Prophylaxis with beta blockers as a performance measure of quality health care in cirrhosis

Abstract

Wilbur K, Sidhu K (Faculty of Pharmaceutical Sciences, University of British Columbia, Vancouver General Hospital, Vancouver, British Columbia, Canada). Beta blocker prophylaxis for patients with variceal hemorrhage. J Clin Gastroenterol 2005;39:435–440.Cirrhosis and complications from portal hypertension now represents the 12th leading cause of death in adults residing within the United States for the year 2000 (Natl Vital Stat Rep 2002;50:1–119). An estimated 26,500 deaths were reported, of which over 50% were related specifically to alcohol-related liver injury. When accounting for all complications of cirrhosis including the development of hepatocellular carcinoma, an estimated 40,000 deaths can be identified from cirrhosis, elevating it to the eighth leading cause of death in 2000 (Hepatology 2002;36:S30–S34). In turn, cirrhosis is the fifth leading cause of mortality in persons between ages 45 and 54 years. It remains the second leading cause of gastrointestinal death after colorectal neoplasia (Gastroenterology 2004;126:1448–1453).The burden of disease from cirrhosis also includes significant health care resource utilization. In 2002, the estimated number of discharges for patients hospitalized with cirrhosis and related complications was 153,783 with 26% of discharges occurring for individuals age 65 and over. The mean charge per episode of hospitalization was greater than $25,000 resulting in an overall total charge of $3.9 billion for this calendar year. The overall in-hospital mortality rate continues to be high at 9% and comparable to death rates among patients hospitalized for congestive heart failure. For patients who are 85 years and older, the in-hospital mortality rate exceeds 14% (see http://hcup.ahrq.gov/HCUPnet.asp).In addition to hepatocellular carcinoma, the major causes of liver-related death in patients with cirrhosis include infection/sepsis, variceal bleeding, progressive liver failure, and hepatorenal syndrome (Liver 1987;7:316–324). Among these life-threatening complications, only variceal bleeding is amenable to prophylactic interventions, which might prevent this devastating outcome. Despite a reduction in frequency of variceal bleeding over time (Gut 2001;49:682–685), case fatality rates continue to be as high as 30% for patients with advanced liver disease (Hepatology 2004;40:652–659). Based on several high-quality randomized controlled trials, the use of nonselective beta blockers can reduce the risk of initial variceal bleeding by 30%–50% over a 2-year period. Meta-analysis of existing studies also demonstrates a trend for improved survival with beta blocker therapy (Semin Liver Dis 1999;19:475–505). Clinical features among patients who benefit from beta blockers include the presence of large varices, stigmata associated with an increased risk for bleeding including red whale markings, and evidence for advanced liver disease proxied by Child-Turcotte-Pugh classification (Hepatology 1996;24:1047–1052, Am J Gastroenterol 2000;95:2915–2920). For patients with a prior history of hemorrhage from esophageal varices, the need for secondary prophylaxis is of even greater importance because more than 70% of untreated survivors will have recurrent bleeding within 1 year (Lancet 2003;361:952–954).However, the degree to which primary and secondary prophylaxis strategies are used for various populations affected by cirrhosis and portal hypertension remains largely unknown. Recent data, however, raise serious concerns that a significant degree of unwarranted variation in clinical practice exists for providing this potentially beneficial treatment (Am J Gastroenterol 2003;98:653–659, Hepatology 2003;38:599–612, Am J Gastroenterol 2003;98:2424–2434, J Hepatol 2003;39:509–514, Am J Gastroenterol 2004;99:645–649).In the present article, Wilbur and Sidhu retrospectively evaluated the frequency of beta blocker use for hospitalized patients with suspected variceal hemorrhage over a 3-year period. The study cohort also included patients with a prior history of variceal hemorrhage who were admitted to the hospital yet found to have a nonvariceal source of bleeding. Upon review of medical records and pharmacy databases, patients who were receiving beta blocker therapy before admission and at hospital discharge were identified. Absolute contraindications for initiating beta blocker therapy were considered to include: (1) asthma, (2) second- or third-degree heart block, (3) sick sinus syndrome, or (4) documented allergy or prior intolerance.Among 205 patients whose medical records were screened, a total of 106 individuals were eligible to be in the study cohort. Ninety-two of the 205 patients were misclassified as having liver disease or varices and subsequently excluded. Demographic characteristics include a mean age of 56 years and 50% men. Major hepatic disease etiologies included alcohol (47%) and chronic viral hepatitis (44%). Ascites was present in 42% of patients; however, the frequency of hepatic encephalopathy and spontaneous bacterial peritonitis were less than 10% each, respectively. Nearly 75% of patients met criteria for Child-Pugh class C disease severity while the average model for end-stage liver disease (MELD) score for this cohort was 17. Management strategies for controlling acute variceal hemorrhage included intravenous octreotide with endoscopic band ligation (41%) or intravenous octreotide alone (27%). The average length of stay for the 106 patients was 14 days with an in-hospital mortality rate of 20%.An estimated 58% of patients had known esophageal varices prior to hospitalization. Forty percent of patients had a history of prior variceal hemorrhage with 55% experiencing two or more bleeds. However, the majority of patients with known varices and no prior bleeding were not receiving primary prophylaxis with beta blockers (94%). For individuals with a prior or current history of variceal hemorrhage, only 20% of patients were taking beta blockers on admission. The rate of beta blocker use increased to 48% at discharge. Age and severity of liver disease were not associated with the use of beta blocker therapy as secondary prophylaxis on admission or discharge. Individuals with lower average MELD scores (12 vs 16, P = .04) were more likely to receive beta blocker therapy at discharge. Patients with a history of 2 or more variceal bleeding episodes were also more likely to be on beta blockers at hospital admission compared to those with a less frequent history of hemorrhage (73% vs 41%, P = .04). Hypotension (defined as a systolic blood pressure <110 mmHg) was the most common reason cited in medical records for not providing beta blocker treatment as secondary prophylaxis following acute variceal bleeding.CommentDespite the existence of multiple, high-quality, randomized controlled trials that support the use of beta blocker therapy for primary and secondary prophylaxis against esophageal variceal bleeding, a significant gap exists with the translation of these scientific results into clinical practice. In this population, nearly 95% of individuals with known esophageal varices did not receive beta blocker therapy for primary prophylaxis. Of greater concern was that only 20% of individuals with a prior history of variceal bleeding were using beta blocker therapy prior to hospital admission. An improved rate of administration at 48% was observed upon discharge yet remains lower than acceptable.The Institute of Medicine has defined several criteria which recognize chronic diseases of importance that require quality of care measurement to reduce practice variation and improve clinical outcomes (Acad Med 2002;77:91–99). These include safety, efficacy, patient-centeredness, timeliness, efficiency, and equity. Based on these recommendations, cirrhosis and its complications may be considered one of these conditions. However, there has been scant information to date about how to define and measure quality of care for patients with cirrhosis. This is despite the fact that several data sources exist including clinical practice guidelines, which incorporate recommendations based on evidence-based data (Am J Gastroenterol 1997;92:1081–1091, Hepatology 1998;28:868–880, Gut 2000;46:S3–S4:III1–III15). In turn, a number of potential quality indicators (or performance measures) identified from these guidelines could be studied to determine their reliability and validity for assessing quality of care.In the field of quality improvement, a performance measure is used as an operational term to quantify the extent of quality health care that is provided. The most sound performance measures are those supported by high levels of evidence-based medicine, mainly through large or numerous randomized controlled trials. The failure to adhere to these performance measures is usually associated with a significant reduction in the ability to achieve favorable clinical outcomes (Circulation 2005;111:1703–1712). One possible example in patients with cirrhosis is the use of antibiotic prophylaxis for gastrointestinal bleeding where the number of serious infectious events and mortality rates are significantly reduced compared to no intervention (Hepatology 1999;29:1655–1661). A number of performance measures, however, are considered important components of clinical care yet often lack the weight of evidence from controlled trials. These measures are less easy to use and require methodical approaches to definition (Med Care 1999;37:964–968).The study by Wilbur and Sidhu further highlights the underuse of beta blocker therapy for primary and secondary prophylaxis in patients with cirrhosis and esophageal varices. These data are comparable to 4 reported studies of clinical practice (Am J Gastroenterol 2003;98:653–659, Hepatology 2003;38:599–612, Am J Gastroenterol 2003;98:2424–2434, J Hepatol 2003;39:509–514, Am J Gastroenterol 2004;99:645–649) in which similar rates of beta blocker use for primary prophylaxis (20%–30%) and secondary prophylaxis (45%–80%) are reported. Although a significant majority of patients in this study did not receive beta blocker therapy as primary prophylaxis, a number of factors have to be considered when interpreting the results.The authors describe a lack of knowledge regarding the size of esophageal varices in this patient population. Prior controlled clinical trials showed a small risk reduction (5%) with beta blocker therapy compared to placebo among patients with small esophageal varices (Semin Liver Dis 1999;19:475–505, Hepatology 1997;25:63–70). As a result, the use of beta blockers as primary prophylaxis for these individuals has not been routinely performed in clinical practice. A recent study, however, described clinical benefit for preventing growth of small to large esophageal varices on beta blockers which, in turn, reduced the risk for initial variceal bleeding (Gastroenterology 2004;127:476–484). These data require additional confirmation in various populations before a change in medical practice can be recommended.Data on the frequency of initial and subsequent endoscopy for detecting large esophageal varices for patients in this study was also not available. Thus, a low screening rate may have contributed to the increased number of patients who were not offered beta blocker therapy. More importantly, the presence of low utilization rates with beta blocker therapy as primary prophylaxis after identifying large esophageal varices by screening endoscopy would also be considered a critical gap in quality of care for this population. Unfortunately, the extent to which high-risk patients are actually receiving beta blocker therapy as primary prophylaxis is unknown.A significant proportion of patients with severe advanced liver disease (proxied by Child-Pugh class C) will not tolerate long-term beta blocker therapy due to an enhanced risk for side effects (Semin Liver Dis 1999;19:475–505, Hepatology 1997;25:63–70). Endoscopic variceal bland ligation for primary prophylaxis has now emerged as a viable alternative for patients who are intolerant or have contraindications to beta blocker treatment (Hepatology 2001;33:802–807). In this study, the frequency of band ligation in this population was not available for analysis. Nevertheless, a significant percentage of patients did not have absolute contraindications for the use of beta blocker therapy as prophylaxis against variceal bleeding. In this context, it would be difficult to attribute a low rate of beta blocker use based on patient ineligibility.A striking number of patients received inappropriate forms of pharmacologic therapy for prophylaxis against variceal bleeding. The use of oral nitrate therapy (with or without beta blockers) as primary and secondary prevention was observed, albeit in <5% of patients. However, this should be considered an inappropriate strategy based on a lack of evidence to support its use (Am J Med 2004;116:759–766, Semin Liver Dis 1999;19:475–505). In contrast, an estimated 25%–33% of patients were receiving selective or unproven beta blocker therapies as primary and secondary prophylaxis against variceal bleeding. This also constitutes inappropriate or misuse of evidence-based therapy and should be considered an element of poor quality care in this population.Information regarding duration of treatment, extent of adherence to treatment, and treatment efficacy as measured by hepatic vein pressure gradient measurements were also not available in this study. These factors have previously been described as important variables that contribute to the effectiveness of pharmacologic therapy as prophylaxis (Lancet 2003;361:952–954, Hepatology 2001;34:1096–1102). In turn, this information is also critical for determining extent of quality of medical care for improving clinical outcomes.Despite limitations in study design and missing data elements, this study further highlights the gap in using beta blockers as primary and secondary prophylaxis against esophageal variceal hemorrhage in patients with cirrhosis. These data are consistent with reports that at least 50% of individuals with chronic disease are undertreated with existing medical therapies deemed to be effective (N Engl J Med 2003;348:2635–2645). What can be done to improve these rates? Several interventions have been tried including the passive inoculation of clinical practice guidelines as well as formal instruction through continuing medical education formats. Although not specifically investigated for patients with cirrhosis, these methods have been found ineffective in changing provider attitude and reducing practice variation (JAMA 2004;291:2466–2470). With decision support tools and the electronic medical record, a strategy using timely reminders can be implemented to ensure that diagnostic endoscopy for identifying large esophageal varices is performed in all patients with cirrhosis. Similarly, the use of surveillance endoscopy performed at appropriate intervals for individuals with no or small varices can also be facilitated. Active interventions supported by opinion leaders and patient care teams do appear to improve quality of care in other fields (Health Affairs 2003;22:196–201) but require study to determine their true effectiveness in hepatology. Wilbur K, Sidhu K (Faculty of Pharmaceutical Sciences, University of British Columbia, Vancouver General Hospital, Vancouver, British Columbia, Canada). Beta blocker prophylaxis for patients with variceal hemorrhage. J Clin Gastroenterol 2005;39:435–440. Cirrhosis and complications from portal hypertension now represents the 12th leading cause of death in adults residing within the United States for the year 2000 (Natl Vital Stat Rep 2002;50:1–119). An estimated 26,500 deaths were reported, of which over 50% were related specifically to alcohol-related liver injury. When accounting for all complications of cirrhosis including the development of hepatocellular carcinoma, an estimated 40,000 deaths can be identified from cirrhosis, elevating it to the eighth leading cause of death in 2000 (Hepatology 2002;36:S30–S34). In turn, cirrhosis is the fifth leading cause of mortality in persons between ages 45 and 54 years. It remains the second leading cause of gastrointestinal death after colorectal neoplasia (Gastroenterology 2004;126:1448–1453). The burden of disease from cirrhosis also includes significant health care resource utilization. In 2002, the estimated number of discharges for patients hospitalized with cirrhosis and related complications was 153,783 with 26% of discharges occurring for individuals age 65 and over. The mean charge per episode of hospitalization was greater than $25,000 resulting in an overall total charge of $3.9 billion for this calendar year. The overall in-hospital mortality rate continues to be high at 9% and comparable to death rates among patients hospitalized for congestive heart failure. For patients who are 85 years and older, the in-hospital mortality rate exceeds 14% (see http://hcup.ahrq.gov/HCUPnet.asp). In addition to hepatocellular carcinoma, the major causes of liver-related death in patients with cirrhosis include infection/sepsis, variceal bleeding, progressive liver failure, and hepatorenal syndrome (Liver 1987;7:316–324). Among these life-threatening complications, only variceal bleeding is amenable to prophylactic interventions, which might prevent this devastating outcome. Despite a reduction in frequency of variceal bleeding over time (Gut 2001;49:682–685), case fatality rates continue to be as high as 30% for patients with advanced liver disease (Hepatology 2004;40:652–659). Based on several high-quality randomized controlled trials, the use of nonselective beta blockers can reduce the risk of initial variceal bleeding by 30%–50% over a 2-year period. Meta-analysis of existing studies also demonstrates a trend for improved survival with beta blocker therapy (Semin Liver Dis 1999;19:475–505). Clinical features among patients who benefit from beta blockers include the presence of large varices, stigmata associated with an increased risk for bleeding including red whale markings, and evidence for advanced liver disease proxied by Child-Turcotte-Pugh classification (Hepatology 1996;24:1047–1052, Am J Gastroenterol 2000;95:2915–2920). For patients with a prior history of hemorrhage from esophageal varices, the need for secondary prophylaxis is of even greater importance because more than 70% of untreated survivors will have recurrent bleeding within 1 year (Lancet 2003;361:952–954). However, the degree to which primary and secondary prophylaxis strategies are used for various populations affected by cirrhosis and portal hypertension remains largely unknown. Recent data, however, raise serious concerns that a significant degree of unwarranted variation in clinical practice exists for providing this potentially beneficial treatment (Am J Gastroenterol 2003;98:653–659, Hepatology 2003;38:599–612, Am J Gastroenterol 2003;98:2424–2434, J Hepatol 2003;39:509–514, Am J Gastroenterol 2004;99:645–649). In the present article, Wilbur and Sidhu retrospectively evaluated the frequency of beta blocker use for hospitalized patients with suspected variceal hemorrhage over a 3-year period. The study cohort also included patients with a prior history of variceal hemorrhage who were admitted to the hospital yet found to have a nonvariceal source of bleeding. Upon review of medical records and pharmacy databases, patients who were receiving beta blocker therapy before admission and at hospital discharge were identified. Absolute contraindications for initiating beta blocker therapy were considered to include: (1) asthma, (2) second- or third-degree heart block, (3) sick sinus syndrome, or (4) documented allergy or prior intolerance. Among 205 patients whose medical records were screened, a total of 106 individuals were eligible to be in the study cohort. Ninety-two of the 205 patients were misclassified as having liver disease or varices and subsequently excluded. Demographic characteristics include a mean age of 56 years and 50% men. Major hepatic disease etiologies included alcohol (47%) and chronic viral hepatitis (44%). Ascites was present in 42% of patients; however, the frequency of hepatic encephalopathy and spontaneous bacterial peritonitis were less than 10% each, respectively. Nearly 75% of patients met criteria for Child-Pugh class C disease severity while the average model for end-stage liver disease (MELD) score for this cohort was 17. Management strategies for controlling acute variceal hemorrhage included intravenous octreotide with endoscopic band ligation (41%) or intravenous octreotide alone (27%). The average length of stay for the 106 patients was 14 days with an in-hospital mortality rate of 20%. An estimated 58% of patients had known esophageal varices prior to hospitalization. Forty percent of patients had a history of prior variceal hemorrhage with 55% experiencing two or more bleeds. However, the majority of patients with known varices and no prior bleeding were not receiving primary prophylaxis with beta blockers (94%). For individuals with a prior or current history of variceal hemorrhage, only 20% of patients were taking beta blockers on admission. The rate of beta blocker use increased to 48% at discharge. Age and severity of liver disease were not associated with the use of beta blocker therapy as secondary prophylaxis on admission or discharge. Individuals with lower average MELD scores (12 vs 16, P = .04) were more likely to receive beta blocker therapy at discharge. Patients with a history of 2 or more variceal bleeding episodes were also more likely to be on beta blockers at hospital admission compared to those with a less frequent history of hemorrhage (73% vs 41%, P = .04). Hypotension (defined as a systolic blood pressure <110 mmHg) was the most common reason cited in medical records for not providing beta blocker treatment as secondary prophylaxis following acute variceal bleeding. CommentDespite the existence of multiple, high-quality, randomized controlled trials that support the use of beta blocker therapy for primary and secondary prophylaxis against esophageal variceal bleeding, a significant gap exists with the translation of these scientific results into clinical practice. In this population, nearly 95% of individuals with known esophageal varices did not receive beta blocker therapy for primary prophylaxis. Of greater concern was that only 20% of individuals with a prior history of variceal bleeding were using beta blocker therapy prior to hospital admission. An improved rate of administration at 48% was observed upon discharge yet remains lower than acceptable.The Institute of Medicine has defined several criteria which recognize chronic diseases of importance that require quality of care measurement to reduce practice variation and improve clinical outcomes (Acad Med 2002;77:91–99). These include safety, efficacy, patient-centeredness, timeliness, efficiency, and equity. Based on these recommendations, cirrhosis and its complications may be considered one of these conditions. However, there has been scant information to date about how to define and measure quality of care for patients with cirrhosis. This is despite the fact that several data sources exist including clinical practice guidelines, which incorporate recommendations based on evidence-based data (Am J Gastroenterol 1997;92:1081–1091, Hepatology 1998;28:868–880, Gut 2000;46:S3–S4:III1–III15). In turn, a number of potential quality indicators (or performance measures) identified from these guidelines could be studied to determine their reliability and validity for assessing quality of care.In the field of quality improvement, a performance measure is used as an operational term to quantify the extent of quality health care that is provided. The most sound performance measures are those supported by high levels of evidence-based medicine, mainly through large or numerous randomized controlled trials. The failure to adhere to these performance measures is usually associated with a significant reduction in the ability to achieve favorable clinical outcomes (Circulation 2005;111:1703–1712). One possible example in patients with cirrhosis is the use of antibiotic prophylaxis for gastrointestinal bleeding where the number of serious infectious events and mortality rates are significantly reduced compared to no intervention (Hepatology 1999;29:1655–1661). A number of performance measures, however, are considered important components of clinical care yet often lack the weight of evidence from controlled trials. These measures are less easy to use and require methodical approaches to definition (Med Care 1999;37:964–968).The study by Wilbur and Sidhu further highlights the underuse of beta blocker therapy for primary and secondary prophylaxis in patients with cirrhosis and esophageal varices. These data are comparable to 4 reported studies of clinical practice (Am J Gastroenterol 2003;98:653–659, Hepatology 2003;38:599–612, Am J Gastroenterol 2003;98:2424–2434, J Hepatol 2003;39:509–514, Am J Gastroenterol 2004;99:645–649) in which similar rates of beta blocker use for primary prophylaxis (20%–30%) and secondary prophylaxis (45%–80%) are reported. Although a significant majority of patients in this study did not receive beta blocker therapy as primary prophylaxis, a number of factors have to be considered when interpreting the results.The authors describe a lack of knowledge regarding the size of esophageal varices in this patient population. Prior controlled clinical trials showed a small risk reduction (5%) with beta blocker therapy compared to placebo among patients with small esophageal varices (Semin Liver Dis 1999;19:475–505, Hepatology 1997;25:63–70). As a result, the use of beta blockers as primary prophylaxis for these individuals has not been routinely performed in clinical practice. A recent study, however, described clinical benefit for preventing growth of small to large esophageal varices on beta blockers which, in turn, reduced the risk for initial variceal bleeding (Gastroenterology 2004;127:476–484). These data require additional confirmation in various populations before a change in medical practice can be recommended.Data on the frequency of initial and subsequent endoscopy for detecting large esophageal varices for patients in this study was also not available. Thus, a low screening rate may have contributed to the increased number of patients who were not offered beta blocker therapy. More importantly, the presence of low utilization rates with beta blocker therapy as primary prophylaxis after identifying large esophageal varices by screening endoscopy would also be considered a critical gap in quality of care for this population. Unfortunately, the extent to which high-risk patients are actually receiving beta blocker therapy as primary prophylaxis is unknown.A significant proportion of patients with severe advanced liver disease (proxied by Child-Pugh class C) will not tolerate long-term beta blocker therapy due to an enhanced risk for side effects (Semin Liver Dis 1999;19:475–505, Hepatology 1997;25:63–70). Endoscopic variceal bland ligation for primary prophylaxis has now emerged as a viable alternative for patients who are intolerant or have contraindications to beta blocker treatment (Hepatology 2001;33:802–807). In this study, the frequency of band ligation in this population was not available for analysis. Nevertheless, a significant percentage of patients did not have absolute contraindications for the use of beta blocker therapy as prophylaxis against variceal bleeding. In this context, it would be difficult to attribute a low rate of beta blocker use based on patient ineligibility.A striking number of patients received inappropriate forms of pharmacologic therapy for prophylaxis against variceal bleeding. The use of oral nitrate therapy (with or without beta blockers) as primary and secondary prevention was observed, albeit in <5% of patients. However, this should be considered an inappropriate strategy based on a lack of evidence to support its use (Am J Med 2004;116:759–766, Semin Liver Dis 1999;19:475–505). In contrast, an estimated 25%–33% of patients were receiving selective or unproven beta blocker therapies as primary and secondary prophylaxis against variceal bleeding. This also constitutes inappropriate or misuse of evidence-based therapy and should be considered an element of poor quality care in this population.Information regarding duration of treatment, extent of adherence to treatment, and treatment efficacy as measured by hepatic vein pressure gradient measurements were also not available in this study. These factors have previously been described as important variables that contribute to the effectiveness of pharmacologic therapy as prophylaxis (Lancet 2003;361:952–954, Hepatology 2001;34:1096–1102). In turn, this information is also critical for determining extent of quality of medical care for improving clinical outcomes.Despite limitations in study design and missing data elements, this study further highlights the gap in using beta blockers as primary and secondary prophylaxis against esophageal variceal hemorrhage in patients with cirrhosis. These data are consistent with reports that at least 50% of individuals with chronic disease are undertreated with existing medical therapies deemed to be effective (N Engl J Med 2003;348:2635–2645). What can be done to improve these rates? Several interventions have been tried including the passive inoculation of clinical practice guidelines as well as formal instruction through continuing medical education formats. Although not specifically investigated for patients with cirrhosis, these methods have been found ineffective in changing provider attitude and reducing practice variation (JAMA 2004;291:2466–2470). With decision support tools and the electronic medical record, a strategy using timely reminders can be implemented to ensure that diagnostic endoscopy for identifying large esophageal varices is performed in all patients with cirrhosis. Similarly, the use of surveillance endoscopy performed at appropriate intervals for individuals with no or small varices can also be facilitated. Active interventions supported by opinion leaders and patient care teams do appear to improve quality of care in other fields (Health Affairs 2003;22:196–201) but require study to determine their true effectiveness in hepatology. Despite the existence of multiple, high-quality, randomized controlled trials that support the use of beta blocker therapy for primary and secondary prophylaxis against esophageal variceal bleeding, a significant gap exists with the translation of these scientific results into clinical practice. In this population, nearly 95% of individuals with known esophageal varices did not receive beta blocker therapy for primary prophylaxis. Of greater concern was that only 20% of individuals with a prior history of variceal bleeding were using beta blocker therapy prior to hospital admission. An improved rate of administration at 48% was observed upon discharge yet remains lower than acceptable. The Institute of Medicine has defined several criteria which recognize chronic diseases of importance that require quality of care measurement to reduce practice variation and improve clinical outcomes (Acad Med 2002;77:91–99). These include safety, efficacy, patient-centeredness, timeliness, efficiency, and equity. Based on these recommendations, cirrhosis and its complications may be considered one of these conditions. However, there has been scant information to date about how to define and measure quality of care for patients with cirrhosis. This is despite the fact that several data sources exist including clinical practice guidelines, which incorporate recommendations based on evidence-based data (Am J Gastroenterol 1997;92:1081–1091, Hepatology 1998;28:868–880, Gut 2000;46:S3–S4:III1–III15). In turn, a number of potential quality indicators (or performance measures) identified from these guidelines could be studied to determine their reliability and validity for assessing quality of care. In the field of quality improvement, a performance measure is used as an operational term to quantify the extent of quality health care that is provided. The most sound performance measures are those supported by high levels of evidence-based medicine, mainly through large or numerous randomized controlled trials. The failure to adhere to these performance measures is usually associated with a significant reduction in the ability to achieve favorable clinical outcomes (Circulation 2005;111:1703–1712). One possible example in patients with cirrhosis is the use of antibiotic prophylaxis for gastrointestinal bleeding where the number of serious infectious events and mortality rates are significantly reduced compared to no intervention (Hepatology 1999;29:1655–1661). A number of performance measures, however, are considered important components of clinical care yet often lack the weight of evidence from controlled trials. These measures are less easy to use and require methodical approaches to definition (Med Care 1999;37:964–968). The study by Wilbur and Sidhu further highlights the underuse of beta blocker therapy for primary and secondary prophylaxis in patients with cirrhosis and esophageal varices. These data are comparable to 4 reported studies of clinical practice (Am J Gastroenterol 2003;98:653–659, Hepatology 2003;38:599–612, Am J Gastroenterol 2003;98:2424–2434, J Hepatol 2003;39:509–514, Am J Gastroenterol 2004;99:645–649) in which similar rates of beta blocker use for primary prophylaxis (20%–30%) and secondary prophylaxis (45%–80%) are reported. Although a significant majority of patients in this study did not receive beta blocker therapy as primary prophylaxis, a number of factors have to be considered when interpreting the results. The authors describe a lack of knowledge regarding the size of esophageal varices in this patient population. Prior controlled clinical trials showed a small risk reduction (5%) with beta blocker therapy compared to placebo among patients with small esophageal varices (Semin Liver Dis 1999;19:475–505, Hepatology 1997;25:63–70). As a result, the use of beta blockers as primary prophylaxis for these individuals has not been routinely performed in clinical practice. A recent study, however, described clinical benefit for preventing growth of small to large esophageal varices on beta blockers which, in turn, reduced the risk for initial variceal bleeding (Gastroenterology 2004;127:476–484). These data require additional confirmation in various populations before a change in medical practice can be recommended. Data on the frequency of initial and subsequent endoscopy for detecting large esophageal varices for patients in this study was also not available. Thus, a low screening rate may have contributed to the increased number of patients who were not offered beta blocker therapy. More importantly, the presence of low utilization rates with beta blocker therapy as primary prophylaxis after identifying large esophageal varices by screening endoscopy would also be considered a critical gap in quality of care for this population. Unfortunately, the extent to which high-risk patients are actually receiving beta blocker therapy as primary prophylaxis is unknown. A significant proportion of patients with severe advanced liver disease (proxied by Child-Pugh class C) will not tolerate long-term beta blocker therapy due to an enhanced risk for side effects (Semin Liver Dis 1999;19:475–505, Hepatology 1997;25:63–70). Endoscopic variceal bland ligation for primary prophylaxis has now emerged as a viable alternative for patients who are intolerant or have contraindications to beta blocker treatment (Hepatology 2001;33:802–807). In this study, the frequency of band ligation in this population was not available for analysis. Nevertheless, a significant percentage of patients did not have absolute contraindications for the use of beta blocker therapy as prophylaxis against variceal bleeding. In this context, it would be difficult to attribute a low rate of beta blocker use based on patient ineligibility. A striking number of patients received inappropriate forms of pharmacologic therapy for prophylaxis against variceal bleeding. The use of oral nitrate therapy (with or without beta blockers) as primary and secondary prevention was observed, albeit in <5% of patients. However, this should be considered an inappropriate strategy based on a lack of evidence to support its use (Am J Med 2004;116:759–766, Semin Liver Dis 1999;19:475–505). In contrast, an estimated 25%–33% of patients were receiving selective or unproven beta blocker therapies as primary and secondary prophylaxis against variceal bleeding. This also constitutes inappropriate or misuse of evidence-based therapy and should be considered an element of poor quality care in this population. Information regarding duration of treatment, extent of adherence to treatment, and treatment efficacy as measured by hepatic vein pressure gradient measurements were also not available in this study. These factors have previously been described as important variables that contribute to the effectiveness of pharmacologic therapy as prophylaxis (Lancet 2003;361:952–954, Hepatology 2001;34:1096–1102). In turn, this information is also critical for determining extent of quality of medical care for improving clinical outcomes. Despite limitations in study design and missing data elements, this study further highlights the gap in using beta blockers as primary and secondary prophylaxis against esophageal variceal hemorrhage in patients with cirrhosis. These data are consistent with reports that at least 50% of individuals with chronic disease are undertreated with existing medical therapies deemed to be effective (N Engl J Med 2003;348:2635–2645). What can be done to improve these rates? Several interventions have been tried including the passive inoculation of clinical practice guidelines as well as formal instruction through continuing medical education formats. Although not specifically investigated for patients with cirrhosis, these methods have been found ineffective in changing provider attitude and reducing practice variation (JAMA 2004;291:2466–2470). With decision support tools and the electronic medical record, a strategy using timely reminders can be implemented to ensure that diagnostic endoscopy for identifying large esophageal varices is performed in all patients with cirrhosis. Similarly, the use of surveillance endoscopy performed at appropriate intervals for individuals with no or small varices can also be facilitated. Active interventions supported by opinion leaders and patient care teams do appear to improve quality of care in other fields (Health Affairs 2003;22:196–201) but require study to determine their true effectiveness in hepatology.