-based dose, fixed dose, and administered dose were compared to infer dosing strategy used at the time of ordering. If the patient’s weight-based dose was within 10% of the fixed dose, the order was categorized as converging because the doses were too similar to determine which dosing strategy was intended.

After determination of dosing strategy, the nivolumab orders were evaluated for actual vs missed cost savings. The cost-savings evaluation included only nivolumab orders because nivolumab is available in a 40-mg, 100-mg, and 240-mg vials and, therefore, has more potential for dose-rounding opportunities with minimal drug waste compared with pembrolizumab, which is available only in a 100-mg vial. Actual cost savings included patients who weighed ≤ 73.3 kg and received nivolumab dose based on 3 mg/kg or patients who weighed ≥ 86.7 kg and received nivolumab 240 mg (fixed dose). Missed cost savings comprised patients who weighed ≤ 73.3 kg who received 240 mg nivolumab or patients who weighed ≥ 86.7 kg and received a nivolumab dose > 240 mg. The cost difference between the dose given and theoretical cost-effective dose was calculated to determine actual and potential cost savings. Converging orders were not included in the cost-savings analysis as the intended nivolumab dose could not be determined. An additional cost analysis of nivolumab orders prescribed between September 1, 2016 and July 1, 2017 was also performed because nivolumab fixed dosing was FDA-approved for most solid tumor indications in September 2016.

To determine the incidence of immunemediated hypothyroidism for patients who received a dose of a PD-1 inhibitor at a VAMC, a CDW report with thyroid function laboratory values (TSH, T4, or T3), including reference range values based on specific VA site, and levothyroxine prescriptions issued during PD-1 inhibitor therapy was obtained. A patient was considered to have experienced immune-mediated hypothyroidism if the patient’s laboratory values demonstrated T4 below the LLN, TSH above the ULN, or if the medication fill history demonstrated levothyroxine initiation or a levothyroxine dose increase.

Results 

The CDW report identified 32,769 total PD-1 inhibitor orders. There were 3982 orders that did not meet inclusion criteria or inadequate data were obtained with CDW report and were excluded (Figure). The remaining 28,787 PD-1 inhibitor orders were evaluated for actual or missed cost savings. The distribution of dosing strategies can be found in Table 1.

Nivolumab accounted for 81.5% of all PD-1 inhibitor orders. Using the most cost-effective nivolumab dosing, the actual cost savings was estimated to be $8,514,300 with potential additional $5,591,250 of missed cost-savings opportunities. There were 8013 nivolumab orders written between September 1, 2016 and July 1, 2017. Cost-effective dosing was used in 4687 of these orders, which accounted for a cost savings of $5,198,570. The remaining 3326 orders had a missed cost-savings opportunity, which accounted for an additional $2,907,180 potential cost savings (Table 2).

PD-1 inhibitors were used for the treatment of 3249 unique patients. Based on abnormal thyroid function tests and levothyroxine initiation or dose increase, it is estimated that 514 (15.8%) patients experienced hypothyroidism during PD-1 inhibitor therapy. However, prior to PD-1 inhibitor therapy, 274 patients were receiving levothyroxine, suggesting baseline thyroid dysfunction. Of these patients, 152 (55.5%) patients maintained the same levothyroxine dose during PD-1 inhibitor therapy, but 91 (33.2%) required a levothyroxine dose increase. There were 187 patients who initiated levothyroxine during PD-1 inhibitor therapy (Table 3).

Discussion

Changes in FDA-approved dosing for PD-1 inhibitors allowed a combination of dosing strategies. Depending on patient weight, a weight-based or fixed-dosing strategy can be used to reduce drug cost while maintaining equivalent efficacy. This study evaluated use of dose rounding for PD-1 inhibitors within the VA health care system to identify actual and potential cost savings. To our knowledge, this is the first study to demonstrate cost savings through use of a combination of PD-1 inhibitor dosing strategies. Using CDW, researchers were able to review PD-1 dosing from all VAMCs and include a larger number of orders in a single retrospective study.

Nivolumab was the primary agent used within VAMCs. Depending on the indication, pembrolizumab requires PD-1 expression testing prior to its use in several solid tumor indications. Consequently, additional testing and patient eligibility is needed prior to use. Both PD-1 inhibitors were primarily dosed based on patient weight since this was the first FDAapproved dosing strategy. Nivolumab had more orders categorized as converging, which may be due to the therapeutic weight-based dose of 3 mg/kg for nivolumab vs 2 mg/kg for pembrolizumab. The calculated weight-based dose of nivolumab for an 80-kg patient is 240 mg, which also is the fixed dose. A 80-kg patient on pembrolizumab at 2 mg/kg would receive a 160-mg dose, whereas the fixed dose of pembrolizumab is 200 mg. Pembrolizumab is available only in a 100-mg vial, which limits opportunities for dose rounding without drug waste and could explain the higher amount of pembrolizumab orders in the fixed-dose category.

In this review of PD-1 inhibitor orders over approximately a 2.5-year study period, we identified $8,514,300 estimated cost savings with $5,591,250 estimated missed cost savings. When looking at orders administered after FDA approval for nivolumab-fixed dosing in September 2016, there was substantial cost savings of $5,198,570 with the potential for an additional $2,907,180 missed cost savings. Due to lower drug acquisition costs within the VA health care system, there may be higher cost-savings opportunities within other health care systems.

Through review of abnormal thyroid laboratory values and levothyroxine initiation or dose changes, this study estimated the incidence of hypothyroidism in patients receiving PD-1 inhibitor therapy at the VA. The incidence of primary hypothyroidism identified in this study was slightly higher at 15.8% compared with the 8.5 to 9.0% incidence reported from clinical trials.^1,2 There are several reasons why the incidence of hypothyroidism appeared higher in this study. Abnormal laboratory values were not assessed for the degree of deviation from the reference range; any TSH above the ULN, T4 below the LLN, or levothyroxine dose increase was included as thyroid dysfunction in this review. There is also the potential for endogenous age-related thyroid fluctuation, and the development of hypothyroidism may not have been related to PD-1 inhibitor therapy. Within this patient population, 8.4% were receiving levothyroxine prior to PD-1 inhibitor initiation indicating baseline thyroid dysfunction, and it is unclear whether levothyroxine dose increases were due to PD-1 inhibitor or endogenous fluctuation.

Limitations

There are several limitations to acknowledge. The dosing strategy and apparent dose rounding was determined by investigator inference and may not accurately represent the intended dosing strategy. This study did not address efficacy of PD-1 inhibitor and dosing strategy; however, clinical trials have demonstrated equivalent efficacy to generate the change in FDA-approved dosing. Additionally, FDA approval for nivolumab fixed dosing was indication specific. Starting in September 2016, many solid tumor indications had fixed dosing approved, but this approval was not necessarily all encompassing.

While the use of CDW allowed for a greater number of PD-1 inhibitor orders to be included in retrospective review, there also were limitations of the CDW report. The patient weight was limited to weight at time of therapy initiation. Due to the potential for weight changes, nivolumab dosing may have seemed inappropriate to investigators, and thereby excluded. Based on data available from CDW reports, hypothyroidism could not be graded according to NCI Common Terminology Criteria for Adverse Events, and the incidence of clinically significant hypothyroidism could not be determined.

Conclusions

With increasing drug acquisition costs, particularly among antineoplastic agents, health care systems frequently seek out cost-savings opportunities. Using a combination of weightbased and fixed-dosing strategies for PD-1 inhibitors can be a mechanism to achieve costsavings. Through the identification of the dosing strategy used for PD-1 inhibitors, we were able to identify and report instances for potential cost-savings opportunities among veterans treated within VA health care system. Use of CDW allows for data from all VAMCs to be evaluated in a single retrospective chart review, which allows for the inclusion of a larger sample size. This study identified a substantial cost savings for nivolumab through a combination of weight-based and fixed-dosing strategies. Due to the novel mechanism of action, ongoing realworld evaluation of adverse events and IrAEs is warranted.

Dosing strategies with nivolumab and pembrolizumab continue to evolve. In March 2018, nivolumab 480 mg IV every 4 weeks was FDA approved and in April 2020, pembrolizumab 400 mg IV every 6 weeks was FDA approved.^13,14 While the drug costs will remain the same, extended interval dosing strategies have cost avoidance such as fewer clinic appointments, resulting in decreased staffing costs and decreased patient travel. Additional studies will be needed to evaluate the cost and safety of the recently approved dosing strategies

Due to the high cost of newer chemotherapy agents, institutions search for strategies to minimize drug cost and drug waste. Programmed death-1 (PD-1) inhibitors, nivolumab and pembrolizumab, are commonly used in the treatment of solid tumors; however, the agents cost thousands of dollars per dose. Nivolumab and pembrolizumab were initially approved using weight-based dosing, but package labeling for both agents now includes fixed dosing.^1,2 A combination of these 2 dosing strategies could be used by institutions depending on individual patient’s weight to maximize cost savings, minimize drug waste, and maintain safety and efficacy of PD-1 inhibitors. Irrespective of dosing strategy, the development of immune-related adverse events (IrAEs) has been demonstrated with PD-1 inhibitors as a result of the mechanism of action.

PD-1 expression suppresses T cell activity to prevent the development of autoimmunity; however, this is also a mechanism in which tumor cells can evade the host immune system.^3-5 Binding of PD-1 and programmed death-ligand 1 (PD-L1) suppresses T cell activity, whereas the inhibition of PD-1 and PD-L1 results in T cell activation.^4,5 Increased T cell activity elicits the anticancer effect, but also contributes to the development of IrAEs^.4,5 Hypothyroidism is one of the most common IrAEs, with a reported incidence of 9% with nivolumab therapy and 8.5% with pembrolizumab.^1,2

Data from the US Department of Veterans Affairs (VA) medical centers is stored in the centralized Corporate Data Warehouse (CDW). VA researchers can obtain approval to use CDW data, which allows for large scale retrospective review of veterans who have received care at VA medical centers (VAMCs). This study aimed to describe the PD-1 inhibitor dosing used within VAMCs and identify actual and potential cost savings. Due to the frequency of immunemediated hypothyroidism and objective data that can be obtained from CDW reports, the study estimated the incidence of immune-mediated hypothyroidism within the veteran population as a safety outcome.

Background

The US Food and Drug Administration (FDA) initially approved dosing for IV nivolumab at 3 mg/kg of patient body weight every 2 weeks and for IV pembrolizumab 2 mg/kg of patient body weight every 3 weeks.^1,2 Subsequent pharmacokinetic studies found that these agents have similar exposure and efficacy with fixed doses of nivolumab 240 mg IV every 2 weeks and pembrolizumab 200 mg IV every 3 weeks; in 2016, FDA labeling shifted from weight-based dosing to fixed dosing for most solid tumor indications.^6-9 Depending on patient weight, a combination of weightbased and fixed dosing could be used by institutions to maximize cost-savings opportunities, minimize drug waste, and maintain clinical efficacy with PD-1 inhibitors. For example, a patient initiating nivolumab who weighs 80 kg would receive 240 mg for both weight-based (3 mg/kg x 80 kg = 240 mg) and fixed dosing; therefore, no cost-savings opportunities would be available. However, for a patient who weighs ≤ 73.3 kg, it would be more costeffective to use weight-based dosing vs the fixed dose. Since nivolumab is available in 40- mg, 100-mg, and 240-mg vials with similar unit prices, a combination of vial sizes could be used to minimize drug waste. Alternatively, for a patient who weighs ≥ 86.7 kg, it would be more cost-effective to administer the fixed, 240 mg dose when compared with the weightbased dose. Pembrolizumab is available only in a 100-mg vial; therefore, weight-based dosing may result in drug waste.

IrAEs can be seen with PD-1 inhibitors due to increased T cell activity, which is independent of dosing strategy and can affect any organ system. However, immune-mediated hypothyroidism has been commonly seen with PD-1 inhibitors. For patients with immunemediated hypothyroidism, levothyroxine can be considered for asymptomatic patients with thyroid- stimulating hormone (TSH) > 10 uIU/mL with normal thyroxine (T4), or patients with clinical primary hypothyroidism (TSH > 10 uIU/mL with low free T4 and clinical symptoms). Additionally, since hypothyroidism usually follows immunotherapy induced thyrotoxicosis, thyroid function tests should be monitored and levothyroxine initiated if TSH is > 10 uIU/mL for these patients.^10,11

Hypothyroidism also can be graded according to the National Cancer Institute (NCI) Common Terminology Criteria for Adverse Events. Hypothyroidism is considered grade 1 when hypothyroidism is demonstrated through clinical or diagnostic observations only and the patient is asymptomatic and no intervention needed. Grade 2 occurs when the patient is symptomatic and limits instrumental activities of daily living (ADLs), prompting thyroid replacement therapy. In grade 3, patients experience severe symptoms that restrict self-care ADLs, and hospitalization is indicated. Grade 4 has life-threatening consequences, and urgent intervention is indicated. Grade 5 results in the death of the patient.¹²

Electronic health records (EHRs) of veterans who receive care at a VAMC are stored in CDW and available through the VA Informatics and Computing Infrastructure (VINCI), which provides access to data while ensuring veterans’ privacy and data security. This feature of the VA EHR allows for analysis of data across the VA health care system, and larger data sets can be used for retrospective chart reviews.

Using reports from CDW, the primary objective of this study was to describe the dosing strategy used for PD-1 inhibitors, and the primary safety outcome was to determine the incidence of immune-mediated hypothyroidism. The secondary objective was to estimate potential cost-savings opportunities using a combination of PD-1 inhibitor dosing strategies.

Methods

This was a retrospective study including data stored in CDW. The study was approved by the Durham VA Health Care System Institutional Review Board and VINCI/Data Access request tracker. Data were limited to nivolumab and pembrolizumab because they received earlier FDA approval, had multiple solid tumor indications, and 2 FDA-approved dosing strategies. The incidence of IrAEs was limited to hypothyroidism, which could be objectively verified with laboratory monitoring of thyroid function tests, including TSH, free or total T4, and triiodothyronine (T3), all of which were available in CDW data. Additionally, most patients with hypothyroidism initiate treatment with levothyroxine. Prescription refill history could also be retrieved using CDW reports.

Hypothyroidism was defined as T4 below lower limit of normal (LLN), TSH above upper limit of normal (ULN), or any increase in levothyroxine dosage. Patients were excluded if they received PD-1 inhibitor for an indication other than solid tumor treatment, such as hematologic malignancy, or if dosing did not follow weight-based or fixed-dosing strategies, such as nivolumab 1 mg/kg when used in combination with ipilimumab, or pembrolizumab 10 mg/kg. The primary endpoint was the percentage of orders for each dosing strategy, and the primary safety outcome was the incidence of immune-mediated hypothyroidism. Secondary endpoints included estimated cost savings and cost-savings opportunities through nivolumab dose rounding and incidence of levothyroxine initiation or dose change. Descriptive statistics were used for the primary and secondary endpoints.

A report in CDW identified patients who received a dose of nivolumab or pembrolizumab between January 1, 2015 and July 1, 2017 at any VAMC. The CDW report obtained weight at time of PD-1 inhibitor therapy initiation, dose of PD-1 inhibitor given, administration date of PD-1 inhibitor, and VA site. Depending on PD-1 inhibitor administered, weight in kg was multiplied by 3 mg/kg or 2 mg/kg to obtain patient’s anticipated weight-based nivolumab and pembrolizumab dose, respectively. The calculated weight-based dose, fixed dose, and administered dose were compared to infer dosing strategy used at the time of ordering. If the patient’s weight-based dose was within 10% of the fixed dose, the order was categorized as converging because the doses were too similar to determine which dosing strategy was intended.

After determination of dosing strategy, the nivolumab orders were evaluated for actual vs missed cost savings. The cost-savings evaluation included only nivolumab orders because nivolumab is available in a 40-mg, 100-mg, and 240-mg vials and, therefore, has more potential for dose-rounding opportunities with minimal drug waste compared with pembrolizumab, which is available only in a 100-mg vial. Actual cost savings included patients who weighed ≤ 73.3 kg and received nivolumab dose based on 3 mg/kg or patients who weighed ≥ 86.7 kg and received nivolumab 240 mg (fixed dose). Missed cost savings comprised patients who weighed ≤ 73.3 kg who received 240 mg nivolumab or patients who weighed ≥ 86.7 kg and received a nivolumab dose > 240 mg. The cost difference between the dose given and theoretical cost-effective dose was calculated to determine actual and potential cost savings. Converging orders were not included in the cost-savings analysis as the intended nivolumab dose could not be determined. An additional cost analysis of nivolumab orders prescribed between September 1, 2016 and July 1, 2017 was also performed because nivolumab fixed dosing was FDA-approved for most solid tumor indications in September 2016.

To determine the incidence of immunemediated hypothyroidism for patients who received a dose of a PD-1 inhibitor at a VAMC, a CDW report with thyroid function laboratory values (TSH, T4, or T3), including reference range values based on specific VA site, and levothyroxine prescriptions issued during PD-1 inhibitor therapy was obtained. A patient was considered to have experienced immune-mediated hypothyroidism if the patient’s laboratory values demonstrated T4 below the LLN, TSH above the ULN, or if the medication fill history demonstrated levothyroxine initiation or a levothyroxine dose increase.

Results 

The CDW report identified 32,769 total PD-1 inhibitor orders. There were 3982 orders that did not meet inclusion criteria or inadequate data were obtained with CDW report and were excluded (Figure). The remaining 28,787 PD-1 inhibitor orders were evaluated for actual or missed cost savings. The distribution of dosing strategies can be found in Table 1.

Nivolumab accounted for 81.5% of all PD-1 inhibitor orders. Using the most cost-effective nivolumab dosing, the actual cost savings was estimated to be $8,514,300 with potential additional $5,591,250 of missed cost-savings opportunities. There were 8013 nivolumab orders written between September 1, 2016 and July 1, 2017. Cost-effective dosing was used in 4687 of these orders, which accounted for a cost savings of $5,198,570. The remaining 3326 orders had a missed cost-savings opportunity, which accounted for an additional $2,907,180 potential cost savings (Table 2).

PD-1 inhibitors were used for the treatment of 3249 unique patients. Based on abnormal thyroid function tests and levothyroxine initiation or dose increase, it is estimated that 514 (15.8%) patients experienced hypothyroidism during PD-1 inhibitor therapy. However, prior to PD-1 inhibitor therapy, 274 patients were receiving levothyroxine, suggesting baseline thyroid dysfunction. Of these patients, 152 (55.5%) patients maintained the same levothyroxine dose during PD-1 inhibitor therapy, but 91 (33.2%) required a levothyroxine dose increase. There were 187 patients who initiated levothyroxine during PD-1 inhibitor therapy (Table 3).

Discussion

Changes in FDA-approved dosing for PD-1 inhibitors allowed a combination of dosing strategies. Depending on patient weight, a weight-based or fixed-dosing strategy can be used to reduce drug cost while maintaining equivalent efficacy. This study evaluated use of dose rounding for PD-1 inhibitors within the VA health care system to identify actual and potential cost savings. To our knowledge, this is the first study to demonstrate cost savings through use of a combination of PD-1 inhibitor dosing strategies. Using CDW, researchers were able to review PD-1 dosing from all VAMCs and include a larger number of orders in a single retrospective study.

Nivolumab was the primary agent used within VAMCs. Depending on the indication, pembrolizumab requires PD-1 expression testing prior to its use in several solid tumor indications. Consequently, additional testing and patient eligibility is needed prior to use. Both PD-1 inhibitors were primarily dosed based on patient weight since this was the first FDAapproved dosing strategy. Nivolumab had more orders categorized as converging, which may be due to the therapeutic weight-based dose of 3 mg/kg for nivolumab vs 2 mg/kg for pembrolizumab. The calculated weight-based dose of nivolumab for an 80-kg patient is 240 mg, which also is the fixed dose. A 80-kg patient on pembrolizumab at 2 mg/kg would receive a 160-mg dose, whereas the fixed dose of pembrolizumab is 200 mg. Pembrolizumab is available only in a 100-mg vial, which limits opportunities for dose rounding without drug waste and could explain the higher amount of pembrolizumab orders in the fixed-dose category.

In this review of PD-1 inhibitor orders over approximately a 2.5-year study period, we identified $8,514,300 estimated cost savings with $5,591,250 estimated missed cost savings. When looking at orders administered after FDA approval for nivolumab-fixed dosing in September 2016, there was substantial cost savings of $5,198,570 with the potential for an additional $2,907,180 missed cost savings. Due to lower drug acquisition costs within the VA health care system, there may be higher cost-savings opportunities within other health care systems.

Through review of abnormal thyroid laboratory values and levothyroxine initiation or dose changes, this study estimated the incidence of hypothyroidism in patients receiving PD-1 inhibitor therapy at the VA. The incidence of primary hypothyroidism identified in this study was slightly higher at 15.8% compared with the 8.5 to 9.0% incidence reported from clinical trials.^1,2 There are several reasons why the incidence of hypothyroidism appeared higher in this study. Abnormal laboratory values were not assessed for the degree of deviation from the reference range; any TSH above the ULN, T4 below the LLN, or levothyroxine dose increase was included as thyroid dysfunction in this review. There is also the potential for endogenous age-related thyroid fluctuation, and the development of hypothyroidism may not have been related to PD-1 inhibitor therapy. Within this patient population, 8.4% were receiving levothyroxine prior to PD-1 inhibitor initiation indicating baseline thyroid dysfunction, and it is unclear whether levothyroxine dose increases were due to PD-1 inhibitor or endogenous fluctuation.

Limitations

There are several limitations to acknowledge. The dosing strategy and apparent dose rounding was determined by investigator inference and may not accurately represent the intended dosing strategy. This study did not address efficacy of PD-1 inhibitor and dosing strategy; however, clinical trials have demonstrated equivalent efficacy to generate the change in FDA-approved dosing. Additionally, FDA approval for nivolumab fixed dosing was indication specific. Starting in September 2016, many solid tumor indications had fixed dosing approved, but this approval was not necessarily all encompassing.

While the use of CDW allowed for a greater number of PD-1 inhibitor orders to be included in retrospective review, there also were limitations of the CDW report. The patient weight was limited to weight at time of therapy initiation. Due to the potential for weight changes, nivolumab dosing may have seemed inappropriate to investigators, and thereby excluded. Based on data available from CDW reports, hypothyroidism could not be graded according to NCI Common Terminology Criteria for Adverse Events, and the incidence of clinically significant hypothyroidism could not be determined.

Conclusions

With increasing drug acquisition costs, particularly among antineoplastic agents, health care systems frequently seek out cost-savings opportunities. Using a combination of weightbased and fixed-dosing strategies for PD-1 inhibitors can be a mechanism to achieve costsavings. Through the identification of the dosing strategy used for PD-1 inhibitors, we were able to identify and report instances for potential cost-savings opportunities among veterans treated within VA health care system. Use of CDW allows for data from all VAMCs to be evaluated in a single retrospective chart review, which allows for the inclusion of a larger sample size. This study identified a substantial cost savings for nivolumab through a combination of weight-based and fixed-dosing strategies. Due to the novel mechanism of action, ongoing realworld evaluation of adverse events and IrAEs is warranted.

Dosing strategies with nivolumab and pembrolizumab continue to evolve. In March 2018, nivolumab 480 mg IV every 4 weeks was FDA approved and in April 2020, pembrolizumab 400 mg IV every 6 weeks was FDA approved.^13,14 While the drug costs will remain the same, extended interval dosing strategies have cost avoidance such as fewer clinic appointments, resulting in decreased staffing costs and decreased patient travel. Additional studies will be needed to evaluate the cost and safety of the recently approved dosing strategies

Due to the high cost of newer chemotherapy agents, institutions search for strategies to minimize drug cost and drug waste. Programmed death-1 (PD-1) inhibitors, nivolumab and pembrolizumab, are commonly used in the treatment of solid tumors; however, the agents cost thousands of dollars per dose. Nivolumab and pembrolizumab were initially approved using weight-based dosing, but package labeling for both agents now includes fixed dosing.^1,2 A combination of these 2 dosing strategies could be used by institutions depending on individual patient’s weight to maximize cost savings, minimize drug waste, and maintain safety and efficacy of PD-1 inhibitors. Irrespective of dosing strategy, the development of immune-related adverse events (IrAEs) has been demonstrated with PD-1 inhibitors as a result of the mechanism of action.

PD-1 expression suppresses T cell activity to prevent the development of autoimmunity; however, this is also a mechanism in which tumor cells can evade the host immune system.^3-5 Binding of PD-1 and programmed death-ligand 1 (PD-L1) suppresses T cell activity, whereas the inhibition of PD-1 and PD-L1 results in T cell activation.^4,5 Increased T cell activity elicits the anticancer effect, but also contributes to the development of IrAEs^.4,5 Hypothyroidism is one of the most common IrAEs, with a reported incidence of 9% with nivolumab therapy and 8.5% with pembrolizumab.^1,2

Data from the US Department of Veterans Affairs (VA) medical centers is stored in the centralized Corporate Data Warehouse (CDW). VA researchers can obtain approval to use CDW data, which allows for large scale retrospective review of veterans who have received care at VA medical centers (VAMCs). This study aimed to describe the PD-1 inhibitor dosing used within VAMCs and identify actual and potential cost savings. Due to the frequency of immunemediated hypothyroidism and objective data that can be obtained from CDW reports, the study estimated the incidence of immune-mediated hypothyroidism within the veteran population as a safety outcome.

Background

The US Food and Drug Administration (FDA) initially approved dosing for IV nivolumab at 3 mg/kg of patient body weight every 2 weeks and for IV pembrolizumab 2 mg/kg of patient body weight every 3 weeks.^1,2 Subsequent pharmacokinetic studies found that these agents have similar exposure and efficacy with fixed doses of nivolumab 240 mg IV every 2 weeks and pembrolizumab 200 mg IV every 3 weeks; in 2016, FDA labeling shifted from weight-based dosing to fixed dosing for most solid tumor indications.^6-9 Depending on patient weight, a combination of weightbased and fixed dosing could be used by institutions to maximize cost-savings opportunities, minimize drug waste, and maintain clinical efficacy with PD-1 inhibitors. For example, a patient initiating nivolumab who weighs 80 kg would receive 240 mg for both weight-based (3 mg/kg x 80 kg = 240 mg) and fixed dosing; therefore, no cost-savings opportunities would be available. However, for a patient who weighs ≤ 73.3 kg, it would be more costeffective to use weight-based dosing vs the fixed dose. Since nivolumab is available in 40- mg, 100-mg, and 240-mg vials with similar unit prices, a combination of vial sizes could be used to minimize drug waste. Alternatively, for a patient who weighs ≥ 86.7 kg, it would be more cost-effective to administer the fixed, 240 mg dose when compared with the weightbased dose. Pembrolizumab is available only in a 100-mg vial; therefore, weight-based dosing may result in drug waste.

IrAEs can be seen with PD-1 inhibitors due to increased T cell activity, which is independent of dosing strategy and can affect any organ system. However, immune-mediated hypothyroidism has been commonly seen with PD-1 inhibitors. For patients with immunemediated hypothyroidism, levothyroxine can be considered for asymptomatic patients with thyroid- stimulating hormone (TSH) > 10 uIU/mL with normal thyroxine (T4), or patients with clinical primary hypothyroidism (TSH > 10 uIU/mL with low free T4 and clinical symptoms). Additionally, since hypothyroidism usually follows immunotherapy induced thyrotoxicosis, thyroid function tests should be monitored and levothyroxine initiated if TSH is > 10 uIU/mL for these patients.^10,11

Hypothyroidism also can be graded according to the National Cancer Institute (NCI) Common Terminology Criteria for Adverse Events. Hypothyroidism is considered grade 1 when hypothyroidism is demonstrated through clinical or diagnostic observations only and the patient is asymptomatic and no intervention needed. Grade 2 occurs when the patient is symptomatic and limits instrumental activities of daily living (ADLs), prompting thyroid replacement therapy. In grade 3, patients experience severe symptoms that restrict self-care ADLs, and hospitalization is indicated. Grade 4 has life-threatening consequences, and urgent intervention is indicated. Grade 5 results in the death of the patient.¹²

Electronic health records (EHRs) of veterans who receive care at a VAMC are stored in CDW and available through the VA Informatics and Computing Infrastructure (VINCI), which provides access to data while ensuring veterans’ privacy and data security. This feature of the VA EHR allows for analysis of data across the VA health care system, and larger data sets can be used for retrospective chart reviews.

Using reports from CDW, the primary objective of this study was to describe the dosing strategy used for PD-1 inhibitors, and the primary safety outcome was to determine the incidence of immune-mediated hypothyroidism. The secondary objective was to estimate potential cost-savings opportunities using a combination of PD-1 inhibitor dosing strategies.

Methods

This was a retrospective study including data stored in CDW. The study was approved by the Durham VA Health Care System Institutional Review Board and VINCI/Data Access request tracker. Data were limited to nivolumab and pembrolizumab because they received earlier FDA approval, had multiple solid tumor indications, and 2 FDA-approved dosing strategies. The incidence of IrAEs was limited to hypothyroidism, which could be objectively verified with laboratory monitoring of thyroid function tests, including TSH, free or total T4, and triiodothyronine (T3), all of which were available in CDW data. Additionally, most patients with hypothyroidism initiate treatment with levothyroxine. Prescription refill history could also be retrieved using CDW reports.

Hypothyroidism was defined as T4 below lower limit of normal (LLN), TSH above upper limit of normal (ULN), or any increase in levothyroxine dosage. Patients were excluded if they received PD-1 inhibitor for an indication other than solid tumor treatment, such as hematologic malignancy, or if dosing did not follow weight-based or fixed-dosing strategies, such as nivolumab 1 mg/kg when used in combination with ipilimumab, or pembrolizumab 10 mg/kg. The primary endpoint was the percentage of orders for each dosing strategy, and the primary safety outcome was the incidence of immune-mediated hypothyroidism. Secondary endpoints included estimated cost savings and cost-savings opportunities through nivolumab dose rounding and incidence of levothyroxine initiation or dose change. Descriptive statistics were used for the primary and secondary endpoints.

A report in CDW identified patients who received a dose of nivolumab or pembrolizumab between January 1, 2015 and July 1, 2017 at any VAMC. The CDW report obtained weight at time of PD-1 inhibitor therapy initiation, dose of PD-1 inhibitor given, administration date of PD-1 inhibitor, and VA site. Depending on PD-1 inhibitor administered, weight in kg was multiplied by 3 mg/kg or 2 mg/kg to obtain patient’s anticipated weight-based nivolumab and pembrolizumab dose, respectively. The calculated weight-based dose, fixed dose, and administered dose were compared to infer dosing strategy used at the time of ordering. If the patient’s weight-based dose was within 10% of the fixed dose, the order was categorized as converging because the doses were too similar to determine which dosing strategy was intended.

After determination of dosing strategy, the nivolumab orders were evaluated for actual vs missed cost savings. The cost-savings evaluation included only nivolumab orders because nivolumab is available in a 40-mg, 100-mg, and 240-mg vials and, therefore, has more potential for dose-rounding opportunities with minimal drug waste compared with pembrolizumab, which is available only in a 100-mg vial. Actual cost savings included patients who weighed ≤ 73.3 kg and received nivolumab dose based on 3 mg/kg or patients who weighed ≥ 86.7 kg and received nivolumab 240 mg (fixed dose). Missed cost savings comprised patients who weighed ≤ 73.3 kg who received 240 mg nivolumab or patients who weighed ≥ 86.7 kg and received a nivolumab dose > 240 mg. The cost difference between the dose given and theoretical cost-effective dose was calculated to determine actual and potential cost savings. Converging orders were not included in the cost-savings analysis as the intended nivolumab dose could not be determined. An additional cost analysis of nivolumab orders prescribed between September 1, 2016 and July 1, 2017 was also performed because nivolumab fixed dosing was FDA-approved for most solid tumor indications in September 2016.

To determine the incidence of immunemediated hypothyroidism for patients who received a dose of a PD-1 inhibitor at a VAMC, a CDW report with thyroid function laboratory values (TSH, T4, or T3), including reference range values based on specific VA site, and levothyroxine prescriptions issued during PD-1 inhibitor therapy was obtained. A patient was considered to have experienced immune-mediated hypothyroidism if the patient’s laboratory values demonstrated T4 below the LLN, TSH above the ULN, or if the medication fill history demonstrated levothyroxine initiation or a levothyroxine dose increase.

Results 

The CDW report identified 32,769 total PD-1 inhibitor orders. There were 3982 orders that did not meet inclusion criteria or inadequate data were obtained with CDW report and were excluded (Figure). The remaining 28,787 PD-1 inhibitor orders were evaluated for actual or missed cost savings. The distribution of dosing strategies can be found in Table 1.

Nivolumab accounted for 81.5% of all PD-1 inhibitor orders. Using the most cost-effective nivolumab dosing, the actual cost savings was estimated to be $8,514,300 with potential additional $5,591,250 of missed cost-savings opportunities. There were 8013 nivolumab orders written between September 1, 2016 and July 1, 2017. Cost-effective dosing was used in 4687 of these orders, which accounted for a cost savings of $5,198,570. The remaining 3326 orders had a missed cost-savings opportunity, which accounted for an additional $2,907,180 potential cost savings (Table 2).

PD-1 inhibitors were used for the treatment of 3249 unique patients. Based on abnormal thyroid function tests and levothyroxine initiation or dose increase, it is estimated that 514 (15.8%) patients experienced hypothyroidism during PD-1 inhibitor therapy. However, prior to PD-1 inhibitor therapy, 274 patients were receiving levothyroxine, suggesting baseline thyroid dysfunction. Of these patients, 152 (55.5%) patients maintained the same levothyroxine dose during PD-1 inhibitor therapy, but 91 (33.2%) required a levothyroxine dose increase. There were 187 patients who initiated levothyroxine during PD-1 inhibitor therapy (Table 3).

Discussion

Changes in FDA-approved dosing for PD-1 inhibitors allowed a combination of dosing strategies. Depending on patient weight, a weight-based or fixed-dosing strategy can be used to reduce drug cost while maintaining equivalent efficacy. This study evaluated use of dose rounding for PD-1 inhibitors within the VA health care system to identify actual and potential cost savings. To our knowledge, this is the first study to demonstrate cost savings through use of a combination of PD-1 inhibitor dosing strategies. Using CDW, researchers were able to review PD-1 dosing from all VAMCs and include a larger number of orders in a single retrospective study.

Nivolumab was the primary agent used within VAMCs. Depending on the indication, pembrolizumab requires PD-1 expression testing prior to its use in several solid tumor indications. Consequently, additional testing and patient eligibility is needed prior to use. Both PD-1 inhibitors were primarily dosed based on patient weight since this was the first FDAapproved dosing strategy. Nivolumab had more orders categorized as converging, which may be due to the therapeutic weight-based dose of 3 mg/kg for nivolumab vs 2 mg/kg for pembrolizumab. The calculated weight-based dose of nivolumab for an 80-kg patient is 240 mg, which also is the fixed dose. A 80-kg patient on pembrolizumab at 2 mg/kg would receive a 160-mg dose, whereas the fixed dose of pembrolizumab is 200 mg. Pembrolizumab is available only in a 100-mg vial, which limits opportunities for dose rounding without drug waste and could explain the higher amount of pembrolizumab orders in the fixed-dose category.

In this review of PD-1 inhibitor orders over approximately a 2.5-year study period, we identified $8,514,300 estimated cost savings with $5,591,250 estimated missed cost savings. When looking at orders administered after FDA approval for nivolumab-fixed dosing in September 2016, there was substantial cost savings of $5,198,570 with the potential for an additional $2,907,180 missed cost savings. Due to lower drug acquisition costs within the VA health care system, there may be higher cost-savings opportunities within other health care systems.

Through review of abnormal thyroid laboratory values and levothyroxine initiation or dose changes, this study estimated the incidence of hypothyroidism in patients receiving PD-1 inhibitor therapy at the VA. The incidence of primary hypothyroidism identified in this study was slightly higher at 15.8% compared with the 8.5 to 9.0% incidence reported from clinical trials.^1,2 There are several reasons why the incidence of hypothyroidism appeared higher in this study. Abnormal laboratory values were not assessed for the degree of deviation from the reference range; any TSH above the ULN, T4 below the LLN, or levothyroxine dose increase was included as thyroid dysfunction in this review. There is also the potential for endogenous age-related thyroid fluctuation, and the development of hypothyroidism may not have been related to PD-1 inhibitor therapy. Within this patient population, 8.4% were receiving levothyroxine prior to PD-1 inhibitor initiation indicating baseline thyroid dysfunction, and it is unclear whether levothyroxine dose increases were due to PD-1 inhibitor or endogenous fluctuation.

Limitations

There are several limitations to acknowledge. The dosing strategy and apparent dose rounding was determined by investigator inference and may not accurately represent the intended dosing strategy. This study did not address efficacy of PD-1 inhibitor and dosing strategy; however, clinical trials have demonstrated equivalent efficacy to generate the change in FDA-approved dosing. Additionally, FDA approval for nivolumab fixed dosing was indication specific. Starting in September 2016, many solid tumor indications had fixed dosing approved, but this approval was not necessarily all encompassing.

While the use of CDW allowed for a greater number of PD-1 inhibitor orders to be included in retrospective review, there also were limitations of the CDW report. The patient weight was limited to weight at time of therapy initiation. Due to the potential for weight changes, nivolumab dosing may have seemed inappropriate to investigators, and thereby excluded. Based on data available from CDW reports, hypothyroidism could not be graded according to NCI Common Terminology Criteria for Adverse Events, and the incidence of clinically significant hypothyroidism could not be determined.

Conclusions

With increasing drug acquisition costs, particularly among antineoplastic agents, health care systems frequently seek out cost-savings opportunities. Using a combination of weightbased and fixed-dosing strategies for PD-1 inhibitors can be a mechanism to achieve costsavings. Through the identification of the dosing strategy used for PD-1 inhibitors, we were able to identify and report instances for potential cost-savings opportunities among veterans treated within VA health care system. Use of CDW allows for data from all VAMCs to be evaluated in a single retrospective chart review, which allows for the inclusion of a larger sample size. This study identified a substantial cost savings for nivolumab through a combination of weight-based and fixed-dosing strategies. Due to the novel mechanism of action, ongoing realworld evaluation of adverse events and IrAEs is warranted.

Dosing strategies with nivolumab and pembrolizumab continue to evolve. In March 2018, nivolumab 480 mg IV every 4 weeks was FDA approved and in April 2020, pembrolizumab 400 mg IV every 6 weeks was FDA approved.^13,14 While the drug costs will remain the same, extended interval dosing strategies have cost avoidance such as fewer clinic appointments, resulting in decreased staffing costs and decreased patient travel. Additional studies will be needed to evaluate the cost and safety of the recently approved dosing strategies

Globally, the United States has the highest per-capita cost of health care; total costs are expected to account for approximately 20% of the nation’s gross domestic product by 2025.¹ These rising costs have prompted residency programs and medical schools to incorporate business education into their curricula.^2-5 Although medical training is demanding—with little room to add curricular components—these business-focused curricula have consistently received positive feedback from residents.^5,6

In dermatology, more than 50% of residents opt to join a private practice upon graduation.⁷ In the United States, there also is an upward trend of practice acquisition and consolidation by private equity firms. Therefore, dermatology trainees are uniquely positioned to benefit from business education to make well-informed decisions about joining or starting a practice.Furthermore, whether in a private or academic setting, knowledge of foundational economics, business strategy, finance, marketing, and health care policy can equip dermatologists to more effectively advocate for local and national policies that benefit their patient population.⁷

We conducted a survey of dermatology program directors (PDs) to determine the availability of and perceptions regarding business education during residency training.

Materials and Methods

Institutional review board (Vanderbilt University Medical Center, Nashville, Tennessee) approval was obtained. The survey was distributed weekly during a 5-week period from July 2020 to August 2020 through the Research Electronic Data Capture survey application (www.project-redcap.org). Program director email addresses were obtained through the Accreditation Council for Graduate Medical Education (ACGME) program list. A PD was included in the survey if they were employed by an accredited US osteopathic or allopathic program and their email address was provided in the ACGME program list or on their program’s faculty web page; a PD was excluded if an email address was not provided in the ACGME program list or on their program’s faculty web page.

The 8-part questionnaire was designed to assess the following characteristics: details about the respondent’s residency program (institutional affiliation, number of residents), the respondent’s professional background (number of years as a PD, business training experience), resources for business education provided by the program, the respondent’s opinion about business education for residents, and the respondent’s perception of the most important topics to include in a dermatology curriculum’s business education component, which included economics/finance, health care policy/government, management, marketing, negotiation, private equity involvement in health care, business strategy, supply chain/operations, and technology/product development. Responses were kept anonymous. Categorical and continuous variables were analyzed with medians and proportions.

Results

Of the 139 surveys distributed, 35 were completed and returned (response rate, 25.2%). Most programs were university-affiliated (71.4%) or community-affiliated (22.9%). The median number of residents was 12. The respondents had a median of 5 years’ experience in their role. Most respondents (65.7%) had no business training, although 20.0% had completed undergraduate business coursework, and 8.6% had attended formal seminars on business topics; 5.7% were self-taught on business topics.

Business Education Availability
Approximately half (51.4%) of programs offered business training to residents, primarily through seminars or lectures (94.4%) and take-home modules (16.7%). None of the programs offered a formal gap year during which residents could pursue a professional business degree. Most respondents thought business education during residency was important (82.8%) and that programs should implement more training (57.1%). When asked whether residents were competent to handle business aspects of dermatology upon graduation, most respondents disagreed somewhat (22.9%) or were neutral (40.0%).

Topics for Business Education
The most important topics identified for inclusion in a business curriculum were economics or finance (68.6%), management (68.6%), and health care policy or government (57.1%). Other identified topics included negotiation (40.0%), private equity involvement in health care (40.0%), strategy (11.4%), supply chain or operations (11.4%), marketing (2.9%), and technology (2.9%).

Comment

Residency programs and medical schools in the United States have started to integrate formal business training into their curricula; however, the state of business training in dermatology has not been characterized. Overall, this survey revealed largely positive perceptions about business education and identified a demand for more resources.

Whereas most PDs identified business education as important, only one half (51.4%) of the representative programs offered structured training. Notably, most PDs did not agree that graduating residents were competent to handle the business demands of dermatology practice. These responses highlight a gap in the demand and resources available for business training.

Identifying Curricular Resources
During an already demanding residency, additional curricular components need to be beneficial and worthwhile. To avoid significant disruption, business training could take place in the form of online lectures or take-home modules. Most programs represented in the survey responses had an academic affiliation and therefore commonly have access to an affiliated graduate business school and/or hospital administrators who have clinical and business training.

Community dermatologists who own or run their own practice also are uniquely positioned to provide residents with practical, dermatology-specific business education. Programs can utilize their institutional and local colleagues to aid in curricular design and implementation. In addition, a potential long-term solution to obtaining resources for business education is to coordinate with a national dermatology organization to create standardized modules that are available to all residency programs.

Key Curriculum Topics
Our survey identified the most important topics to include in a business curriculum for dermatology residents. Economics and finance, management, and health care policy would be valuable to a trainee regardless of whether they ultimately choose a career in academia or private practice. A thorough understanding of complex health care policy reinforces knowledge about insurance and regional and national regulations, which could ultimately benefit patient care. As an example, the American Academy of Dermatology outlines several advocacy priorities such as Medicare reimbursement policies, access to dermatologic care through public and private insurance, medication access and pricing, and preservation of private practice in the setting of market consolidation. Having a better understanding of health care policy and business could better equip dermatologists to lead these often business-driven advocacy efforts to ultimately improve patient care and advance the specialty.⁸

Limitations
There were notable limitations to this survey, primarily related to its design. With a 25% response rate, there was the potential for response and selection biases; therefore, these results might not be generalizable to all programs. In addition, views held by PDs might not be consistent with those of other members of the dermatology community; for example, surveying residents, other faculty members, and dermatologists in private practice would have provided a more comprehensive characterization of the topic.

Conclusion

This study assessed residency program directors’ perceptions of business education in dermatology training. There appears to be an imbalance between the perceived importance of such education and the resources that are available to provide it. More attention is needed to address this gap to ensure that dermatologists are prepared to manage a rapidly changing health care environment. Results of this survey should encourage efforts to establish (1) a standardized, dermatology-specific business curriculum and (2) a plan to make that curriculum accessible to trainees and other members of the dermatology community.

Globally, the United States has the highest per-capita cost of health care; total costs are expected to account for approximately 20% of the nation’s gross domestic product by 2025.¹ These rising costs have prompted residency programs and medical schools to incorporate business education into their curricula.^2-5 Although medical training is demanding—with little room to add curricular components—these business-focused curricula have consistently received positive feedback from residents.^5,6

In dermatology, more than 50% of residents opt to join a private practice upon graduation.⁷ In the United States, there also is an upward trend of practice acquisition and consolidation by private equity firms. Therefore, dermatology trainees are uniquely positioned to benefit from business education to make well-informed decisions about joining or starting a practice.Furthermore, whether in a private or academic setting, knowledge of foundational economics, business strategy, finance, marketing, and health care policy can equip dermatologists to more effectively advocate for local and national policies that benefit their patient population.⁷

We conducted a survey of dermatology program directors (PDs) to determine the availability of and perceptions regarding business education during residency training.

Materials and Methods

Institutional review board (Vanderbilt University Medical Center, Nashville, Tennessee) approval was obtained. The survey was distributed weekly during a 5-week period from July 2020 to August 2020 through the Research Electronic Data Capture survey application (www.project-redcap.org). Program director email addresses were obtained through the Accreditation Council for Graduate Medical Education (ACGME) program list. A PD was included in the survey if they were employed by an accredited US osteopathic or allopathic program and their email address was provided in the ACGME program list or on their program’s faculty web page; a PD was excluded if an email address was not provided in the ACGME program list or on their program’s faculty web page.

The 8-part questionnaire was designed to assess the following characteristics: details about the respondent’s residency program (institutional affiliation, number of residents), the respondent’s professional background (number of years as a PD, business training experience), resources for business education provided by the program, the respondent’s opinion about business education for residents, and the respondent’s perception of the most important topics to include in a dermatology curriculum’s business education component, which included economics/finance, health care policy/government, management, marketing, negotiation, private equity involvement in health care, business strategy, supply chain/operations, and technology/product development. Responses were kept anonymous. Categorical and continuous variables were analyzed with medians and proportions.

Results

Of the 139 surveys distributed, 35 were completed and returned (response rate, 25.2%). Most programs were university-affiliated (71.4%) or community-affiliated (22.9%). The median number of residents was 12. The respondents had a median of 5 years’ experience in their role. Most respondents (65.7%) had no business training, although 20.0% had completed undergraduate business coursework, and 8.6% had attended formal seminars on business topics; 5.7% were self-taught on business topics.

Business Education Availability
Approximately half (51.4%) of programs offered business training to residents, primarily through seminars or lectures (94.4%) and take-home modules (16.7%). None of the programs offered a formal gap year during which residents could pursue a professional business degree. Most respondents thought business education during residency was important (82.8%) and that programs should implement more training (57.1%). When asked whether residents were competent to handle business aspects of dermatology upon graduation, most respondents disagreed somewhat (22.9%) or were neutral (40.0%).

Topics for Business Education
The most important topics identified for inclusion in a business curriculum were economics or finance (68.6%), management (68.6%), and health care policy or government (57.1%). Other identified topics included negotiation (40.0%), private equity involvement in health care (40.0%), strategy (11.4%), supply chain or operations (11.4%), marketing (2.9%), and technology (2.9%).

Comment

Residency programs and medical schools in the United States have started to integrate formal business training into their curricula; however, the state of business training in dermatology has not been characterized. Overall, this survey revealed largely positive perceptions about business education and identified a demand for more resources.

Whereas most PDs identified business education as important, only one half (51.4%) of the representative programs offered structured training. Notably, most PDs did not agree that graduating residents were competent to handle the business demands of dermatology practice. These responses highlight a gap in the demand and resources available for business training.

Identifying Curricular Resources
During an already demanding residency, additional curricular components need to be beneficial and worthwhile. To avoid significant disruption, business training could take place in the form of online lectures or take-home modules. Most programs represented in the survey responses had an academic affiliation and therefore commonly have access to an affiliated graduate business school and/or hospital administrators who have clinical and business training.

Community dermatologists who own or run their own practice also are uniquely positioned to provide residents with practical, dermatology-specific business education. Programs can utilize their institutional and local colleagues to aid in curricular design and implementation. In addition, a potential long-term solution to obtaining resources for business education is to coordinate with a national dermatology organization to create standardized modules that are available to all residency programs.

Key Curriculum Topics
Our survey identified the most important topics to include in a business curriculum for dermatology residents. Economics and finance, management, and health care policy would be valuable to a trainee regardless of whether they ultimately choose a career in academia or private practice. A thorough understanding of complex health care policy reinforces knowledge about insurance and regional and national regulations, which could ultimately benefit patient care. As an example, the American Academy of Dermatology outlines several advocacy priorities such as Medicare reimbursement policies, access to dermatologic care through public and private insurance, medication access and pricing, and preservation of private practice in the setting of market consolidation. Having a better understanding of health care policy and business could better equip dermatologists to lead these often business-driven advocacy efforts to ultimately improve patient care and advance the specialty.⁸

Limitations
There were notable limitations to this survey, primarily related to its design. With a 25% response rate, there was the potential for response and selection biases; therefore, these results might not be generalizable to all programs. In addition, views held by PDs might not be consistent with those of other members of the dermatology community; for example, surveying residents, other faculty members, and dermatologists in private practice would have provided a more comprehensive characterization of the topic.

Conclusion

This study assessed residency program directors’ perceptions of business education in dermatology training. There appears to be an imbalance between the perceived importance of such education and the resources that are available to provide it. More attention is needed to address this gap to ensure that dermatologists are prepared to manage a rapidly changing health care environment. Results of this survey should encourage efforts to establish (1) a standardized, dermatology-specific business curriculum and (2) a plan to make that curriculum accessible to trainees and other members of the dermatology community.

Globally, the United States has the highest per-capita cost of health care; total costs are expected to account for approximately 20% of the nation’s gross domestic product by 2025.¹ These rising costs have prompted residency programs and medical schools to incorporate business education into their curricula.^2-5 Although medical training is demanding—with little room to add curricular components—these business-focused curricula have consistently received positive feedback from residents.^5,6

In dermatology, more than 50% of residents opt to join a private practice upon graduation.⁷ In the United States, there also is an upward trend of practice acquisition and consolidation by private equity firms. Therefore, dermatology trainees are uniquely positioned to benefit from business education to make well-informed decisions about joining or starting a practice.Furthermore, whether in a private or academic setting, knowledge of foundational economics, business strategy, finance, marketing, and health care policy can equip dermatologists to more effectively advocate for local and national policies that benefit their patient population.⁷

We conducted a survey of dermatology program directors (PDs) to determine the availability of and perceptions regarding business education during residency training.

Materials and Methods

Institutional review board (Vanderbilt University Medical Center, Nashville, Tennessee) approval was obtained. The survey was distributed weekly during a 5-week period from July 2020 to August 2020 through the Research Electronic Data Capture survey application (www.project-redcap.org). Program director email addresses were obtained through the Accreditation Council for Graduate Medical Education (ACGME) program list. A PD was included in the survey if they were employed by an accredited US osteopathic or allopathic program and their email address was provided in the ACGME program list or on their program’s faculty web page; a PD was excluded if an email address was not provided in the ACGME program list or on their program’s faculty web page.

The 8-part questionnaire was designed to assess the following characteristics: details about the respondent’s residency program (institutional affiliation, number of residents), the respondent’s professional background (number of years as a PD, business training experience), resources for business education provided by the program, the respondent’s opinion about business education for residents, and the respondent’s perception of the most important topics to include in a dermatology curriculum’s business education component, which included economics/finance, health care policy/government, management, marketing, negotiation, private equity involvement in health care, business strategy, supply chain/operations, and technology/product development. Responses were kept anonymous. Categorical and continuous variables were analyzed with medians and proportions.

Results

Of the 139 surveys distributed, 35 were completed and returned (response rate, 25.2%). Most programs were university-affiliated (71.4%) or community-affiliated (22.9%). The median number of residents was 12. The respondents had a median of 5 years’ experience in their role. Most respondents (65.7%) had no business training, although 20.0% had completed undergraduate business coursework, and 8.6% had attended formal seminars on business topics; 5.7% were self-taught on business topics.

Business Education Availability
Approximately half (51.4%) of programs offered business training to residents, primarily through seminars or lectures (94.4%) and take-home modules (16.7%). None of the programs offered a formal gap year during which residents could pursue a professional business degree. Most respondents thought business education during residency was important (82.8%) and that programs should implement more training (57.1%). When asked whether residents were competent to handle business aspects of dermatology upon graduation, most respondents disagreed somewhat (22.9%) or were neutral (40.0%).

Topics for Business Education
The most important topics identified for inclusion in a business curriculum were economics or finance (68.6%), management (68.6%), and health care policy or government (57.1%). Other identified topics included negotiation (40.0%), private equity involvement in health care (40.0%), strategy (11.4%), supply chain or operations (11.4%), marketing (2.9%), and technology (2.9%).

Comment

Residency programs and medical schools in the United States have started to integrate formal business training into their curricula; however, the state of business training in dermatology has not been characterized. Overall, this survey revealed largely positive perceptions about business education and identified a demand for more resources.

Whereas most PDs identified business education as important, only one half (51.4%) of the representative programs offered structured training. Notably, most PDs did not agree that graduating residents were competent to handle the business demands of dermatology practice. These responses highlight a gap in the demand and resources available for business training.

Identifying Curricular Resources
During an already demanding residency, additional curricular components need to be beneficial and worthwhile. To avoid significant disruption, business training could take place in the form of online lectures or take-home modules. Most programs represented in the survey responses had an academic affiliation and therefore commonly have access to an affiliated graduate business school and/or hospital administrators who have clinical and business training.

Community dermatologists who own or run their own practice also are uniquely positioned to provide residents with practical, dermatology-specific business education. Programs can utilize their institutional and local colleagues to aid in curricular design and implementation. In addition, a potential long-term solution to obtaining resources for business education is to coordinate with a national dermatology organization to create standardized modules that are available to all residency programs.

Key Curriculum Topics
Our survey identified the most important topics to include in a business curriculum for dermatology residents. Economics and finance, management, and health care policy would be valuable to a trainee regardless of whether they ultimately choose a career in academia or private practice. A thorough understanding of complex health care policy reinforces knowledge about insurance and regional and national regulations, which could ultimately benefit patient care. As an example, the American Academy of Dermatology outlines several advocacy priorities such as Medicare reimbursement policies, access to dermatologic care through public and private insurance, medication access and pricing, and preservation of private practice in the setting of market consolidation. Having a better understanding of health care policy and business could better equip dermatologists to lead these often business-driven advocacy efforts to ultimately improve patient care and advance the specialty.⁸

Limitations
There were notable limitations to this survey, primarily related to its design. With a 25% response rate, there was the potential for response and selection biases; therefore, these results might not be generalizable to all programs. In addition, views held by PDs might not be consistent with those of other members of the dermatology community; for example, surveying residents, other faculty members, and dermatologists in private practice would have provided a more comprehensive characterization of the topic.

Conclusion

This study assessed residency program directors’ perceptions of business education in dermatology training. There appears to be an imbalance between the perceived importance of such education and the resources that are available to provide it. More attention is needed to address this gap to ensure that dermatologists are prepared to manage a rapidly changing health care environment. Results of this survey should encourage efforts to establish (1) a standardized, dermatology-specific business curriculum and (2) a plan to make that curriculum accessible to trainees and other members of the dermatology community.

Biologic medications are highly effective in treating moderate to severe psoriasis, yet many patients are apprehensive about taking a biologic medication for a variety of reasons, such as hearing negative information about the drug from friends or family, being nervous about injection, or seeing the drug or its side effects negatively portrayed in the media.^1-3 Because biologic medications are costly, many patients may fear needing to discontinue use of the medication owing to lack of affordability, which may result in subsequent rebound of psoriasis. Because patients’ fear of a drug is inherently subjective, it can be modified with appropriate reassurance and presentation of evidence. By understanding what information increases patients’ confidence in their willingness to take a biologic medication, patients may be more willing to initiate use of the drug and improve treatment outcomes.

There are mixed findings about whether statistical evidence or an anecdote is more effective in persuasion.^4-6 The specific context in which the persuasion takes place may be important in determining which method is superior. In most nonthreatening situations, people appear to be more easily persuaded by statistical evidence rather than an anecdote. However, in circumstances where emotional engagement is high, such as regarding one’s own health, an anecdote tends to be more persuasive compared to statistical evidence.⁷ The purpose of this study was to evaluate patients’ willingness to take a biologic medication for the management of their psoriasis if presented with either clinical trial evidence of the agent’s efficacy and safety, an anecdote of a single patient’s positive experience, or both.

Methods

Patient Inclusion Criteria
Following Wake Forest School of Medicine institutional review board approval, a prospective parallel-arm survey study was performed on eligible patients 18 years or older with a self-reported diagnosis of psoriasis. Patients were required to have a working knowledge of English and not have been previously prescribed a biologic medication for their psoriasis. If patients did not meet inclusion criteria after answering the survey eligibility screening questions, then they were unable to complete the remainder of the survey and were excluded from the analysis.

Survey Administration
A total of 222 patients were recruited through Amazon Mechanical Turk, an online crowdsourcing platform. (Amazon Mechanical Turk is a validated tool in conducting research in psychology and other social sciences and is considered as diverse as and perhaps more representative than traditional samples.^8,9) Patients received a fact sheet and were taken to the survey hosted on Qualtrics, a secure web-based survey software that supports data collection for research studies. Amazon Mechanical Turk requires some amount of compensation to patients; therefore, recruited patients were compensated $0.03.

Statistical Analysis
Patients were randomized using SPSS Statistics version 23.0 (IBM) in a 1:1 ratio to assess how willing they would be to take a biologic medication for their psoriasis if presented with one of the following: (1) a control that queried patients about their willingness to take treatment without having been informed on its efficacy or safety, (2) clinical trial evidence of the agent’s efficacy and safety, (3) an anecdote of a single patient’s positive experience, or (4) both clinical trial evidence of the agent’s efficacy and safety and an anecdote of a single patient’s positive experience (Table 1). Demographic information including sex, age, ethnicity, and education level was collected, in addition to other baseline characteristics such as having friends or family with a history of psoriasis, history of participation in a clinical trial with use of an experimental drug, and the number of years since clinical diagnosis of psoriasis.

Results

There were no statistically significant differences among the baseline characteristics of the 4 information assignment groups (Table 2). Patients in the control group not given either clinical trial evidence of a biologic medication’s efficacy and safety or anecdote of a single patient’s positive experience had the lowest reported willingness to take treatment (median, 4.0)(Figure). Patients in the group given clinical trial evidence had an intermediate level of reported willingness to take treatment (median, 7.0). Patients in the groups given an anecdote or clinical trial evidence and anecdote had the highest reported willingness to take treatment (median, 8.0). Patients presented with an anecdote or clinical trial evidence and anecdote were significantly more likely to report willingness to take treatment than those not presented with either clinical trial evidence or anecdote (P<.001). Although patients presented with an anecdote or clinical trial evidence and anecdote were more likely to take a biologic medication than those presented with only clinical trial evidence, the difference between these groups was not significant (P=.75).

Based on regression analysis, age, sex, and having friends or family with a history of psoriasis were not significantly associated with patients’ responses (eTable). The number of years since clinical diagnosis of psoriasis (P=.034) and history of participation in a clinical trial with use of an experimental drug (P=.018) were significantly associated with the willingness of patients presented with an anecdote to take a biologic medication.

Comment

Anecdotal Reassurance
The presentation of clinical trial and/or anecdotal evidence had a strong effect on patients’ willingness to take a biologic medication for their psoriasis. Human perception of a treatment is inherently subjective, and such perceptions can be modified with appropriate reassurance and presentation of evidence.¹ Across the population we studied, presenting a brief anecdote of a single patient’s positive experience is a quick and efficient means—and as or more effective as giving details on efficacy and safety—to help patients decide to take a treatment for their psoriasis.

Anecdotal reassurance is powerful. Both health care providers and patients have a natural tendency to focus on anecdotal experiences rather than statistical reasoning when making treatment decisions.^10-12 Although negative anecdotal experiences may make patients unwilling to take a medication (or may make them overly desirous of an inappropriate treatment), clinicians can harness this psychological phenomenon to both increase patient willingness to take potentially beneficial treatments or to deter them from engaging in activities that can be harmful to their health, such as tanning and smoking.

Psoriasis Duration and Willingness to Take a Biologic Medication
In general, patient demographics did not appear to have an association with reported willingness to take a biologic medication for psoriasis. However, the number of years since clinical diagnosis of psoriasis had an effect on willingness to take a biologic medication, with patients with a longer personal history of psoriasis showing a higher willingness to take a treatment after being presented with an anecdote than patients with a shorter personal history of psoriasis. We can only speculate on the reasons why. Patients with a longer personal history of psoriasis may have tried and failed more treatments and therefore have a distrust in the validity of clinical trial evidence. These patients may feel their psoriasis is different than that of other clinical trial participants and thus may be more willing to rely on the success stories of individual patients.

Prior participation in a clinical trial with use of an experimental drug was associated with a lower willingness to choose treatment after being presented with anecdotal reassurance. This finding may be attributable to these patients understanding the subjective nature of anecdotes and preferring more objective information in the form of randomized clinical trials in making treatment decisions. Overall, the presentation of evidence about the efficacy and safety of biologic medications in the treatment of psoriasis has a greater impact on patient decision-making than patients’ age, sex, and having friends or family with a history of psoriasis.

Limitations
Limitations of the study were typical of survey-based research. With closed-ended questions, patients were not able to explain their responses. In addition, hypothetical informational statements of a biologic’s efficacy and safety may not always imitate clinical reality. However, we believe the study is valid in exploring the power of an anecdote in influencing patients’ willingness to take biologic medications for psoriasis. Furthermore, educational level and ethnicity were excluded from the ordinal regression analysis because the assumption of parallel lines was not met.

Ethics Behind an Anecdote
An important consideration is the ethical implications of sharing an anecdote to guide patients’ perceptions of treatment and behavior. Although clinicians rely heavily on the available data to determine the best course of treatment, providing patients with comprehensive information on all risks and benefits is rarely, if ever, feasible. Moreover, even objective clinical data will inevitably be subjectively interpreted by patients. For example, describing a medication side effect as occurring in 1 in 100 patients may discourage patients from pursuing treatment, whereas describing that risk as not occurring in 99 in 100 patients may encourage patients, despite these 2 choices being mathematically identical.¹³ Because the subjective interpretation of data is inevitable, presenting patients with subjective information in the form of an anecdote to help them overcome fears of starting treatment and achieve their desired clinical outcomes may be one of the appropriate approaches to present what is objectively the best option, particularly if the anecdote is representative of the expected treatment response. Clinicians can harness this understanding of human psychology to better educate patients about their treatment options while fulfilling their ethical duty to act in their patients’ best interest.

Conclusion

Using an anecdote to help patients overcome fears of starting a biologic medication may be appropriate if the anecdote is reasonably representative of an expected treatment outcome. Patients should have an accurate understanding of the common risks and benefits of a medication for purposes of shared decision-making.

Biologic medications are highly effective in treating moderate to severe psoriasis, yet many patients are apprehensive about taking a biologic medication for a variety of reasons, such as hearing negative information about the drug from friends or family, being nervous about injection, or seeing the drug or its side effects negatively portrayed in the media.^1-3 Because biologic medications are costly, many patients may fear needing to discontinue use of the medication owing to lack of affordability, which may result in subsequent rebound of psoriasis. Because patients’ fear of a drug is inherently subjective, it can be modified with appropriate reassurance and presentation of evidence. By understanding what information increases patients’ confidence in their willingness to take a biologic medication, patients may be more willing to initiate use of the drug and improve treatment outcomes.

There are mixed findings about whether statistical evidence or an anecdote is more effective in persuasion.^4-6 The specific context in which the persuasion takes place may be important in determining which method is superior. In most nonthreatening situations, people appear to be more easily persuaded by statistical evidence rather than an anecdote. However, in circumstances where emotional engagement is high, such as regarding one’s own health, an anecdote tends to be more persuasive compared to statistical evidence.⁷ The purpose of this study was to evaluate patients’ willingness to take a biologic medication for the management of their psoriasis if presented with either clinical trial evidence of the agent’s efficacy and safety, an anecdote of a single patient’s positive experience, or both.

Methods

Patient Inclusion Criteria
Following Wake Forest School of Medicine institutional review board approval, a prospective parallel-arm survey study was performed on eligible patients 18 years or older with a self-reported diagnosis of psoriasis. Patients were required to have a working knowledge of English and not have been previously prescribed a biologic medication for their psoriasis. If patients did not meet inclusion criteria after answering the survey eligibility screening questions, then they were unable to complete the remainder of the survey and were excluded from the analysis.

Survey Administration
A total of 222 patients were recruited through Amazon Mechanical Turk, an online crowdsourcing platform. (Amazon Mechanical Turk is a validated tool in conducting research in psychology and other social sciences and is considered as diverse as and perhaps more representative than traditional samples.^8,9) Patients received a fact sheet and were taken to the survey hosted on Qualtrics, a secure web-based survey software that supports data collection for research studies. Amazon Mechanical Turk requires some amount of compensation to patients; therefore, recruited patients were compensated $0.03.

Statistical Analysis
Patients were randomized using SPSS Statistics version 23.0 (IBM) in a 1:1 ratio to assess how willing they would be to take a biologic medication for their psoriasis if presented with one of the following: (1) a control that queried patients about their willingness to take treatment without having been informed on its efficacy or safety, (2) clinical trial evidence of the agent’s efficacy and safety, (3) an anecdote of a single patient’s positive experience, or (4) both clinical trial evidence of the agent’s efficacy and safety and an anecdote of a single patient’s positive experience (Table 1). Demographic information including sex, age, ethnicity, and education level was collected, in addition to other baseline characteristics such as having friends or family with a history of psoriasis, history of participation in a clinical trial with use of an experimental drug, and the number of years since clinical diagnosis of psoriasis.

Results

There were no statistically significant differences among the baseline characteristics of the 4 information assignment groups (Table 2). Patients in the control group not given either clinical trial evidence of a biologic medication’s efficacy and safety or anecdote of a single patient’s positive experience had the lowest reported willingness to take treatment (median, 4.0)(Figure). Patients in the group given clinical trial evidence had an intermediate level of reported willingness to take treatment (median, 7.0). Patients in the groups given an anecdote or clinical trial evidence and anecdote had the highest reported willingness to take treatment (median, 8.0). Patients presented with an anecdote or clinical trial evidence and anecdote were significantly more likely to report willingness to take treatment than those not presented with either clinical trial evidence or anecdote (P<.001). Although patients presented with an anecdote or clinical trial evidence and anecdote were more likely to take a biologic medication than those presented with only clinical trial evidence, the difference between these groups was not significant (P=.75).

Based on regression analysis, age, sex, and having friends or family with a history of psoriasis were not significantly associated with patients’ responses (eTable). The number of years since clinical diagnosis of psoriasis (P=.034) and history of participation in a clinical trial with use of an experimental drug (P=.018) were significantly associated with the willingness of patients presented with an anecdote to take a biologic medication.

Comment

Anecdotal Reassurance
The presentation of clinical trial and/or anecdotal evidence had a strong effect on patients’ willingness to take a biologic medication for their psoriasis. Human perception of a treatment is inherently subjective, and such perceptions can be modified with appropriate reassurance and presentation of evidence.¹ Across the population we studied, presenting a brief anecdote of a single patient’s positive experience is a quick and efficient means—and as or more effective as giving details on efficacy and safety—to help patients decide to take a treatment for their psoriasis.

Anecdotal reassurance is powerful. Both health care providers and patients have a natural tendency to focus on anecdotal experiences rather than statistical reasoning when making treatment decisions.^10-12 Although negative anecdotal experiences may make patients unwilling to take a medication (or may make them overly desirous of an inappropriate treatment), clinicians can harness this psychological phenomenon to both increase patient willingness to take potentially beneficial treatments or to deter them from engaging in activities that can be harmful to their health, such as tanning and smoking.

Psoriasis Duration and Willingness to Take a Biologic Medication
In general, patient demographics did not appear to have an association with reported willingness to take a biologic medication for psoriasis. However, the number of years since clinical diagnosis of psoriasis had an effect on willingness to take a biologic medication, with patients with a longer personal history of psoriasis showing a higher willingness to take a treatment after being presented with an anecdote than patients with a shorter personal history of psoriasis. We can only speculate on the reasons why. Patients with a longer personal history of psoriasis may have tried and failed more treatments and therefore have a distrust in the validity of clinical trial evidence. These patients may feel their psoriasis is different than that of other clinical trial participants and thus may be more willing to rely on the success stories of individual patients.

Prior participation in a clinical trial with use of an experimental drug was associated with a lower willingness to choose treatment after being presented with anecdotal reassurance. This finding may be attributable to these patients understanding the subjective nature of anecdotes and preferring more objective information in the form of randomized clinical trials in making treatment decisions. Overall, the presentation of evidence about the efficacy and safety of biologic medications in the treatment of psoriasis has a greater impact on patient decision-making than patients’ age, sex, and having friends or family with a history of psoriasis.

Limitations
Limitations of the study were typical of survey-based research. With closed-ended questions, patients were not able to explain their responses. In addition, hypothetical informational statements of a biologic’s efficacy and safety may not always imitate clinical reality. However, we believe the study is valid in exploring the power of an anecdote in influencing patients’ willingness to take biologic medications for psoriasis. Furthermore, educational level and ethnicity were excluded from the ordinal regression analysis because the assumption of parallel lines was not met.

Ethics Behind an Anecdote
An important consideration is the ethical implications of sharing an anecdote to guide patients’ perceptions of treatment and behavior. Although clinicians rely heavily on the available data to determine the best course of treatment, providing patients with comprehensive information on all risks and benefits is rarely, if ever, feasible. Moreover, even objective clinical data will inevitably be subjectively interpreted by patients. For example, describing a medication side effect as occurring in 1 in 100 patients may discourage patients from pursuing treatment, whereas describing that risk as not occurring in 99 in 100 patients may encourage patients, despite these 2 choices being mathematically identical.¹³ Because the subjective interpretation of data is inevitable, presenting patients with subjective information in the form of an anecdote to help them overcome fears of starting treatment and achieve their desired clinical outcomes may be one of the appropriate approaches to present what is objectively the best option, particularly if the anecdote is representative of the expected treatment response. Clinicians can harness this understanding of human psychology to better educate patients about their treatment options while fulfilling their ethical duty to act in their patients’ best interest.

Conclusion

Using an anecdote to help patients overcome fears of starting a biologic medication may be appropriate if the anecdote is reasonably representative of an expected treatment outcome. Patients should have an accurate understanding of the common risks and benefits of a medication for purposes of shared decision-making.

Biologic medications are highly effective in treating moderate to severe psoriasis, yet many patients are apprehensive about taking a biologic medication for a variety of reasons, such as hearing negative information about the drug from friends or family, being nervous about injection, or seeing the drug or its side effects negatively portrayed in the media.^1-3 Because biologic medications are costly, many patients may fear needing to discontinue use of the medication owing to lack of affordability, which may result in subsequent rebound of psoriasis. Because patients’ fear of a drug is inherently subjective, it can be modified with appropriate reassurance and presentation of evidence. By understanding what information increases patients’ confidence in their willingness to take a biologic medication, patients may be more willing to initiate use of the drug and improve treatment outcomes.

There are mixed findings about whether statistical evidence or an anecdote is more effective in persuasion.^4-6 The specific context in which the persuasion takes place may be important in determining which method is superior. In most nonthreatening situations, people appear to be more easily persuaded by statistical evidence rather than an anecdote. However, in circumstances where emotional engagement is high, such as regarding one’s own health, an anecdote tends to be more persuasive compared to statistical evidence.⁷ The purpose of this study was to evaluate patients’ willingness to take a biologic medication for the management of their psoriasis if presented with either clinical trial evidence of the agent’s efficacy and safety, an anecdote of a single patient’s positive experience, or both.

Methods

Patient Inclusion Criteria
Following Wake Forest School of Medicine institutional review board approval, a prospective parallel-arm survey study was performed on eligible patients 18 years or older with a self-reported diagnosis of psoriasis. Patients were required to have a working knowledge of English and not have been previously prescribed a biologic medication for their psoriasis. If patients did not meet inclusion criteria after answering the survey eligibility screening questions, then they were unable to complete the remainder of the survey and were excluded from the analysis.

Survey Administration
A total of 222 patients were recruited through Amazon Mechanical Turk, an online crowdsourcing platform. (Amazon Mechanical Turk is a validated tool in conducting research in psychology and other social sciences and is considered as diverse as and perhaps more representative than traditional samples.^8,9) Patients received a fact sheet and were taken to the survey hosted on Qualtrics, a secure web-based survey software that supports data collection for research studies. Amazon Mechanical Turk requires some amount of compensation to patients; therefore, recruited patients were compensated $0.03.

Statistical Analysis
Patients were randomized using SPSS Statistics version 23.0 (IBM) in a 1:1 ratio to assess how willing they would be to take a biologic medication for their psoriasis if presented with one of the following: (1) a control that queried patients about their willingness to take treatment without having been informed on its efficacy or safety, (2) clinical trial evidence of the agent’s efficacy and safety, (3) an anecdote of a single patient’s positive experience, or (4) both clinical trial evidence of the agent’s efficacy and safety and an anecdote of a single patient’s positive experience (Table 1). Demographic information including sex, age, ethnicity, and education level was collected, in addition to other baseline characteristics such as having friends or family with a history of psoriasis, history of participation in a clinical trial with use of an experimental drug, and the number of years since clinical diagnosis of psoriasis.

Results

There were no statistically significant differences among the baseline characteristics of the 4 information assignment groups (Table 2). Patients in the control group not given either clinical trial evidence of a biologic medication’s efficacy and safety or anecdote of a single patient’s positive experience had the lowest reported willingness to take treatment (median, 4.0)(Figure). Patients in the group given clinical trial evidence had an intermediate level of reported willingness to take treatment (median, 7.0). Patients in the groups given an anecdote or clinical trial evidence and anecdote had the highest reported willingness to take treatment (median, 8.0). Patients presented with an anecdote or clinical trial evidence and anecdote were significantly more likely to report willingness to take treatment than those not presented with either clinical trial evidence or anecdote (P<.001). Although patients presented with an anecdote or clinical trial evidence and anecdote were more likely to take a biologic medication than those presented with only clinical trial evidence, the difference between these groups was not significant (P=.75).

Based on regression analysis, age, sex, and having friends or family with a history of psoriasis were not significantly associated with patients’ responses (eTable). The number of years since clinical diagnosis of psoriasis (P=.034) and history of participation in a clinical trial with use of an experimental drug (P=.018) were significantly associated with the willingness of patients presented with an anecdote to take a biologic medication.

Comment

Anecdotal Reassurance
The presentation of clinical trial and/or anecdotal evidence had a strong effect on patients’ willingness to take a biologic medication for their psoriasis. Human perception of a treatment is inherently subjective, and such perceptions can be modified with appropriate reassurance and presentation of evidence.¹ Across the population we studied, presenting a brief anecdote of a single patient’s positive experience is a quick and efficient means—and as or more effective as giving details on efficacy and safety—to help patients decide to take a treatment for their psoriasis.

Anecdotal reassurance is powerful. Both health care providers and patients have a natural tendency to focus on anecdotal experiences rather than statistical reasoning when making treatment decisions.^10-12 Although negative anecdotal experiences may make patients unwilling to take a medication (or may make them overly desirous of an inappropriate treatment), clinicians can harness this psychological phenomenon to both increase patient willingness to take potentially beneficial treatments or to deter them from engaging in activities that can be harmful to their health, such as tanning and smoking.

Psoriasis Duration and Willingness to Take a Biologic Medication
In general, patient demographics did not appear to have an association with reported willingness to take a biologic medication for psoriasis. However, the number of years since clinical diagnosis of psoriasis had an effect on willingness to take a biologic medication, with patients with a longer personal history of psoriasis showing a higher willingness to take a treatment after being presented with an anecdote than patients with a shorter personal history of psoriasis. We can only speculate on the reasons why. Patients with a longer personal history of psoriasis may have tried and failed more treatments and therefore have a distrust in the validity of clinical trial evidence. These patients may feel their psoriasis is different than that of other clinical trial participants and thus may be more willing to rely on the success stories of individual patients.

Prior participation in a clinical trial with use of an experimental drug was associated with a lower willingness to choose treatment after being presented with anecdotal reassurance. This finding may be attributable to these patients understanding the subjective nature of anecdotes and preferring more objective information in the form of randomized clinical trials in making treatment decisions. Overall, the presentation of evidence about the efficacy and safety of biologic medications in the treatment of psoriasis has a greater impact on patient decision-making than patients’ age, sex, and having friends or family with a history of psoriasis.

Limitations
Limitations of the study were typical of survey-based research. With closed-ended questions, patients were not able to explain their responses. In addition, hypothetical informational statements of a biologic’s efficacy and safety may not always imitate clinical reality. However, we believe the study is valid in exploring the power of an anecdote in influencing patients’ willingness to take biologic medications for psoriasis. Furthermore, educational level and ethnicity were excluded from the ordinal regression analysis because the assumption of parallel lines was not met.

Ethics Behind an Anecdote
An important consideration is the ethical implications of sharing an anecdote to guide patients’ perceptions of treatment and behavior. Although clinicians rely heavily on the available data to determine the best course of treatment, providing patients with comprehensive information on all risks and benefits is rarely, if ever, feasible. Moreover, even objective clinical data will inevitably be subjectively interpreted by patients. For example, describing a medication side effect as occurring in 1 in 100 patients may discourage patients from pursuing treatment, whereas describing that risk as not occurring in 99 in 100 patients may encourage patients, despite these 2 choices being mathematically identical.¹³ Because the subjective interpretation of data is inevitable, presenting patients with subjective information in the form of an anecdote to help them overcome fears of starting treatment and achieve their desired clinical outcomes may be one of the appropriate approaches to present what is objectively the best option, particularly if the anecdote is representative of the expected treatment response. Clinicians can harness this understanding of human psychology to better educate patients about their treatment options while fulfilling their ethical duty to act in their patients’ best interest.

Conclusion

Using an anecdote to help patients overcome fears of starting a biologic medication may be appropriate if the anecdote is reasonably representative of an expected treatment outcome. Patients should have an accurate understanding of the common risks and benefits of a medication for purposes of shared decision-making.

Nearly 50,000 veterans are diagnosed with cancer within the Veterans Health Administration annually with prostate cancer (PC) being the most frequently diagnosed, accounting for 29% of all cancers diagnosed.¹ The treatment of PC depends on the stage and risk group at presentation and patient preference. Men with early stage, localized PC can be managed with prostatectomy, radiation therapy, or active surveillance.²

Within the Veterans Health Administration, more patients are treated with radiation therapy than with radical prostatectomy.³ This is in contrast to the civil health system, where more patients are treated with radical prostatectomy than with radiation therapy.^4,5 Radiation therapy for PC can be given externally with external beam radiation therapy or internally with brachytherapy (BT). BT is categorized by the rate at which the radiation dose is delivered and generally grouped as low-dose rate (LDR) or high-dose rate (HDR). LDRBT consists of permanently implanting radioactive seeds, which slowly deliver a radiation dose over an extended period. HDRBT consists of implanting catheters that allow delivery of a radioactive source to be placed temporarily in the prostate and removed after treatment. The utilization of HDRBT has become more common as treatment has evolved to consist of fewer, larger fractions in a shorter time, making it a convenient treatment option for men with PC.⁶ The veteran population has singular medical challenges. These patients differ from the general population and are often underrepresented in medical research and published studies.⁷ There are no studies exploring the treatment-associated toxicities from HDRBT treatment for PC specifically in the veteran population. The objective of this study is to report our findings regarding the veteran-reported and physician-graded toxicities associated with HDRBT as monotherapy in veterans treated through the US Department of Veterans Affairs (VA) for PC.

Methods

We performed a retrospective cohort study of a prospectively maintained, institutional review board-approved database of patients treated with HDRBT for PC. Veterans were seen in consultation at Edward Hines, Jr. VA Hospital (EHJVAH) in Hines, Illinois. This is the only VA hospital in Illinois that offers radiation therapy, so it acted as a tertiary center, receiving referrals from other, neighboring VA hospitals. If the veteran was deemed a good BT candidate and elected to proceed with HDRBT, HDR treatment was performed at a partnering academic institution equipped to provide HDRBT (Loyola University Medical Center).

We selected patients with National Cancer Center Network (NCCN) low- or intermediate-risk PC undergoing definitive HDRBT as monotherapy using 13.5 Gy x 2 fractions delivered over 2 implants that were 1 to 2 weeks apart. Patients who received androgen deprivation therapy (ADT) were excluded from this study. No patients received supplemental external beam radiation. Men with unfavorable intermediate risk PC were offered ADT and BT in accordance with NCCN guidelines. However, patients with unfavorable intermediate-risk PC who declined ADT or who were deemed poor ADT candidates due to comorbidities were treated with HDR as monotherapy and included in this study.⁸

HDR Treatment

Our HDRBT implant procedure and treatment planning details have been previously described.⁹ In brief, patients were implanted with between 17 and 22 catheters based on gland size under transrectal ultrasound guidance. After implantation, computed tomography and, when possible, magnetic resonance imaging of the prostate were obtained and registered for target delineation. The prostate was segmented, and an asymmetric planning target volume of 0 to 5 mm was created and extended to encompass the proximal seminal vesicles. The second fraction was given 1 to 2 weeks after initial treatment, based on patient, physician, and operating room availability.

Health-Related Quality of Life Assessment

Veteran-reported genitourinary (GU), gastrointestinal (GI), and sexual health-related quality of life (hrQOL) were assessed using the validated International Prostate Symptom Score (IPSS) and the Expanded Prostate Cancer Index Composite Short Form (EPIC-26) instruments.^10,11 Baseline veteran-reported hrQOL scores in the GU, GI, and sexual domains were obtained prior to each veteran’s first HDR treatment. Veteran-reported hrQOL scores were assessed at each of the patient’s follow-up appointments. Physician-graded toxicity was assessed Common Terminology Criteria for Adverse Events (CTCAE) v 4.03 criteria.¹² Physician-graded toxicity was assessed at each follow-up visit and reported as the highest grade reported during any follow-up examination.

Follow-up appointments typically occurred at 1 month, 3 months, 6 months, 12 months, and subsequently every 6 months after the second HDR treatment. Follow-up appointments were conducted in the radiation oncology department at EHJVAH.

Minimal Clinically Important Differences

To evaluate the veteran-reported hrQOL, we characterized statistically significant differences in IPSS or EPIC-26 scores over time as compared with baseline values as clinically important or not clinically important through the use of reported minimal clinically important difference (MCID) assessments.^13-15 For the IPSS, we used reported data that showed a change of ≥ 3.0 points represented a clinically meaningful change in urinary function.¹⁴ For the EPIC-26 scores, we used reported data that showed a change of ≥ 6 points for urinary incontinence score, ≥ 5 points for urinary obstruction score, ≥ 4 points for bowel score, and ≥ 10 points for sexual score to represent an MCID.¹⁵

Statistical Analysis

Changes in veteran-reported hrQOL over time were compared using mixed linear effects models, with the time since the last BT implant serving as the fixed variable. Effects were deemed statistically significant if P < .05. If a statistically significant difference from baseline was found at any time point, additional evaluation was done to see if the numerical difference in the assessment led to an MCID as described above. IBM SPSS Statistics for Windows, version 25.0 was used for data analysis.

Results

Seventy-four veterans were included in the study. The median follow-up was 18 months (range 1-43). The demographic and oncologic specifics of the treated veterans are outlined in Table 1.

There was a significant increase in IPSS (P < .001) with reciprocal decline in EPIC-26 urinary incontinence (P = .008) and EPIC-26 urinary obstruction scores (P = .001) from baseline over time (Table 2 and Figure 1). At the 18-month follow-up assessment, there was no longer a significant difference in the EPIC-26 urinary obstruction score from baseline (88.7 vs 84.0, P = .31). The increases in IPSS at the 1-, 3-, and 6-month assessments met the criteria for MCID. The decrease in EPIC-26 urinary incontinence scores at the 1-, 3-, 6-, 12-, and 18-month assessments were found to be an MCID, as were the decrease in EPIC-26 urinary obstruction scores at the 1-, 3-, 6-, and 12-month assessments.

Discussion

We performed a retrospective study of veterans with low- or intermediate-risk PC undergoing definitive HDR prostate BT as monotherapy. We found that veterans experienced immediate declines in GU, GI, and sexual hrQOL after treatment. However, each trended toward a return to baseline over time, with the EPIC-26 urinary obstruction and the EPIC-26 bowel scores showing no difference from the baseline value within 18 months and 12 months, respectively. The physician-reported toxicities were low, with only 1 incidence of grade 3 GU toxicity, no grade 3 GI or sexual toxicities, and no grade 4 or 5 toxicity. This suggests that HDRBT is a well-tolerated and safe, definitive treatment for veterans with localized PC.

In a series similar to ours, Gaudet and colleagues reported on their single institutional results of treating 30 low- or intermediate-risk PC patients with HDRBT as monotherapy.¹⁶ Patients included in their study were civilians from the general population, treated in a similar fashion to the veterans treated in our study. Each patient received 27 Gy in 2 fractions given over 2 implants. The authors collected patient-reported hrQOL results using the IPSS and EPIC questionnaires and found that 57% of patients treated experienced moderate-to-severe urinary symptoms at the 1-month assessment after implantation, with a rapid recovery toward baseline over time. In contrast, GI symptoms did not change from baseline, while sexual symptoms worsened after implantation and failed to return to baseline.

Limitations

To the authors’ knowledge, there are no other studies exploring HDR prostate BT toxicity in a veteran-specific population, and our study is novel in addressing this question. One limitation of the study is the relatively short median follow-up time of 18 months. With this limitation, our data were not yet sufficiently mature to perform biochemical control or overall survival analyses. The next step in our study is to calculate these clinical endpoints from our data after longer follow-up.

An additional limitation to our study is the single institutional nature of the design. While veterans from neighboring VA hospitals were included in the study by way of referral and treatment at our center, the only VA hospital in the state to provide radiation therapy, our patient population remains limited. Further multi-institutional and prospective data are needed to validate our findings.

Conclusions

HDR prostate BT as monotherapy is feasible with a favorable veteran-reported hrQOL and physician-graded toxicity profile. Veterans should be educated about this treatment modality when considering the optimal treatment for their localized prostate cancer.

Nearly 50,000 veterans are diagnosed with cancer within the Veterans Health Administration annually with prostate cancer (PC) being the most frequently diagnosed, accounting for 29% of all cancers diagnosed.¹ The treatment of PC depends on the stage and risk group at presentation and patient preference. Men with early stage, localized PC can be managed with prostatectomy, radiation therapy, or active surveillance.²

Within the Veterans Health Administration, more patients are treated with radiation therapy than with radical prostatectomy.³ This is in contrast to the civil health system, where more patients are treated with radical prostatectomy than with radiation therapy.^4,5 Radiation therapy for PC can be given externally with external beam radiation therapy or internally with brachytherapy (BT). BT is categorized by the rate at which the radiation dose is delivered and generally grouped as low-dose rate (LDR) or high-dose rate (HDR). LDRBT consists of permanently implanting radioactive seeds, which slowly deliver a radiation dose over an extended period. HDRBT consists of implanting catheters that allow delivery of a radioactive source to be placed temporarily in the prostate and removed after treatment. The utilization of HDRBT has become more common as treatment has evolved to consist of fewer, larger fractions in a shorter time, making it a convenient treatment option for men with PC.⁶ The veteran population has singular medical challenges. These patients differ from the general population and are often underrepresented in medical research and published studies.⁷ There are no studies exploring the treatment-associated toxicities from HDRBT treatment for PC specifically in the veteran population. The objective of this study is to report our findings regarding the veteran-reported and physician-graded toxicities associated with HDRBT as monotherapy in veterans treated through the US Department of Veterans Affairs (VA) for PC.

Methods

We performed a retrospective cohort study of a prospectively maintained, institutional review board-approved database of patients treated with HDRBT for PC. Veterans were seen in consultation at Edward Hines, Jr. VA Hospital (EHJVAH) in Hines, Illinois. This is the only VA hospital in Illinois that offers radiation therapy, so it acted as a tertiary center, receiving referrals from other, neighboring VA hospitals. If the veteran was deemed a good BT candidate and elected to proceed with HDRBT, HDR treatment was performed at a partnering academic institution equipped to provide HDRBT (Loyola University Medical Center).

We selected patients with National Cancer Center Network (NCCN) low- or intermediate-risk PC undergoing definitive HDRBT as monotherapy using 13.5 Gy x 2 fractions delivered over 2 implants that were 1 to 2 weeks apart. Patients who received androgen deprivation therapy (ADT) were excluded from this study. No patients received supplemental external beam radiation. Men with unfavorable intermediate risk PC were offered ADT and BT in accordance with NCCN guidelines. However, patients with unfavorable intermediate-risk PC who declined ADT or who were deemed poor ADT candidates due to comorbidities were treated with HDR as monotherapy and included in this study.⁸

HDR Treatment

Our HDRBT implant procedure and treatment planning details have been previously described.⁹ In brief, patients were implanted with between 17 and 22 catheters based on gland size under transrectal ultrasound guidance. After implantation, computed tomography and, when possible, magnetic resonance imaging of the prostate were obtained and registered for target delineation. The prostate was segmented, and an asymmetric planning target volume of 0 to 5 mm was created and extended to encompass the proximal seminal vesicles. The second fraction was given 1 to 2 weeks after initial treatment, based on patient, physician, and operating room availability.

Health-Related Quality of Life Assessment

Veteran-reported genitourinary (GU), gastrointestinal (GI), and sexual health-related quality of life (hrQOL) were assessed using the validated International Prostate Symptom Score (IPSS) and the Expanded Prostate Cancer Index Composite Short Form (EPIC-26) instruments.^10,11 Baseline veteran-reported hrQOL scores in the GU, GI, and sexual domains were obtained prior to each veteran’s first HDR treatment. Veteran-reported hrQOL scores were assessed at each of the patient’s follow-up appointments. Physician-graded toxicity was assessed Common Terminology Criteria for Adverse Events (CTCAE) v 4.03 criteria.¹² Physician-graded toxicity was assessed at each follow-up visit and reported as the highest grade reported during any follow-up examination.

Follow-up appointments typically occurred at 1 month, 3 months, 6 months, 12 months, and subsequently every 6 months after the second HDR treatment. Follow-up appointments were conducted in the radiation oncology department at EHJVAH.

Minimal Clinically Important Differences

To evaluate the veteran-reported hrQOL, we characterized statistically significant differences in IPSS or EPIC-26 scores over time as compared with baseline values as clinically important or not clinically important through the use of reported minimal clinically important difference (MCID) assessments.^13-15 For the IPSS, we used reported data that showed a change of ≥ 3.0 points represented a clinically meaningful change in urinary function.¹⁴ For the EPIC-26 scores, we used reported data that showed a change of ≥ 6 points for urinary incontinence score, ≥ 5 points for urinary obstruction score, ≥ 4 points for bowel score, and ≥ 10 points for sexual score to represent an MCID.¹⁵

Statistical Analysis

Changes in veteran-reported hrQOL over time were compared using mixed linear effects models, with the time since the last BT implant serving as the fixed variable. Effects were deemed statistically significant if P < .05. If a statistically significant difference from baseline was found at any time point, additional evaluation was done to see if the numerical difference in the assessment led to an MCID as described above. IBM SPSS Statistics for Windows, version 25.0 was used for data analysis.

Results

Seventy-four veterans were included in the study. The median follow-up was 18 months (range 1-43). The demographic and oncologic specifics of the treated veterans are outlined in Table 1.

There was a significant increase in IPSS (P < .001) with reciprocal decline in EPIC-26 urinary incontinence (P = .008) and EPIC-26 urinary obstruction scores (P = .001) from baseline over time (Table 2 and Figure 1). At the 18-month follow-up assessment, there was no longer a significant difference in the EPIC-26 urinary obstruction score from baseline (88.7 vs 84.0, P = .31). The increases in IPSS at the 1-, 3-, and 6-month assessments met the criteria for MCID. The decrease in EPIC-26 urinary incontinence scores at the 1-, 3-, 6-, 12-, and 18-month assessments were found to be an MCID, as were the decrease in EPIC-26 urinary obstruction scores at the 1-, 3-, 6-, and 12-month assessments.

Discussion

We performed a retrospective study of veterans with low- or intermediate-risk PC undergoing definitive HDR prostate BT as monotherapy. We found that veterans experienced immediate declines in GU, GI, and sexual hrQOL after treatment. However, each trended toward a return to baseline over time, with the EPIC-26 urinary obstruction and the EPIC-26 bowel scores showing no difference from the baseline value within 18 months and 12 months, respectively. The physician-reported toxicities were low, with only 1 incidence of grade 3 GU toxicity, no grade 3 GI or sexual toxicities, and no grade 4 or 5 toxicity. This suggests that HDRBT is a well-tolerated and safe, definitive treatment for veterans with localized PC.

In a series similar to ours, Gaudet and colleagues reported on their single institutional results of treating 30 low- or intermediate-risk PC patients with HDRBT as monotherapy.¹⁶ Patients included in their study were civilians from the general population, treated in a similar fashion to the veterans treated in our study. Each patient received 27 Gy in 2 fractions given over 2 implants. The authors collected patient-reported hrQOL results using the IPSS and EPIC questionnaires and found that 57% of patients treated experienced moderate-to-severe urinary symptoms at the 1-month assessment after implantation, with a rapid recovery toward baseline over time. In contrast, GI symptoms did not change from baseline, while sexual symptoms worsened after implantation and failed to return to baseline.

Limitations

To the authors’ knowledge, there are no other studies exploring HDR prostate BT toxicity in a veteran-specific population, and our study is novel in addressing this question. One limitation of the study is the relatively short median follow-up time of 18 months. With this limitation, our data were not yet sufficiently mature to perform biochemical control or overall survival analyses. The next step in our study is to calculate these clinical endpoints from our data after longer follow-up.

An additional limitation to our study is the single institutional nature of the design. While veterans from neighboring VA hospitals were included in the study by way of referral and treatment at our center, the only VA hospital in the state to provide radiation therapy, our patient population remains limited. Further multi-institutional and prospective data are needed to validate our findings.

Conclusions

HDR prostate BT as monotherapy is feasible with a favorable veteran-reported hrQOL and physician-graded toxicity profile. Veterans should be educated about this treatment modality when considering the optimal treatment for their localized prostate cancer.

Nearly 50,000 veterans are diagnosed with cancer within the Veterans Health Administration annually with prostate cancer (PC) being the most frequently diagnosed, accounting for 29% of all cancers diagnosed.¹ The treatment of PC depends on the stage and risk group at presentation and patient preference. Men with early stage, localized PC can be managed with prostatectomy, radiation therapy, or active surveillance.²

Within the Veterans Health Administration, more patients are treated with radiation therapy than with radical prostatectomy.³ This is in contrast to the civil health system, where more patients are treated with radical prostatectomy than with radiation therapy.^4,5 Radiation therapy for PC can be given externally with external beam radiation therapy or internally with brachytherapy (BT). BT is categorized by the rate at which the radiation dose is delivered and generally grouped as low-dose rate (LDR) or high-dose rate (HDR). LDRBT consists of permanently implanting radioactive seeds, which slowly deliver a radiation dose over an extended period. HDRBT consists of implanting catheters that allow delivery of a radioactive source to be placed temporarily in the prostate and removed after treatment. The utilization of HDRBT has become more common as treatment has evolved to consist of fewer, larger fractions in a shorter time, making it a convenient treatment option for men with PC.⁶ The veteran population has singular medical challenges. These patients differ from the general population and are often underrepresented in medical research and published studies.⁷ There are no studies exploring the treatment-associated toxicities from HDRBT treatment for PC specifically in the veteran population. The objective of this study is to report our findings regarding the veteran-reported and physician-graded toxicities associated with HDRBT as monotherapy in veterans treated through the US Department of Veterans Affairs (VA) for PC.

Methods

We performed a retrospective cohort study of a prospectively maintained, institutional review board-approved database of patients treated with HDRBT for PC. Veterans were seen in consultation at Edward Hines, Jr. VA Hospital (EHJVAH) in Hines, Illinois. This is the only VA hospital in Illinois that offers radiation therapy, so it acted as a tertiary center, receiving referrals from other, neighboring VA hospitals. If the veteran was deemed a good BT candidate and elected to proceed with HDRBT, HDR treatment was performed at a partnering academic institution equipped to provide HDRBT (Loyola University Medical Center).

We selected patients with National Cancer Center Network (NCCN) low- or intermediate-risk PC undergoing definitive HDRBT as monotherapy using 13.5 Gy x 2 fractions delivered over 2 implants that were 1 to 2 weeks apart. Patients who received androgen deprivation therapy (ADT) were excluded from this study. No patients received supplemental external beam radiation. Men with unfavorable intermediate risk PC were offered ADT and BT in accordance with NCCN guidelines. However, patients with unfavorable intermediate-risk PC who declined ADT or who were deemed poor ADT candidates due to comorbidities were treated with HDR as monotherapy and included in this study.⁸

HDR Treatment

Our HDRBT implant procedure and treatment planning details have been previously described.⁹ In brief, patients were implanted with between 17 and 22 catheters based on gland size under transrectal ultrasound guidance. After implantation, computed tomography and, when possible, magnetic resonance imaging of the prostate were obtained and registered for target delineation. The prostate was segmented, and an asymmetric planning target volume of 0 to 5 mm was created and extended to encompass the proximal seminal vesicles. The second fraction was given 1 to 2 weeks after initial treatment, based on patient, physician, and operating room availability.

Health-Related Quality of Life Assessment

Veteran-reported genitourinary (GU), gastrointestinal (GI), and sexual health-related quality of life (hrQOL) were assessed using the validated International Prostate Symptom Score (IPSS) and the Expanded Prostate Cancer Index Composite Short Form (EPIC-26) instruments.^10,11 Baseline veteran-reported hrQOL scores in the GU, GI, and sexual domains were obtained prior to each veteran’s first HDR treatment. Veteran-reported hrQOL scores were assessed at each of the patient’s follow-up appointments. Physician-graded toxicity was assessed Common Terminology Criteria for Adverse Events (CTCAE) v 4.03 criteria.¹² Physician-graded toxicity was assessed at each follow-up visit and reported as the highest grade reported during any follow-up examination.

Follow-up appointments typically occurred at 1 month, 3 months, 6 months, 12 months, and subsequently every 6 months after the second HDR treatment. Follow-up appointments were conducted in the radiation oncology department at EHJVAH.

Minimal Clinically Important Differences

To evaluate the veteran-reported hrQOL, we characterized statistically significant differences in IPSS or EPIC-26 scores over time as compared with baseline values as clinically important or not clinically important through the use of reported minimal clinically important difference (MCID) assessments.^13-15 For the IPSS, we used reported data that showed a change of ≥ 3.0 points represented a clinically meaningful change in urinary function.¹⁴ For the EPIC-26 scores, we used reported data that showed a change of ≥ 6 points for urinary incontinence score, ≥ 5 points for urinary obstruction score, ≥ 4 points for bowel score, and ≥ 10 points for sexual score to represent an MCID.¹⁵

Statistical Analysis

Changes in veteran-reported hrQOL over time were compared using mixed linear effects models, with the time since the last BT implant serving as the fixed variable. Effects were deemed statistically significant if P < .05. If a statistically significant difference from baseline was found at any time point, additional evaluation was done to see if the numerical difference in the assessment led to an MCID as described above. IBM SPSS Statistics for Windows, version 25.0 was used for data analysis.

Results

Seventy-four veterans were included in the study. The median follow-up was 18 months (range 1-43). The demographic and oncologic specifics of the treated veterans are outlined in Table 1.

There was a significant increase in IPSS (P < .001) with reciprocal decline in EPIC-26 urinary incontinence (P = .008) and EPIC-26 urinary obstruction scores (P = .001) from baseline over time (Table 2 and Figure 1). At the 18-month follow-up assessment, there was no longer a significant difference in the EPIC-26 urinary obstruction score from baseline (88.7 vs 84.0, P = .31). The increases in IPSS at the 1-, 3-, and 6-month assessments met the criteria for MCID. The decrease in EPIC-26 urinary incontinence scores at the 1-, 3-, 6-, 12-, and 18-month assessments were found to be an MCID, as were the decrease in EPIC-26 urinary obstruction scores at the 1-, 3-, 6-, and 12-month assessments.

Discussion

We performed a retrospective study of veterans with low- or intermediate-risk PC undergoing definitive HDR prostate BT as monotherapy. We found that veterans experienced immediate declines in GU, GI, and sexual hrQOL after treatment. However, each trended toward a return to baseline over time, with the EPIC-26 urinary obstruction and the EPIC-26 bowel scores showing no difference from the baseline value within 18 months and 12 months, respectively. The physician-reported toxicities were low, with only 1 incidence of grade 3 GU toxicity, no grade 3 GI or sexual toxicities, and no grade 4 or 5 toxicity. This suggests that HDRBT is a well-tolerated and safe, definitive treatment for veterans with localized PC.

In a series similar to ours, Gaudet and colleagues reported on their single institutional results of treating 30 low- or intermediate-risk PC patients with HDRBT as monotherapy.¹⁶ Patients included in their study were civilians from the general population, treated in a similar fashion to the veterans treated in our study. Each patient received 27 Gy in 2 fractions given over 2 implants. The authors collected patient-reported hrQOL results using the IPSS and EPIC questionnaires and found that 57% of patients treated experienced moderate-to-severe urinary symptoms at the 1-month assessment after implantation, with a rapid recovery toward baseline over time. In contrast, GI symptoms did not change from baseline, while sexual symptoms worsened after implantation and failed to return to baseline.

Limitations

To the authors’ knowledge, there are no other studies exploring HDR prostate BT toxicity in a veteran-specific population, and our study is novel in addressing this question. One limitation of the study is the relatively short median follow-up time of 18 months. With this limitation, our data were not yet sufficiently mature to perform biochemical control or overall survival analyses. The next step in our study is to calculate these clinical endpoints from our data after longer follow-up.

An additional limitation to our study is the single institutional nature of the design. While veterans from neighboring VA hospitals were included in the study by way of referral and treatment at our center, the only VA hospital in the state to provide radiation therapy, our patient population remains limited. Further multi-institutional and prospective data are needed to validate our findings.

Conclusions

HDR prostate BT as monotherapy is feasible with a favorable veteran-reported hrQOL and physician-graded toxicity profile. Veterans should be educated about this treatment modality when considering the optimal treatment for their localized prostate cancer.