Peter C. Smith, MD

Evidence summary

Truly isolated hypertriglyceridemia is rare. To date, no good trials directly address the effect of reducing truly isolated hypertriglyceridemia on cardiovascular morbidity or mortality. High triglycerides are usually accompanied by other features of the “metabolic syndrome” (low HDL, high LDL, insulin resistance, diabetes, hypertension, and obesity), making it almost impossible to look at these in isolation or attribute risk to a specific component.¹

Whether high triglyceride levels pose risk in the true absence of these other metabolic factors is controversial. One meta-analysis of 17 population-based prospective studies of triglycerides and cardiovascular disease (including 57,000 patients) showed high triglyceride levels to be predictive of cardiac events, even when adjusted for HDL and other risk factors (age, total and LDL cholesterol, smoking, body mass index, and blood pressure).² After adjusting for these other risk factors, the authors found an increased risk for all cardiac endpoints (myocardial infarction, death, etc) of 14% for men and 32% for women (Men: relative risk [RR]=1.14; 95% confidence interval [CI], 1.05–1.28; Women: RR=1.37; 95% CI, 1.13–1.66).

Another meta-analysis of 3 prospective intervention trials with 15,880 enrolled subjects found that triglyceride levels did not provide any clinically meaningful information about risk beyond that provided by other cholesterol subfractions.³

In treatment trials, the most impressive risk reductions come from the groups who fit the lipid triad of low HDL, high LDL, and high triglycerides. Low levels of HDL appear to interact with hypertriglyceridemia to increase coronary risk, and all studies showing improved outcomes have simultaneously increased HDL while lowering triglycerides.^4-6 In 3 large-scale prospective, placebo-controlled trials (the Helsinki Heart Study, a primary prevention study, and the VA-HIT and Bezafibrate Infarction Prevention trials, both secondary prevention studies), lowering triglycerides and raising HDL concurrently improved outcomes.⁵ Successful dietary and medical interventions, especially with statins and fibrates, improved overall lipid profiles—not just triglyceride levels.

Accordingly, elevated triglycerides should prompt providers to rigorously identify these other risk factors for cardiovascular morbidity and mortality, which may not be immediately obvious. In the absence of such other factors, no evidence exists to guide therapy. Expert opinion^7,8 supported by epidemiologic evidence⁹ suggests that patients with triglyceride levels of 500 to 1000 mg/dL may have an increased risk of pancreatitis. Accordingly, providers should consider therapy to lower triglycerides to less than 500 in these patients, regardless of accompanying risk factors.

Recommendations from others

The American College of Physicians, the European Society of Cardiology, and the US Preventive Services Task Force do not recommend screening for hypertriglyceridemia. Clinical guidelines of the National Cholesterol Education Program/Adult Treatment Panel III (NCEP/ATP III), American Heart Association/American College of Cardiology, and the American Diabetes Association all support LDL lowering as the primary target of therapy based on the patients risk profile.¹⁰ NCEP/ATP III has identified triglyceride levels of <150 as normal, 150–199 as borderline high, 200–499 as high, and 500 as very high.7

A patient with high triglycerides should prompt a search for components of the “meta-bolic syndrome” and secondary causes, including high dietary fat, high alcohol intake, drugs (steroids, beta-blockers, high-estrogen oral contraceptives), medical conditions (hypothyroidism, nephrosis, renal failure, liver disease, Cushing disease, and lupus), and rare familial dyslipidemias.^7,10,11

Evidence summary

Truly isolated hypertriglyceridemia is rare. To date, no good trials directly address the effect of reducing truly isolated hypertriglyceridemia on cardiovascular morbidity or mortality. High triglycerides are usually accompanied by other features of the “metabolic syndrome” (low HDL, high LDL, insulin resistance, diabetes, hypertension, and obesity), making it almost impossible to look at these in isolation or attribute risk to a specific component.¹

Whether high triglyceride levels pose risk in the true absence of these other metabolic factors is controversial. One meta-analysis of 17 population-based prospective studies of triglycerides and cardiovascular disease (including 57,000 patients) showed high triglyceride levels to be predictive of cardiac events, even when adjusted for HDL and other risk factors (age, total and LDL cholesterol, smoking, body mass index, and blood pressure).² After adjusting for these other risk factors, the authors found an increased risk for all cardiac endpoints (myocardial infarction, death, etc) of 14% for men and 32% for women (Men: relative risk [RR]=1.14; 95% confidence interval [CI], 1.05–1.28; Women: RR=1.37; 95% CI, 1.13–1.66).

Another meta-analysis of 3 prospective intervention trials with 15,880 enrolled subjects found that triglyceride levels did not provide any clinically meaningful information about risk beyond that provided by other cholesterol subfractions.³

In treatment trials, the most impressive risk reductions come from the groups who fit the lipid triad of low HDL, high LDL, and high triglycerides. Low levels of HDL appear to interact with hypertriglyceridemia to increase coronary risk, and all studies showing improved outcomes have simultaneously increased HDL while lowering triglycerides.^4-6 In 3 large-scale prospective, placebo-controlled trials (the Helsinki Heart Study, a primary prevention study, and the VA-HIT and Bezafibrate Infarction Prevention trials, both secondary prevention studies), lowering triglycerides and raising HDL concurrently improved outcomes.⁵ Successful dietary and medical interventions, especially with statins and fibrates, improved overall lipid profiles—not just triglyceride levels.

Accordingly, elevated triglycerides should prompt providers to rigorously identify these other risk factors for cardiovascular morbidity and mortality, which may not be immediately obvious. In the absence of such other factors, no evidence exists to guide therapy. Expert opinion^7,8 supported by epidemiologic evidence⁹ suggests that patients with triglyceride levels of 500 to 1000 mg/dL may have an increased risk of pancreatitis. Accordingly, providers should consider therapy to lower triglycerides to less than 500 in these patients, regardless of accompanying risk factors.

Recommendations from others

The American College of Physicians, the European Society of Cardiology, and the US Preventive Services Task Force do not recommend screening for hypertriglyceridemia. Clinical guidelines of the National Cholesterol Education Program/Adult Treatment Panel III (NCEP/ATP III), American Heart Association/American College of Cardiology, and the American Diabetes Association all support LDL lowering as the primary target of therapy based on the patients risk profile.¹⁰ NCEP/ATP III has identified triglyceride levels of <150 as normal, 150–199 as borderline high, 200–499 as high, and 500 as very high.7

A patient with high triglycerides should prompt a search for components of the “meta-bolic syndrome” and secondary causes, including high dietary fat, high alcohol intake, drugs (steroids, beta-blockers, high-estrogen oral contraceptives), medical conditions (hypothyroidism, nephrosis, renal failure, liver disease, Cushing disease, and lupus), and rare familial dyslipidemias.^7,10,11

Evidence summary

Truly isolated hypertriglyceridemia is rare. To date, no good trials directly address the effect of reducing truly isolated hypertriglyceridemia on cardiovascular morbidity or mortality. High triglycerides are usually accompanied by other features of the “metabolic syndrome” (low HDL, high LDL, insulin resistance, diabetes, hypertension, and obesity), making it almost impossible to look at these in isolation or attribute risk to a specific component.¹

Whether high triglyceride levels pose risk in the true absence of these other metabolic factors is controversial. One meta-analysis of 17 population-based prospective studies of triglycerides and cardiovascular disease (including 57,000 patients) showed high triglyceride levels to be predictive of cardiac events, even when adjusted for HDL and other risk factors (age, total and LDL cholesterol, smoking, body mass index, and blood pressure).² After adjusting for these other risk factors, the authors found an increased risk for all cardiac endpoints (myocardial infarction, death, etc) of 14% for men and 32% for women (Men: relative risk [RR]=1.14; 95% confidence interval [CI], 1.05–1.28; Women: RR=1.37; 95% CI, 1.13–1.66).

Another meta-analysis of 3 prospective intervention trials with 15,880 enrolled subjects found that triglyceride levels did not provide any clinically meaningful information about risk beyond that provided by other cholesterol subfractions.³

In treatment trials, the most impressive risk reductions come from the groups who fit the lipid triad of low HDL, high LDL, and high triglycerides. Low levels of HDL appear to interact with hypertriglyceridemia to increase coronary risk, and all studies showing improved outcomes have simultaneously increased HDL while lowering triglycerides.^4-6 In 3 large-scale prospective, placebo-controlled trials (the Helsinki Heart Study, a primary prevention study, and the VA-HIT and Bezafibrate Infarction Prevention trials, both secondary prevention studies), lowering triglycerides and raising HDL concurrently improved outcomes.⁵ Successful dietary and medical interventions, especially with statins and fibrates, improved overall lipid profiles—not just triglyceride levels.

Accordingly, elevated triglycerides should prompt providers to rigorously identify these other risk factors for cardiovascular morbidity and mortality, which may not be immediately obvious. In the absence of such other factors, no evidence exists to guide therapy. Expert opinion^7,8 supported by epidemiologic evidence⁹ suggests that patients with triglyceride levels of 500 to 1000 mg/dL may have an increased risk of pancreatitis. Accordingly, providers should consider therapy to lower triglycerides to less than 500 in these patients, regardless of accompanying risk factors.

Recommendations from others

The American College of Physicians, the European Society of Cardiology, and the US Preventive Services Task Force do not recommend screening for hypertriglyceridemia. Clinical guidelines of the National Cholesterol Education Program/Adult Treatment Panel III (NCEP/ATP III), American Heart Association/American College of Cardiology, and the American Diabetes Association all support LDL lowering as the primary target of therapy based on the patients risk profile.¹⁰ NCEP/ATP III has identified triglyceride levels of <150 as normal, 150–199 as borderline high, 200–499 as high, and 500 as very high.7

A patient with high triglycerides should prompt a search for components of the “meta-bolic syndrome” and secondary causes, including high dietary fat, high alcohol intake, drugs (steroids, beta-blockers, high-estrogen oral contraceptives), medical conditions (hypothyroidism, nephrosis, renal failure, liver disease, Cushing disease, and lupus), and rare familial dyslipidemias.^7,10,11

The hospitalist movement has promised to improve the quality of inpatient care, increase patient satisfaction, and decrease costs.¹ Many hospitals, practices, and managed care corporations have adopted this model of care,² but whether this model has fulfilled its promises is unknown. Those who favor hospitalists have argued that hospitalists offer more efficient care by increasing quality and decreasing costs. Detractors are concerned about potential substandard quality through aggressive discharge policies and loss of continuity of care. Unfortunately, both positions are based largely on untested assumptions. We identified 6 peer-reviewed articles directly comparing hospitalists and primary care physicians.^3-8 Another 314 were descriptive studies, editorials, letters, and news pieces arguing about the potential risks and benefits of the hospitalist movement.

Hospitalists have been described as physicians who spend over one fourth of their time exclusively in the hospital caring for other physicians’ patients only during that admission.⁹ Others believe the hospitalist movement more accurately encompasses a broad spectrum of how inpatient care is organized,¹⁰ including primary care physicians managing their own inpatients and seeing clinic patients, primary care physicians sharing week- or month-long periods of exclusive hospital care with partners, or excluding the primary care physician from inpatient care by using dedicated inpatient-only physicians who may be family physicians, internists, or specialists.

The scant literature comparing care provided by hospitalists and primary care physicians has several methodologic constraints including before and after designs that may have time-effect bias,^11-15 inappropriately assigning subspecialists to the primary care group,³ restricting efficiency tools such as nurse managers and discharge planners to the intervention group,^4-6 failing to account for differential involvement of house staff,^7,11 using possibly unreliable outcomes,⁸ and relying exclusively on claims data.^11,14 Two recent studies avoided many of these pitfalls and found no differences between different types of hospitalists, but did not compare them with primary care physicians.^16,17 We designed our study to address multiple methodologic concerns and determine whether differences in outcomes, processes of care, and costs exist between these multiple models of inpatient care.

Methods

Setting

In 1997 a large regional health maintenance organization in Colorado mandated that all its inpatients be admitted by a pulmonology or critical care hospitalist team to the exclusion of their primary care physicians. Rose Medical Center, a 420-bed private community hospital in Denver, Colorado, serves as a family practice residency training site in which residents care for patients under the guidance of resident faculty and community primary care physicians. We recognized the health maintenance orgaization’s program as a natural experiment and an opportunity to address some of the design limitations of prior studies by comparing the care delivered simultaneously by these 3 inpatient models.

Subjects and study design

We conducted a retrospective cohort study of all patients admitted between April 1997 and March 1998 with a primary diagnosis of pneumonia as identified by codes from the International Classification of Diseases, Ninth Revision. We studied pneumonia care because of the high incidence of pneumonia in our institution and the existence of a valid, population-based measure of disease severity, the Pneumonia Severity Index (PSI; see Statistical Methods). In addition, focusing on 1 diagnosis allowed for a direct and detailed analysis of the process of care. To eliminate potential biases produced by different outpatient physician specialties, we excluded patients who did not have a family physician as a primary care provider. Patients also were excluded if they were younger than 18 years, had human immunodeficiency virus or acquired immunodeficiency syndrome, had exclusively nosocomial pneumonia, or had the diagnosis of pneumonia subsequently ruled out. Ninety-seven patients were included. Data were collected by standardized chart abstraction and entered into a Microsoft Access database. The chart abstraction was performed by research assistants who had no knowledge of whether the attending physician was a primary care physician or a hospitalist.

Patients were grouped based on the status of their inpatient attending physicians of record. The critical care hospitalists represented a group of subspecialty critical care pulmonologists contracted by a large health maintenance organization to care for all its inpatients. The family physician hospitalists represented rotating family medicine faculty who worked exclusively in the hospital for 8 weeks each year. Some patients under the care of family physician hospitalists originated from the residency’s continuity clinic, and others came from the practices of approximately 30 community family physicians for whom the family physician hospitalist acts as a hospitalist service. Community family medicine primary care physicians cared for their own patients and continued their outpatient practices. Family medicine house staff was involved in the care of inpatients from all groups.

Patients’ insurance plans rather than physician referral or self-referral determined whether they were assigned to the critical care hospitalist, family physician hospitalist, or primary care physician group. Accordingly, there were instances when a primary care physician had some patients admitted to the critical care hospitalist because the patient was a member of this particular health maintenance organization and also because the physician cared directly for his other patients in the hospital.

Outcome measures

We looked at 5 primary patient- and policy-oriented outcomes that have been validated as indicators of quality of inpatient care: in-hospital mortality, length of stay, hospital charges, 7-day readmissions, and 7- and 30-day returns to the emergency department.^18,19 We included Colorado data from the Healthcare Cost and Utilization Project, when available, to serve as a reference standard.²⁰ In addition, we chose 7 validated secondary “process of care”outcomes²¹ to further describe the practice behaviors of the 3 groups: documentation of lifestyle modification counseling (tobacco cessation, exercise, etc), documentation of end-of-life counseling, compliance with contemporary guidelines from the American Thoracic Society for treating community-acquired pneumonia,²¹ length of stay in intensive care, and the use of chest radiographs and blood and sputum cultures.

Statistical methods

All statistical analyses were performed with SAS version 6.12. Patient demographic and clinical characteristics were analyzed with chi-square test and analysis of variance, when appropriate. We controlled for disease severity with the PSI, a well-validated, population-based severity of illness score for inpatients with pneumonia.^22,23

Multiple and logistic regressions were used to control for disease severity and potential confounders. Our models included the PSI class and those characteristics that were statistically significantly different (ie, diagnosis of hypertension) or showed a trend toward difference and were felt to be potentially clinically significant (diagnosis of diabetes, effusion on chest x-ray, mental status at admission), in addition to sex and age. Logarithmic transformations of non-normal data were conducted, when appropriate. We eliminated as extreme outliers 3 cases (2 under the care of critical care hospitalists and 1 under the care of family physician hospitalists) whose outcomes were 2 standard deviations or more beyond the mean. For example, 1 patient had a rare clotting disorder, stayed in intensive care for 30 days, required an orphan drug at $6000 per dose, and had charges well in excess of $1 million. This study was approved by the Rose Medical Center and Health One Institutional Review Board.

Results

Demographic and clinical characteristics

Table 1 provides the demographic and clinical descriptions of patients admitted by the different admitting physician models. Patients in the different groups were similar, with 2 exceptions: hypertension was significantly more common in the critical care hospitalist group than in the other groups (P < .05), and there was a trend toward more diabetes in the critical care hospitalist group that did not quite reach statistical significance. Nonsignificant trends also existed for PSI, effusion on chest x-ray, and mental status, with more effusions and acute mental status changes occurring in the critical care hospitalist group, more chronic altered mental status in the family physician hospitalist group, and greater severity of illness in the critical care hospitalist group. Otherwise, demographics, disease severity, and comorbidities were comparable.

TABLE 1
Demographic and clinical characteristics

Primary outcomes

After controlling for severity of illness and intergroup differences, we found that the critical care hospitalist team had the highest mean hospital charge ($10,231), followed by the family physician hospitalist ($7699) and the primary care physician ($5680) groups (Figure 1). The difference in charges between the primary care physician and the critical care hospitalist groups was statistically significant (P = .005) and approached significance between the primary care physician and family physician hospitalist groups (P = .08). The critical care hospitalist and family physician hospitalist groups had longer mean lengths of stay than did the primary care physician group (P = .04 and .01, respectively; Figure 2). The other primary outcomes were rare: 1 primary care physician patient died (4.5%), 2 critical care hospitalist patients died (9.5%) and no family physician hospitalist patients died; no primary care physician patients were readmitted, 1 critical care hospitalist patient was readmitted (4.8%), and 2 family physician hospitalist patients were readmitted (3.8%). There was 1 return to the emergency room in the cohort, in the family physician hospitalist group (1.9%). No intergroup comparisons between these unadjusted rates were statistically significant (P > .05).

FIGURE 1
Hospital charges (in US dollars; with 95% confidence intervals)*

FIGURE 2
Length of stay (in days; with 95% confidence intervals)*

Secondary outcomes

After controlling for severity of illness and intergroup differences, we found that the critical care hospitalists were more likely to obtain 2 or more chest x-rays than the primary care physicians. There were nonsignificant trends toward longer stays in intensive care, greater likelihood of obtaining sputum cultures, and documenting end-of-life counseling by the critical care hospitalists compared with the primary care physicians. For the family physician hospitalists, there were nonsignificant trends toward better compliance with American Thoracic Society antibiotic guidelines and greater likelihood of documenting end-of-life and lifestyle modification counseling compared with the primary care physicians (Table 2).

TABLE 2
Secondary “process of care” outcomes*

Discussion

Our study provided a unique perspective on the impact of different models of caring for inpatients on the quality, processes, and cost of care. We believe this is the first study to successfully address several methodologic limitations of previous studies: potential time-effect bias, inappropriate controls, differential assignment of house staff and case management resources, unvalidated outcomes, and lack of clinical data.

In addition, we reduced the potential biases inherent in comparing different hospitals and different outpatient physician specialties and used a standardized chart abstraction instrument to avoid the problems inherent in using claims data. As a result, we were able to examine processes of care and use the PSI, an extensively validated tool, to control for disease severity. This study represents an effectiveness study of a real-life intervention by a health maintenance organization to mandate the use of hospitalists. Although the retrospective design of this study may create the potential for bias, there was inadequate advanced notice of the implementation of this hospitalist plan to allow for prospective analysis. Despite not being randomly assigned to 1 of the 3 groups, patients were quite similar with the exceptions of hypertension and possibly diabetes and PSI, which we controlled for in the statistical analysis.

Unfortunately, this was a small study that lacked sufficient power to detect modest differences between groups because the health maintenance organization sponsoring the critical care hospitalist group abandoned the program after 1 year. In addition, the differences in disease severity might have been significant in a larger sample. However, even after controlling for these differences statistically, we found no large differences for mortality, readmission, or returns to the emergency room.

Despite insufficient power to observe statistically significant differences in these relatively rare but important pneumonia outcomes, we did detect a substantial difference in adjusted hospital charges and a modest difference in length of stay. Subspecialist hospitalists had significantly higher adjusted charges than did primary care family physicians. Although the comparison across groups failed to show statistically significant differences, we did see a trend of increasing charges as the degree of hospitalization increased. These higher costs may be explained in part by primary care physicians advising shorter lengths of stay and the subspecialists’ increased use of multiple chest x-rays and trends toward greater use of other resources (eg, intensive care and blood and sputum cultures). Alternatively, some of the difference in charges may reflect differing levels of continuity; the critical care hospitalists had no outpatient continuity with their inpatients, whereas the family physician hospitalists had continuity relationships with some inpatients and the primary care physicians had relationships with all their inpatients. Thus the primary care physicians and, to a lesser extent, the family physician hospitalists may have had information about prior care. Hence, knowledge of previous antibiotic use might argue for the low yield of blood and sputum cultures, and having obtained an outpatient x-ray might obviate the need for another in the hospital. The critical care hospitalists’ increased length of stay and x-ray use, in conjunction with the trend toward greater use of cultures and intensive care, may in turn reflect different degrees of comfort with uncertainty between family physicians and subspecialists. Also, we examined only hospital charges rather than total costs to the system.

Interestingly, we found a trend showing that family physician hospitalists were more likely to document lifestyle modification counseling than were primary care physicians. This result should be interpreted with some caution. Our findings may indicate a true lack of performance by primary care physicians, or they may show a failure to document advice on the hospital chart, reflecting some aspect of the continuity relationship in which such discussions are relegated to the outpatient setting.We also were surprised to see the trend toward decreased end-of-life counseling by the primary care physicians. This could reflect some adverse effect of continuity, the time constraints imposed on nonhospitalists, or not documenting outpatient counseling on the inpatient record.

There were other potential sources of confounding in this study. All patients in the critical care hospitalist group were members of the same health maintenance organization, which may have introduced unmeasured bias despite our attempts to control for differences between groups. Even though we purposefully avoided differential use of house staff, its involvement in each case may have decreased any potential differences across practices.

We draw 2 important conclusions from our results. First, our findings of increased costs and length of stay for mandated hospitalists without significantly different outcomes support the assertion of the American Academy of Family Physicians, the American College of Physicians–American Society of Internal Medicine, the American Medical Association, and the National Association of Inpatient Physicians: the practice of mandating the use of hospitalists should be abandoned pending larger, more comprehensive contemporaneous trials. Second, if hospitalists are to be employed on a voluntary basis, the use of subspecialists rather than generalists may result in more costly care.

Acknowledgments

The authors thank Ralph B. D’Agonstino, Jr., PhD, of the Bowman Gray School of Medicine for his invaluable assistance in statistical methods and Cary Foster, MD, and Promoda Mahupatra, MD, for their collection of the data.

The hospitalist movement has promised to improve the quality of inpatient care, increase patient satisfaction, and decrease costs.¹ Many hospitals, practices, and managed care corporations have adopted this model of care,² but whether this model has fulfilled its promises is unknown. Those who favor hospitalists have argued that hospitalists offer more efficient care by increasing quality and decreasing costs. Detractors are concerned about potential substandard quality through aggressive discharge policies and loss of continuity of care. Unfortunately, both positions are based largely on untested assumptions. We identified 6 peer-reviewed articles directly comparing hospitalists and primary care physicians.^3-8 Another 314 were descriptive studies, editorials, letters, and news pieces arguing about the potential risks and benefits of the hospitalist movement.

Hospitalists have been described as physicians who spend over one fourth of their time exclusively in the hospital caring for other physicians’ patients only during that admission.⁹ Others believe the hospitalist movement more accurately encompasses a broad spectrum of how inpatient care is organized,¹⁰ including primary care physicians managing their own inpatients and seeing clinic patients, primary care physicians sharing week- or month-long periods of exclusive hospital care with partners, or excluding the primary care physician from inpatient care by using dedicated inpatient-only physicians who may be family physicians, internists, or specialists.

The scant literature comparing care provided by hospitalists and primary care physicians has several methodologic constraints including before and after designs that may have time-effect bias,^11-15 inappropriately assigning subspecialists to the primary care group,³ restricting efficiency tools such as nurse managers and discharge planners to the intervention group,^4-6 failing to account for differential involvement of house staff,^7,11 using possibly unreliable outcomes,⁸ and relying exclusively on claims data.^11,14 Two recent studies avoided many of these pitfalls and found no differences between different types of hospitalists, but did not compare them with primary care physicians.^16,17 We designed our study to address multiple methodologic concerns and determine whether differences in outcomes, processes of care, and costs exist between these multiple models of inpatient care.

Methods

Setting

In 1997 a large regional health maintenance organization in Colorado mandated that all its inpatients be admitted by a pulmonology or critical care hospitalist team to the exclusion of their primary care physicians. Rose Medical Center, a 420-bed private community hospital in Denver, Colorado, serves as a family practice residency training site in which residents care for patients under the guidance of resident faculty and community primary care physicians. We recognized the health maintenance orgaization’s program as a natural experiment and an opportunity to address some of the design limitations of prior studies by comparing the care delivered simultaneously by these 3 inpatient models.

Subjects and study design

We conducted a retrospective cohort study of all patients admitted between April 1997 and March 1998 with a primary diagnosis of pneumonia as identified by codes from the International Classification of Diseases, Ninth Revision. We studied pneumonia care because of the high incidence of pneumonia in our institution and the existence of a valid, population-based measure of disease severity, the Pneumonia Severity Index (PSI; see Statistical Methods). In addition, focusing on 1 diagnosis allowed for a direct and detailed analysis of the process of care. To eliminate potential biases produced by different outpatient physician specialties, we excluded patients who did not have a family physician as a primary care provider. Patients also were excluded if they were younger than 18 years, had human immunodeficiency virus or acquired immunodeficiency syndrome, had exclusively nosocomial pneumonia, or had the diagnosis of pneumonia subsequently ruled out. Ninety-seven patients were included. Data were collected by standardized chart abstraction and entered into a Microsoft Access database. The chart abstraction was performed by research assistants who had no knowledge of whether the attending physician was a primary care physician or a hospitalist.

Patients were grouped based on the status of their inpatient attending physicians of record. The critical care hospitalists represented a group of subspecialty critical care pulmonologists contracted by a large health maintenance organization to care for all its inpatients. The family physician hospitalists represented rotating family medicine faculty who worked exclusively in the hospital for 8 weeks each year. Some patients under the care of family physician hospitalists originated from the residency’s continuity clinic, and others came from the practices of approximately 30 community family physicians for whom the family physician hospitalist acts as a hospitalist service. Community family medicine primary care physicians cared for their own patients and continued their outpatient practices. Family medicine house staff was involved in the care of inpatients from all groups.

Patients’ insurance plans rather than physician referral or self-referral determined whether they were assigned to the critical care hospitalist, family physician hospitalist, or primary care physician group. Accordingly, there were instances when a primary care physician had some patients admitted to the critical care hospitalist because the patient was a member of this particular health maintenance organization and also because the physician cared directly for his other patients in the hospital.

Outcome measures

We looked at 5 primary patient- and policy-oriented outcomes that have been validated as indicators of quality of inpatient care: in-hospital mortality, length of stay, hospital charges, 7-day readmissions, and 7- and 30-day returns to the emergency department.^18,19 We included Colorado data from the Healthcare Cost and Utilization Project, when available, to serve as a reference standard.²⁰ In addition, we chose 7 validated secondary “process of care”outcomes²¹ to further describe the practice behaviors of the 3 groups: documentation of lifestyle modification counseling (tobacco cessation, exercise, etc), documentation of end-of-life counseling, compliance with contemporary guidelines from the American Thoracic Society for treating community-acquired pneumonia,²¹ length of stay in intensive care, and the use of chest radiographs and blood and sputum cultures.

Statistical methods

All statistical analyses were performed with SAS version 6.12. Patient demographic and clinical characteristics were analyzed with chi-square test and analysis of variance, when appropriate. We controlled for disease severity with the PSI, a well-validated, population-based severity of illness score for inpatients with pneumonia.^22,23

Multiple and logistic regressions were used to control for disease severity and potential confounders. Our models included the PSI class and those characteristics that were statistically significantly different (ie, diagnosis of hypertension) or showed a trend toward difference and were felt to be potentially clinically significant (diagnosis of diabetes, effusion on chest x-ray, mental status at admission), in addition to sex and age. Logarithmic transformations of non-normal data were conducted, when appropriate. We eliminated as extreme outliers 3 cases (2 under the care of critical care hospitalists and 1 under the care of family physician hospitalists) whose outcomes were 2 standard deviations or more beyond the mean. For example, 1 patient had a rare clotting disorder, stayed in intensive care for 30 days, required an orphan drug at $6000 per dose, and had charges well in excess of $1 million. This study was approved by the Rose Medical Center and Health One Institutional Review Board.

Results

Demographic and clinical characteristics

Table 1 provides the demographic and clinical descriptions of patients admitted by the different admitting physician models. Patients in the different groups were similar, with 2 exceptions: hypertension was significantly more common in the critical care hospitalist group than in the other groups (P < .05), and there was a trend toward more diabetes in the critical care hospitalist group that did not quite reach statistical significance. Nonsignificant trends also existed for PSI, effusion on chest x-ray, and mental status, with more effusions and acute mental status changes occurring in the critical care hospitalist group, more chronic altered mental status in the family physician hospitalist group, and greater severity of illness in the critical care hospitalist group. Otherwise, demographics, disease severity, and comorbidities were comparable.

TABLE 1
Demographic and clinical characteristics

Primary outcomes

After controlling for severity of illness and intergroup differences, we found that the critical care hospitalist team had the highest mean hospital charge ($10,231), followed by the family physician hospitalist ($7699) and the primary care physician ($5680) groups (Figure 1). The difference in charges between the primary care physician and the critical care hospitalist groups was statistically significant (P = .005) and approached significance between the primary care physician and family physician hospitalist groups (P = .08). The critical care hospitalist and family physician hospitalist groups had longer mean lengths of stay than did the primary care physician group (P = .04 and .01, respectively; Figure 2). The other primary outcomes were rare: 1 primary care physician patient died (4.5%), 2 critical care hospitalist patients died (9.5%) and no family physician hospitalist patients died; no primary care physician patients were readmitted, 1 critical care hospitalist patient was readmitted (4.8%), and 2 family physician hospitalist patients were readmitted (3.8%). There was 1 return to the emergency room in the cohort, in the family physician hospitalist group (1.9%). No intergroup comparisons between these unadjusted rates were statistically significant (P > .05).

FIGURE 1
Hospital charges (in US dollars; with 95% confidence intervals)*

FIGURE 2
Length of stay (in days; with 95% confidence intervals)*

Secondary outcomes

After controlling for severity of illness and intergroup differences, we found that the critical care hospitalists were more likely to obtain 2 or more chest x-rays than the primary care physicians. There were nonsignificant trends toward longer stays in intensive care, greater likelihood of obtaining sputum cultures, and documenting end-of-life counseling by the critical care hospitalists compared with the primary care physicians. For the family physician hospitalists, there were nonsignificant trends toward better compliance with American Thoracic Society antibiotic guidelines and greater likelihood of documenting end-of-life and lifestyle modification counseling compared with the primary care physicians (Table 2).

TABLE 2
Secondary “process of care” outcomes*

Discussion

Our study provided a unique perspective on the impact of different models of caring for inpatients on the quality, processes, and cost of care. We believe this is the first study to successfully address several methodologic limitations of previous studies: potential time-effect bias, inappropriate controls, differential assignment of house staff and case management resources, unvalidated outcomes, and lack of clinical data.

In addition, we reduced the potential biases inherent in comparing different hospitals and different outpatient physician specialties and used a standardized chart abstraction instrument to avoid the problems inherent in using claims data. As a result, we were able to examine processes of care and use the PSI, an extensively validated tool, to control for disease severity. This study represents an effectiveness study of a real-life intervention by a health maintenance organization to mandate the use of hospitalists. Although the retrospective design of this study may create the potential for bias, there was inadequate advanced notice of the implementation of this hospitalist plan to allow for prospective analysis. Despite not being randomly assigned to 1 of the 3 groups, patients were quite similar with the exceptions of hypertension and possibly diabetes and PSI, which we controlled for in the statistical analysis.

Unfortunately, this was a small study that lacked sufficient power to detect modest differences between groups because the health maintenance organization sponsoring the critical care hospitalist group abandoned the program after 1 year. In addition, the differences in disease severity might have been significant in a larger sample. However, even after controlling for these differences statistically, we found no large differences for mortality, readmission, or returns to the emergency room.

Despite insufficient power to observe statistically significant differences in these relatively rare but important pneumonia outcomes, we did detect a substantial difference in adjusted hospital charges and a modest difference in length of stay. Subspecialist hospitalists had significantly higher adjusted charges than did primary care family physicians. Although the comparison across groups failed to show statistically significant differences, we did see a trend of increasing charges as the degree of hospitalization increased. These higher costs may be explained in part by primary care physicians advising shorter lengths of stay and the subspecialists’ increased use of multiple chest x-rays and trends toward greater use of other resources (eg, intensive care and blood and sputum cultures). Alternatively, some of the difference in charges may reflect differing levels of continuity; the critical care hospitalists had no outpatient continuity with their inpatients, whereas the family physician hospitalists had continuity relationships with some inpatients and the primary care physicians had relationships with all their inpatients. Thus the primary care physicians and, to a lesser extent, the family physician hospitalists may have had information about prior care. Hence, knowledge of previous antibiotic use might argue for the low yield of blood and sputum cultures, and having obtained an outpatient x-ray might obviate the need for another in the hospital. The critical care hospitalists’ increased length of stay and x-ray use, in conjunction with the trend toward greater use of cultures and intensive care, may in turn reflect different degrees of comfort with uncertainty between family physicians and subspecialists. Also, we examined only hospital charges rather than total costs to the system.

Interestingly, we found a trend showing that family physician hospitalists were more likely to document lifestyle modification counseling than were primary care physicians. This result should be interpreted with some caution. Our findings may indicate a true lack of performance by primary care physicians, or they may show a failure to document advice on the hospital chart, reflecting some aspect of the continuity relationship in which such discussions are relegated to the outpatient setting.We also were surprised to see the trend toward decreased end-of-life counseling by the primary care physicians. This could reflect some adverse effect of continuity, the time constraints imposed on nonhospitalists, or not documenting outpatient counseling on the inpatient record.

There were other potential sources of confounding in this study. All patients in the critical care hospitalist group were members of the same health maintenance organization, which may have introduced unmeasured bias despite our attempts to control for differences between groups. Even though we purposefully avoided differential use of house staff, its involvement in each case may have decreased any potential differences across practices.

We draw 2 important conclusions from our results. First, our findings of increased costs and length of stay for mandated hospitalists without significantly different outcomes support the assertion of the American Academy of Family Physicians, the American College of Physicians–American Society of Internal Medicine, the American Medical Association, and the National Association of Inpatient Physicians: the practice of mandating the use of hospitalists should be abandoned pending larger, more comprehensive contemporaneous trials. Second, if hospitalists are to be employed on a voluntary basis, the use of subspecialists rather than generalists may result in more costly care.

Acknowledgments

The authors thank Ralph B. D’Agonstino, Jr., PhD, of the Bowman Gray School of Medicine for his invaluable assistance in statistical methods and Cary Foster, MD, and Promoda Mahupatra, MD, for their collection of the data.

The hospitalist movement has promised to improve the quality of inpatient care, increase patient satisfaction, and decrease costs.¹ Many hospitals, practices, and managed care corporations have adopted this model of care,² but whether this model has fulfilled its promises is unknown. Those who favor hospitalists have argued that hospitalists offer more efficient care by increasing quality and decreasing costs. Detractors are concerned about potential substandard quality through aggressive discharge policies and loss of continuity of care. Unfortunately, both positions are based largely on untested assumptions. We identified 6 peer-reviewed articles directly comparing hospitalists and primary care physicians.^3-8 Another 314 were descriptive studies, editorials, letters, and news pieces arguing about the potential risks and benefits of the hospitalist movement.

Hospitalists have been described as physicians who spend over one fourth of their time exclusively in the hospital caring for other physicians’ patients only during that admission.⁹ Others believe the hospitalist movement more accurately encompasses a broad spectrum of how inpatient care is organized,¹⁰ including primary care physicians managing their own inpatients and seeing clinic patients, primary care physicians sharing week- or month-long periods of exclusive hospital care with partners, or excluding the primary care physician from inpatient care by using dedicated inpatient-only physicians who may be family physicians, internists, or specialists.

The scant literature comparing care provided by hospitalists and primary care physicians has several methodologic constraints including before and after designs that may have time-effect bias,^11-15 inappropriately assigning subspecialists to the primary care group,³ restricting efficiency tools such as nurse managers and discharge planners to the intervention group,^4-6 failing to account for differential involvement of house staff,^7,11 using possibly unreliable outcomes,⁸ and relying exclusively on claims data.^11,14 Two recent studies avoided many of these pitfalls and found no differences between different types of hospitalists, but did not compare them with primary care physicians.^16,17 We designed our study to address multiple methodologic concerns and determine whether differences in outcomes, processes of care, and costs exist between these multiple models of inpatient care.

Methods

Setting

In 1997 a large regional health maintenance organization in Colorado mandated that all its inpatients be admitted by a pulmonology or critical care hospitalist team to the exclusion of their primary care physicians. Rose Medical Center, a 420-bed private community hospital in Denver, Colorado, serves as a family practice residency training site in which residents care for patients under the guidance of resident faculty and community primary care physicians. We recognized the health maintenance orgaization’s program as a natural experiment and an opportunity to address some of the design limitations of prior studies by comparing the care delivered simultaneously by these 3 inpatient models.

Subjects and study design

We conducted a retrospective cohort study of all patients admitted between April 1997 and March 1998 with a primary diagnosis of pneumonia as identified by codes from the International Classification of Diseases, Ninth Revision. We studied pneumonia care because of the high incidence of pneumonia in our institution and the existence of a valid, population-based measure of disease severity, the Pneumonia Severity Index (PSI; see Statistical Methods). In addition, focusing on 1 diagnosis allowed for a direct and detailed analysis of the process of care. To eliminate potential biases produced by different outpatient physician specialties, we excluded patients who did not have a family physician as a primary care provider. Patients also were excluded if they were younger than 18 years, had human immunodeficiency virus or acquired immunodeficiency syndrome, had exclusively nosocomial pneumonia, or had the diagnosis of pneumonia subsequently ruled out. Ninety-seven patients were included. Data were collected by standardized chart abstraction and entered into a Microsoft Access database. The chart abstraction was performed by research assistants who had no knowledge of whether the attending physician was a primary care physician or a hospitalist.

Patients were grouped based on the status of their inpatient attending physicians of record. The critical care hospitalists represented a group of subspecialty critical care pulmonologists contracted by a large health maintenance organization to care for all its inpatients. The family physician hospitalists represented rotating family medicine faculty who worked exclusively in the hospital for 8 weeks each year. Some patients under the care of family physician hospitalists originated from the residency’s continuity clinic, and others came from the practices of approximately 30 community family physicians for whom the family physician hospitalist acts as a hospitalist service. Community family medicine primary care physicians cared for their own patients and continued their outpatient practices. Family medicine house staff was involved in the care of inpatients from all groups.

Patients’ insurance plans rather than physician referral or self-referral determined whether they were assigned to the critical care hospitalist, family physician hospitalist, or primary care physician group. Accordingly, there were instances when a primary care physician had some patients admitted to the critical care hospitalist because the patient was a member of this particular health maintenance organization and also because the physician cared directly for his other patients in the hospital.

Outcome measures

We looked at 5 primary patient- and policy-oriented outcomes that have been validated as indicators of quality of inpatient care: in-hospital mortality, length of stay, hospital charges, 7-day readmissions, and 7- and 30-day returns to the emergency department.^18,19 We included Colorado data from the Healthcare Cost and Utilization Project, when available, to serve as a reference standard.²⁰ In addition, we chose 7 validated secondary “process of care”outcomes²¹ to further describe the practice behaviors of the 3 groups: documentation of lifestyle modification counseling (tobacco cessation, exercise, etc), documentation of end-of-life counseling, compliance with contemporary guidelines from the American Thoracic Society for treating community-acquired pneumonia,²¹ length of stay in intensive care, and the use of chest radiographs and blood and sputum cultures.

Statistical methods

All statistical analyses were performed with SAS version 6.12. Patient demographic and clinical characteristics were analyzed with chi-square test and analysis of variance, when appropriate. We controlled for disease severity with the PSI, a well-validated, population-based severity of illness score for inpatients with pneumonia.^22,23

Multiple and logistic regressions were used to control for disease severity and potential confounders. Our models included the PSI class and those characteristics that were statistically significantly different (ie, diagnosis of hypertension) or showed a trend toward difference and were felt to be potentially clinically significant (diagnosis of diabetes, effusion on chest x-ray, mental status at admission), in addition to sex and age. Logarithmic transformations of non-normal data were conducted, when appropriate. We eliminated as extreme outliers 3 cases (2 under the care of critical care hospitalists and 1 under the care of family physician hospitalists) whose outcomes were 2 standard deviations or more beyond the mean. For example, 1 patient had a rare clotting disorder, stayed in intensive care for 30 days, required an orphan drug at $6000 per dose, and had charges well in excess of $1 million. This study was approved by the Rose Medical Center and Health One Institutional Review Board.

Results

Demographic and clinical characteristics

Table 1 provides the demographic and clinical descriptions of patients admitted by the different admitting physician models. Patients in the different groups were similar, with 2 exceptions: hypertension was significantly more common in the critical care hospitalist group than in the other groups (P < .05), and there was a trend toward more diabetes in the critical care hospitalist group that did not quite reach statistical significance. Nonsignificant trends also existed for PSI, effusion on chest x-ray, and mental status, with more effusions and acute mental status changes occurring in the critical care hospitalist group, more chronic altered mental status in the family physician hospitalist group, and greater severity of illness in the critical care hospitalist group. Otherwise, demographics, disease severity, and comorbidities were comparable.

TABLE 1
Demographic and clinical characteristics

Primary outcomes

After controlling for severity of illness and intergroup differences, we found that the critical care hospitalist team had the highest mean hospital charge ($10,231), followed by the family physician hospitalist ($7699) and the primary care physician ($5680) groups (Figure 1). The difference in charges between the primary care physician and the critical care hospitalist groups was statistically significant (P = .005) and approached significance between the primary care physician and family physician hospitalist groups (P = .08). The critical care hospitalist and family physician hospitalist groups had longer mean lengths of stay than did the primary care physician group (P = .04 and .01, respectively; Figure 2). The other primary outcomes were rare: 1 primary care physician patient died (4.5%), 2 critical care hospitalist patients died (9.5%) and no family physician hospitalist patients died; no primary care physician patients were readmitted, 1 critical care hospitalist patient was readmitted (4.8%), and 2 family physician hospitalist patients were readmitted (3.8%). There was 1 return to the emergency room in the cohort, in the family physician hospitalist group (1.9%). No intergroup comparisons between these unadjusted rates were statistically significant (P > .05).

FIGURE 1
Hospital charges (in US dollars; with 95% confidence intervals)*

FIGURE 2
Length of stay (in days; with 95% confidence intervals)*

Secondary outcomes

After controlling for severity of illness and intergroup differences, we found that the critical care hospitalists were more likely to obtain 2 or more chest x-rays than the primary care physicians. There were nonsignificant trends toward longer stays in intensive care, greater likelihood of obtaining sputum cultures, and documenting end-of-life counseling by the critical care hospitalists compared with the primary care physicians. For the family physician hospitalists, there were nonsignificant trends toward better compliance with American Thoracic Society antibiotic guidelines and greater likelihood of documenting end-of-life and lifestyle modification counseling compared with the primary care physicians (Table 2).

TABLE 2
Secondary “process of care” outcomes*

Discussion

Our study provided a unique perspective on the impact of different models of caring for inpatients on the quality, processes, and cost of care. We believe this is the first study to successfully address several methodologic limitations of previous studies: potential time-effect bias, inappropriate controls, differential assignment of house staff and case management resources, unvalidated outcomes, and lack of clinical data.

In addition, we reduced the potential biases inherent in comparing different hospitals and different outpatient physician specialties and used a standardized chart abstraction instrument to avoid the problems inherent in using claims data. As a result, we were able to examine processes of care and use the PSI, an extensively validated tool, to control for disease severity. This study represents an effectiveness study of a real-life intervention by a health maintenance organization to mandate the use of hospitalists. Although the retrospective design of this study may create the potential for bias, there was inadequate advanced notice of the implementation of this hospitalist plan to allow for prospective analysis. Despite not being randomly assigned to 1 of the 3 groups, patients were quite similar with the exceptions of hypertension and possibly diabetes and PSI, which we controlled for in the statistical analysis.

Unfortunately, this was a small study that lacked sufficient power to detect modest differences between groups because the health maintenance organization sponsoring the critical care hospitalist group abandoned the program after 1 year. In addition, the differences in disease severity might have been significant in a larger sample. However, even after controlling for these differences statistically, we found no large differences for mortality, readmission, or returns to the emergency room.

Despite insufficient power to observe statistically significant differences in these relatively rare but important pneumonia outcomes, we did detect a substantial difference in adjusted hospital charges and a modest difference in length of stay. Subspecialist hospitalists had significantly higher adjusted charges than did primary care family physicians. Although the comparison across groups failed to show statistically significant differences, we did see a trend of increasing charges as the degree of hospitalization increased. These higher costs may be explained in part by primary care physicians advising shorter lengths of stay and the subspecialists’ increased use of multiple chest x-rays and trends toward greater use of other resources (eg, intensive care and blood and sputum cultures). Alternatively, some of the difference in charges may reflect differing levels of continuity; the critical care hospitalists had no outpatient continuity with their inpatients, whereas the family physician hospitalists had continuity relationships with some inpatients and the primary care physicians had relationships with all their inpatients. Thus the primary care physicians and, to a lesser extent, the family physician hospitalists may have had information about prior care. Hence, knowledge of previous antibiotic use might argue for the low yield of blood and sputum cultures, and having obtained an outpatient x-ray might obviate the need for another in the hospital. The critical care hospitalists’ increased length of stay and x-ray use, in conjunction with the trend toward greater use of cultures and intensive care, may in turn reflect different degrees of comfort with uncertainty between family physicians and subspecialists. Also, we examined only hospital charges rather than total costs to the system.

Interestingly, we found a trend showing that family physician hospitalists were more likely to document lifestyle modification counseling than were primary care physicians. This result should be interpreted with some caution. Our findings may indicate a true lack of performance by primary care physicians, or they may show a failure to document advice on the hospital chart, reflecting some aspect of the continuity relationship in which such discussions are relegated to the outpatient setting.We also were surprised to see the trend toward decreased end-of-life counseling by the primary care physicians. This could reflect some adverse effect of continuity, the time constraints imposed on nonhospitalists, or not documenting outpatient counseling on the inpatient record.

There were other potential sources of confounding in this study. All patients in the critical care hospitalist group were members of the same health maintenance organization, which may have introduced unmeasured bias despite our attempts to control for differences between groups. Even though we purposefully avoided differential use of house staff, its involvement in each case may have decreased any potential differences across practices.

We draw 2 important conclusions from our results. First, our findings of increased costs and length of stay for mandated hospitalists without significantly different outcomes support the assertion of the American Academy of Family Physicians, the American College of Physicians–American Society of Internal Medicine, the American Medical Association, and the National Association of Inpatient Physicians: the practice of mandating the use of hospitalists should be abandoned pending larger, more comprehensive contemporaneous trials. Second, if hospitalists are to be employed on a voluntary basis, the use of subspecialists rather than generalists may result in more costly care.

Acknowledgments

The authors thank Ralph B. D’Agonstino, Jr., PhD, of the Bowman Gray School of Medicine for his invaluable assistance in statistical methods and Cary Foster, MD, and Promoda Mahupatra, MD, for their collection of the data.

EVIDENCE-BASED ANSWER

Efforts to increase physical activity or decrease sedentary activities have shown some short-term benefit, and adding dietary changes may be more effective. Aiming interventions at parents, intensive family therapy, comprehensive school-based programs, and selecting motivated children for subspecialty care may improve success. (Grade of recommendation: B, based on poor-quality randomized controlled trials [RCTs] and heterogeneous systematic reviews.) Other potentially effective short-term strategies include screening with body mass index (BMI) for age (grade of recommendation: C, extrapolation from cohort studies and ecological research) or dietary counseling (grade of recommendation: D, conflicting poor-quality RCTs). No drugs are currently approved for pediatric obesity therapy in the United States.

Evidence summary

Pediatric obesity increases the risk of adverse outcomes in adulthood, independent of adult BMI.¹ Trials aiming to reduce childhood obesity suffer from serious methodological constraints. No long-term (> 2 years) evidence is available. Many apparently efficacious interventions are beyond the scope of primary care physicians. A detailed summary is available online at http://www.FPIN.org.

Several studies have examined the value of isolated changes in either diet or activity level. Randomized controlled trials and retrospective cohort studies of dietary advice alone show short-term efficacy (weeks to months).¹ Most involve intensive subspecialty care for extremely obese children who are 120% to 140% over their ideal body weight. Trials without careful selection of motivated children had dropout rates up to 87%. One Italian RCT showed a 12% reduction in the number of obese children in schools receiving multimedia dietary advice compared with a 5% to 6% increase in those schools receiving only written or no advice.² Several RCTs reduced obesity by introducing or improving school-based physical activity.¹ Two RCTs that discouraged sedentary activity through counseling or school-based programs also reduced obesity.^3,4

Two larger trials integrated diet and exercise advice into school curricula. One study emphasized improving school menus, but no difference was observed because children compensated by overeating at home.⁵ The other emphasized reducing sedentary activities and found lower obesity rates in the intervention schools, but only for girls.⁶ A third trial provided family-based dietary and behavior counseling, but emphasized either increasing physical activity or decreasing sedentary activity.⁷ Both strategies resulted in reduced obesity compared with controls. A systematic review supported the combined approach of these trials, finding diet and exercise interventions superior to diet interventions alone.⁸

Some evidence supports focusing on the family rather than just the child. A systematic review found that family therapy prevented pediatric obesity.⁹ An RCT found that focusing on parents as the sole change agent was superior to targeting the child.¹⁰

Recommendations from others

Expert consensus promotes screening because of obesity’s increasing incidence and associated morbidity and mortality.¹¹ The Maternal and Child Health Bureau recommends a primary goal of healthy eating and activity. They recommend treating when the body mass index is >95th percentile, and assessing the child and family’s willingness to change. Primary strategies are to begin early, involve the family, promote parenting skills, and increase activity and reduce high-calorie food intake. They also recommend ongoing support to maintain weight loss.⁹

Clinical Commentary by Tess Bobo, MD, at http://www.fpin.org.

EVIDENCE-BASED ANSWER

Efforts to increase physical activity or decrease sedentary activities have shown some short-term benefit, and adding dietary changes may be more effective. Aiming interventions at parents, intensive family therapy, comprehensive school-based programs, and selecting motivated children for subspecialty care may improve success. (Grade of recommendation: B, based on poor-quality randomized controlled trials [RCTs] and heterogeneous systematic reviews.) Other potentially effective short-term strategies include screening with body mass index (BMI) for age (grade of recommendation: C, extrapolation from cohort studies and ecological research) or dietary counseling (grade of recommendation: D, conflicting poor-quality RCTs). No drugs are currently approved for pediatric obesity therapy in the United States.

Evidence summary

Pediatric obesity increases the risk of adverse outcomes in adulthood, independent of adult BMI.¹ Trials aiming to reduce childhood obesity suffer from serious methodological constraints. No long-term (> 2 years) evidence is available. Many apparently efficacious interventions are beyond the scope of primary care physicians. A detailed summary is available online at http://www.FPIN.org.

Several studies have examined the value of isolated changes in either diet or activity level. Randomized controlled trials and retrospective cohort studies of dietary advice alone show short-term efficacy (weeks to months).¹ Most involve intensive subspecialty care for extremely obese children who are 120% to 140% over their ideal body weight. Trials without careful selection of motivated children had dropout rates up to 87%. One Italian RCT showed a 12% reduction in the number of obese children in schools receiving multimedia dietary advice compared with a 5% to 6% increase in those schools receiving only written or no advice.² Several RCTs reduced obesity by introducing or improving school-based physical activity.¹ Two RCTs that discouraged sedentary activity through counseling or school-based programs also reduced obesity.^3,4

Two larger trials integrated diet and exercise advice into school curricula. One study emphasized improving school menus, but no difference was observed because children compensated by overeating at home.⁵ The other emphasized reducing sedentary activities and found lower obesity rates in the intervention schools, but only for girls.⁶ A third trial provided family-based dietary and behavior counseling, but emphasized either increasing physical activity or decreasing sedentary activity.⁷ Both strategies resulted in reduced obesity compared with controls. A systematic review supported the combined approach of these trials, finding diet and exercise interventions superior to diet interventions alone.⁸

Some evidence supports focusing on the family rather than just the child. A systematic review found that family therapy prevented pediatric obesity.⁹ An RCT found that focusing on parents as the sole change agent was superior to targeting the child.¹⁰

Recommendations from others

Expert consensus promotes screening because of obesity’s increasing incidence and associated morbidity and mortality.¹¹ The Maternal and Child Health Bureau recommends a primary goal of healthy eating and activity. They recommend treating when the body mass index is >95th percentile, and assessing the child and family’s willingness to change. Primary strategies are to begin early, involve the family, promote parenting skills, and increase activity and reduce high-calorie food intake. They also recommend ongoing support to maintain weight loss.⁹

Clinical Commentary by Tess Bobo, MD, at http://www.fpin.org.

EVIDENCE-BASED ANSWER

Efforts to increase physical activity or decrease sedentary activities have shown some short-term benefit, and adding dietary changes may be more effective. Aiming interventions at parents, intensive family therapy, comprehensive school-based programs, and selecting motivated children for subspecialty care may improve success. (Grade of recommendation: B, based on poor-quality randomized controlled trials [RCTs] and heterogeneous systematic reviews.) Other potentially effective short-term strategies include screening with body mass index (BMI) for age (grade of recommendation: C, extrapolation from cohort studies and ecological research) or dietary counseling (grade of recommendation: D, conflicting poor-quality RCTs). No drugs are currently approved for pediatric obesity therapy in the United States.

Evidence summary

Pediatric obesity increases the risk of adverse outcomes in adulthood, independent of adult BMI.¹ Trials aiming to reduce childhood obesity suffer from serious methodological constraints. No long-term (> 2 years) evidence is available. Many apparently efficacious interventions are beyond the scope of primary care physicians. A detailed summary is available online at http://www.FPIN.org.

Several studies have examined the value of isolated changes in either diet or activity level. Randomized controlled trials and retrospective cohort studies of dietary advice alone show short-term efficacy (weeks to months).¹ Most involve intensive subspecialty care for extremely obese children who are 120% to 140% over their ideal body weight. Trials without careful selection of motivated children had dropout rates up to 87%. One Italian RCT showed a 12% reduction in the number of obese children in schools receiving multimedia dietary advice compared with a 5% to 6% increase in those schools receiving only written or no advice.² Several RCTs reduced obesity by introducing or improving school-based physical activity.¹ Two RCTs that discouraged sedentary activity through counseling or school-based programs also reduced obesity.^3,4

Two larger trials integrated diet and exercise advice into school curricula. One study emphasized improving school menus, but no difference was observed because children compensated by overeating at home.⁵ The other emphasized reducing sedentary activities and found lower obesity rates in the intervention schools, but only for girls.⁶ A third trial provided family-based dietary and behavior counseling, but emphasized either increasing physical activity or decreasing sedentary activity.⁷ Both strategies resulted in reduced obesity compared with controls. A systematic review supported the combined approach of these trials, finding diet and exercise interventions superior to diet interventions alone.⁸

Some evidence supports focusing on the family rather than just the child. A systematic review found that family therapy prevented pediatric obesity.⁹ An RCT found that focusing on parents as the sole change agent was superior to targeting the child.¹⁰

Recommendations from others

Expert consensus promotes screening because of obesity’s increasing incidence and associated morbidity and mortality.¹¹ The Maternal and Child Health Bureau recommends a primary goal of healthy eating and activity. They recommend treating when the body mass index is >95th percentile, and assessing the child and family’s willingness to change. Primary strategies are to begin early, involve the family, promote parenting skills, and increase activity and reduce high-calorie food intake. They also recommend ongoing support to maintain weight loss.⁹

Clinical Commentary by Tess Bobo, MD, at http://www.fpin.org.

Recommendations from others

Guidelines from other groups regarding mammographic screening of older women are not consistent Table. Neither the American Academy of Family Physicians and the American College of Radiology make specific recommendations.

Evidence-based answer

Screening mammography for women aged up to 70 years is widely recommended. However, its role in older women is unclear, even the though average life expectancy of elderly women is lengthy (eg, 9 years at age 80). Patients with 3 or more major comorbidities (ie, hypertension, diabetes, arthritis, myocardial infarction, stroke, respiratory disease, or other cancers) are 20 times more likely to die of a cause other than breast cancer within 3 years.⁵ Although half of all cancers are diagnosed in women older than 65 years, screening these women poses some risks. For every 1000 women older than 70 years who are screened, 81 mammograms will show an abnormal result, requiring 79 follow-up tests and 26 biopsies to detect 11 cancers.⁶ The picture is further complicated by the possibility that these cancers may be less aggressive. Cancers may grow more slowly with age,⁷ and 25% of all tumors in elderly women are ductal carcinoma in situ, which are unlikely to cause morbidity or death within their life spans.⁸

Unfortunately, the randomized trials of screening mammography have enrolled few women aged 70 to 74 years and none older than 74. Two studies retrospectively compared cohorts of elderly women who did and did not have screening. Smith-Bindman and colleagues⁹ studied 690,000 Medicare beneficiaries in California and found a 43% reduction in the risk of metastatic breast cancer in the screened group. A 13-year Dutch study found a 20% breast cancer mortality reduction that first appeared 7 years after initiating screening. The cohorts were not randomly assigned in either study, leading to possible differences in baseline risk.

In a cost-effectiveness analysis, Kerlikowske and coworkers¹⁰ included BMD in the model, which is positively associated with breast cancer risk. They found a small benefit to screening for women with the highest quartile BMD (highest cancer risk) and virtually no benefit in screening the lowest quartile BMD. Mandeblatt and colleagues ¹¹ incorporated comorbidities in another model and found a small benefit for women older than 69 years that gradually diminished as they approached the age of 85. This benefit was halved in the presence of a major comorbidity. Both of these models assumed that the efficacy of screening in younger women applies to older women.

Since there are no randomized trials in this age group, the true efficacy of screening is unknown. Therefore, other factors must be considered, including the patient’s age, comorbidities, and preferences. The evidence suggests that the best candidates to stop screening mammography are elderly women who have significant comorbidities, poor functional status, low BMD, little interest in preventive care, or an unwillingness to accept the potential harm of screening.

Recommendations from others

Guidelines from other groups regarding mammographic screening of older women are not consistent Table. Neither the American Academy of Family Physicians and the American College of Radiology make specific recommendations.

Evidence-based answer

Screening mammography for women aged up to 70 years is widely recommended. However, its role in older women is unclear, even the though average life expectancy of elderly women is lengthy (eg, 9 years at age 80). Patients with 3 or more major comorbidities (ie, hypertension, diabetes, arthritis, myocardial infarction, stroke, respiratory disease, or other cancers) are 20 times more likely to die of a cause other than breast cancer within 3 years.⁵ Although half of all cancers are diagnosed in women older than 65 years, screening these women poses some risks. For every 1000 women older than 70 years who are screened, 81 mammograms will show an abnormal result, requiring 79 follow-up tests and 26 biopsies to detect 11 cancers.⁶ The picture is further complicated by the possibility that these cancers may be less aggressive. Cancers may grow more slowly with age,⁷ and 25% of all tumors in elderly women are ductal carcinoma in situ, which are unlikely to cause morbidity or death within their life spans.⁸

Unfortunately, the randomized trials of screening mammography have enrolled few women aged 70 to 74 years and none older than 74. Two studies retrospectively compared cohorts of elderly women who did and did not have screening. Smith-Bindman and colleagues⁹ studied 690,000 Medicare beneficiaries in California and found a 43% reduction in the risk of metastatic breast cancer in the screened group. A 13-year Dutch study found a 20% breast cancer mortality reduction that first appeared 7 years after initiating screening. The cohorts were not randomly assigned in either study, leading to possible differences in baseline risk.

In a cost-effectiveness analysis, Kerlikowske and coworkers¹⁰ included BMD in the model, which is positively associated with breast cancer risk. They found a small benefit to screening for women with the highest quartile BMD (highest cancer risk) and virtually no benefit in screening the lowest quartile BMD. Mandeblatt and colleagues ¹¹ incorporated comorbidities in another model and found a small benefit for women older than 69 years that gradually diminished as they approached the age of 85. This benefit was halved in the presence of a major comorbidity. Both of these models assumed that the efficacy of screening in younger women applies to older women.

Since there are no randomized trials in this age group, the true efficacy of screening is unknown. Therefore, other factors must be considered, including the patient’s age, comorbidities, and preferences. The evidence suggests that the best candidates to stop screening mammography are elderly women who have significant comorbidities, poor functional status, low BMD, little interest in preventive care, or an unwillingness to accept the potential harm of screening.

Recommendations from others

Guidelines from other groups regarding mammographic screening of older women are not consistent Table. Neither the American Academy of Family Physicians and the American College of Radiology make specific recommendations.

Evidence-based answer

Screening mammography for women aged up to 70 years is widely recommended. However, its role in older women is unclear, even the though average life expectancy of elderly women is lengthy (eg, 9 years at age 80). Patients with 3 or more major comorbidities (ie, hypertension, diabetes, arthritis, myocardial infarction, stroke, respiratory disease, or other cancers) are 20 times more likely to die of a cause other than breast cancer within 3 years.⁵ Although half of all cancers are diagnosed in women older than 65 years, screening these women poses some risks. For every 1000 women older than 70 years who are screened, 81 mammograms will show an abnormal result, requiring 79 follow-up tests and 26 biopsies to detect 11 cancers.⁶ The picture is further complicated by the possibility that these cancers may be less aggressive. Cancers may grow more slowly with age,⁷ and 25% of all tumors in elderly women are ductal carcinoma in situ, which are unlikely to cause morbidity or death within their life spans.⁸

Unfortunately, the randomized trials of screening mammography have enrolled few women aged 70 to 74 years and none older than 74. Two studies retrospectively compared cohorts of elderly women who did and did not have screening. Smith-Bindman and colleagues⁹ studied 690,000 Medicare beneficiaries in California and found a 43% reduction in the risk of metastatic breast cancer in the screened group. A 13-year Dutch study found a 20% breast cancer mortality reduction that first appeared 7 years after initiating screening. The cohorts were not randomly assigned in either study, leading to possible differences in baseline risk.

In a cost-effectiveness analysis, Kerlikowske and coworkers¹⁰ included BMD in the model, which is positively associated with breast cancer risk. They found a small benefit to screening for women with the highest quartile BMD (highest cancer risk) and virtually no benefit in screening the lowest quartile BMD. Mandeblatt and colleagues ¹¹ incorporated comorbidities in another model and found a small benefit for women older than 69 years that gradually diminished as they approached the age of 85. This benefit was halved in the presence of a major comorbidity. Both of these models assumed that the efficacy of screening in younger women applies to older women.

Since there are no randomized trials in this age group, the true efficacy of screening is unknown. Therefore, other factors must be considered, including the patient’s age, comorbidities, and preferences. The evidence suggests that the best candidates to stop screening mammography are elderly women who have significant comorbidities, poor functional status, low BMD, little interest in preventive care, or an unwillingness to accept the potential harm of screening.

Recommendations from others

The Cochrane Database of Systematic Reviews reports, “In light of the current lack of good trial-based evidence, clinicians should consider restricting surgery for obstructive sleep apnoea to that carried out as part of clinical trials. Where practice is continued, patients should be informed of the experimental nature of the operations. Patients should be told that there is a lack of good trial-based evidence of the efficacy of surgery for obstructive sleep apnoea, a course of action that may restrict the use of these operations.”¹

The American Sleep Disorders Association states, “The presence and severity of obstructive sleep apnea must be determined before initiating surgical therapy… because of the complexity of airway narrowing or collapse during sleep, any one surgical procedure may not eradicate a patient’s sleep apnea. A stepwise approach to surgical management is acceptable if the patient is advised at the onset of treatment about the likelihood of the success of each procedure and that multiple operations may be necessary. After the surgical site has healed, a follow-up evaluation, including an objective measure of respiration and quality of sleep, must be performed to ensure that the abnormalities noted in the original study are corrected.”²

Search strategy

The Cochrane Database of Systematic Reviews, the National Guideline Clearinghouse Database, the Turning Research Into Practice (TRIP) database, British National Health Service Health Technology Assessment Database (HTA), and DARE (Database of Abstracts and Reviews of Effectiveness) were searched, using the keywords “sleep” and “sleep apnea.”

MEDLINE search strategy: {(Sleep Apnea, Obstructive/su [Surgery], OR sleep apnea syndromes/su) AND (clinical trial.pt. OR exp treatment outcome/OR meta-analysis.pt)}

Evidence summary

Obstructive sleep apnea is a common and potentially serious condition. Good evidence now exists supporting the efficacy of CPAP in improving patient-oriented outcomes (eg, daytime sleepiness) and disease-oriented outcomes (eg, nocturnal hypoxia).^3,4,5 However, CPAP is not tolerated by 20% to 30% of patients, not universally effective, and ineffective if not used regularly.⁵ Surgery has held promise as a potentially curative therapy, a promise not yet demonstrated in well-designed randomized controlled trials.

The Cochrane Airways Group performed an extensive review of this topic in October 1999 and found that none of the 594 potentially relevant articles reviewed were methodologically adequate.¹ Our search found 2 randomized controlled trials: 1 apparently missed by the Cochrane review and 1 subsequent to it.

Lojander et al⁶ compared uvulopharyngopalatoplasty (UPPP) to conservative management (weight loss, smoking cessation, and alcohol avoidance advice) in 32 patients. Patients were excluded if they had asthma, obesity (body mass index >40), restless leg syndrome, hypothyroidism, chronic obstructive pulmonary disease, or intolerance of missed work because of somnolence. Remaining patients were selected as possible surgical candidates by a team of hospital-based experts then randomized to surgery or conservative therapy. At 1 year of follow-up, they found no change in nocturnal oxygen desaturation but some improvement in the mean score of a visual analog scale of daytime sleepiness and in the mean self-report of out-of-bed sleeping episodes. Unfortunately these subjective outcome measures have not been validated and are subject to bias. This, in addition to lack of intention-to-treat analysis, renders these results uninterpretable.

The other trial compared UPPP with dental appliances in late 1999 and found dental appliances superior to surgery for improving apnea scores.⁷

Current recommendations in favor of surgical therapy are based solely on case series, poorly designed trials, expert opinion or pathophysiologic reasoning.¹ Although such literature is voluminous, the most optimistic review estimates that less than 50% of patients with obstructive sleep apnea will benefit from surgery.⁸ The nonrandomized design of such trials may overestimate treatment effects by more than 40%.⁹

Given how little is known about its effectiveness, compounded by potential iatrogenesis, surgery should be considered the last resort for patients with obstructive sleep apnea. Doctors should discuss CPAP (the most effective therapy) and condition-specific dental appliances (the best tolerated therapy) with their patients until more is known about the efficacy of surgery. All patients should be counseled regarding healthy lifestyle modification.

Recommendations from others

The Cochrane Database of Systematic Reviews reports, “In light of the current lack of good trial-based evidence, clinicians should consider restricting surgery for obstructive sleep apnoea to that carried out as part of clinical trials. Where practice is continued, patients should be informed of the experimental nature of the operations. Patients should be told that there is a lack of good trial-based evidence of the efficacy of surgery for obstructive sleep apnoea, a course of action that may restrict the use of these operations.”¹

The American Sleep Disorders Association states, “The presence and severity of obstructive sleep apnea must be determined before initiating surgical therapy… because of the complexity of airway narrowing or collapse during sleep, any one surgical procedure may not eradicate a patient’s sleep apnea. A stepwise approach to surgical management is acceptable if the patient is advised at the onset of treatment about the likelihood of the success of each procedure and that multiple operations may be necessary. After the surgical site has healed, a follow-up evaluation, including an objective measure of respiration and quality of sleep, must be performed to ensure that the abnormalities noted in the original study are corrected.”²

Search strategy

The Cochrane Database of Systematic Reviews, the National Guideline Clearinghouse Database, the Turning Research Into Practice (TRIP) database, British National Health Service Health Technology Assessment Database (HTA), and DARE (Database of Abstracts and Reviews of Effectiveness) were searched, using the keywords “sleep” and “sleep apnea.”

MEDLINE search strategy: {(Sleep Apnea, Obstructive/su [Surgery], OR sleep apnea syndromes/su) AND (clinical trial.pt. OR exp treatment outcome/OR meta-analysis.pt)}

Evidence summary

Obstructive sleep apnea is a common and potentially serious condition. Good evidence now exists supporting the efficacy of CPAP in improving patient-oriented outcomes (eg, daytime sleepiness) and disease-oriented outcomes (eg, nocturnal hypoxia).^3,4,5 However, CPAP is not tolerated by 20% to 30% of patients, not universally effective, and ineffective if not used regularly.⁵ Surgery has held promise as a potentially curative therapy, a promise not yet demonstrated in well-designed randomized controlled trials.

The Cochrane Airways Group performed an extensive review of this topic in October 1999 and found that none of the 594 potentially relevant articles reviewed were methodologically adequate.¹ Our search found 2 randomized controlled trials: 1 apparently missed by the Cochrane review and 1 subsequent to it.

Lojander et al⁶ compared uvulopharyngopalatoplasty (UPPP) to conservative management (weight loss, smoking cessation, and alcohol avoidance advice) in 32 patients. Patients were excluded if they had asthma, obesity (body mass index >40), restless leg syndrome, hypothyroidism, chronic obstructive pulmonary disease, or intolerance of missed work because of somnolence. Remaining patients were selected as possible surgical candidates by a team of hospital-based experts then randomized to surgery or conservative therapy. At 1 year of follow-up, they found no change in nocturnal oxygen desaturation but some improvement in the mean score of a visual analog scale of daytime sleepiness and in the mean self-report of out-of-bed sleeping episodes. Unfortunately these subjective outcome measures have not been validated and are subject to bias. This, in addition to lack of intention-to-treat analysis, renders these results uninterpretable.

The other trial compared UPPP with dental appliances in late 1999 and found dental appliances superior to surgery for improving apnea scores.⁷

Current recommendations in favor of surgical therapy are based solely on case series, poorly designed trials, expert opinion or pathophysiologic reasoning.¹ Although such literature is voluminous, the most optimistic review estimates that less than 50% of patients with obstructive sleep apnea will benefit from surgery.⁸ The nonrandomized design of such trials may overestimate treatment effects by more than 40%.⁹

Given how little is known about its effectiveness, compounded by potential iatrogenesis, surgery should be considered the last resort for patients with obstructive sleep apnea. Doctors should discuss CPAP (the most effective therapy) and condition-specific dental appliances (the best tolerated therapy) with their patients until more is known about the efficacy of surgery. All patients should be counseled regarding healthy lifestyle modification.

Recommendations from others

The Cochrane Database of Systematic Reviews reports, “In light of the current lack of good trial-based evidence, clinicians should consider restricting surgery for obstructive sleep apnoea to that carried out as part of clinical trials. Where practice is continued, patients should be informed of the experimental nature of the operations. Patients should be told that there is a lack of good trial-based evidence of the efficacy of surgery for obstructive sleep apnoea, a course of action that may restrict the use of these operations.”¹

The American Sleep Disorders Association states, “The presence and severity of obstructive sleep apnea must be determined before initiating surgical therapy… because of the complexity of airway narrowing or collapse during sleep, any one surgical procedure may not eradicate a patient’s sleep apnea. A stepwise approach to surgical management is acceptable if the patient is advised at the onset of treatment about the likelihood of the success of each procedure and that multiple operations may be necessary. After the surgical site has healed, a follow-up evaluation, including an objective measure of respiration and quality of sleep, must be performed to ensure that the abnormalities noted in the original study are corrected.”²

Search strategy

The Cochrane Database of Systematic Reviews, the National Guideline Clearinghouse Database, the Turning Research Into Practice (TRIP) database, British National Health Service Health Technology Assessment Database (HTA), and DARE (Database of Abstracts and Reviews of Effectiveness) were searched, using the keywords “sleep” and “sleep apnea.”

MEDLINE search strategy: {(Sleep Apnea, Obstructive/su [Surgery], OR sleep apnea syndromes/su) AND (clinical trial.pt. OR exp treatment outcome/OR meta-analysis.pt)}

Evidence summary

Obstructive sleep apnea is a common and potentially serious condition. Good evidence now exists supporting the efficacy of CPAP in improving patient-oriented outcomes (eg, daytime sleepiness) and disease-oriented outcomes (eg, nocturnal hypoxia).^3,4,5 However, CPAP is not tolerated by 20% to 30% of patients, not universally effective, and ineffective if not used regularly.⁵ Surgery has held promise as a potentially curative therapy, a promise not yet demonstrated in well-designed randomized controlled trials.

The Cochrane Airways Group performed an extensive review of this topic in October 1999 and found that none of the 594 potentially relevant articles reviewed were methodologically adequate.¹ Our search found 2 randomized controlled trials: 1 apparently missed by the Cochrane review and 1 subsequent to it.

Lojander et al⁶ compared uvulopharyngopalatoplasty (UPPP) to conservative management (weight loss, smoking cessation, and alcohol avoidance advice) in 32 patients. Patients were excluded if they had asthma, obesity (body mass index >40), restless leg syndrome, hypothyroidism, chronic obstructive pulmonary disease, or intolerance of missed work because of somnolence. Remaining patients were selected as possible surgical candidates by a team of hospital-based experts then randomized to surgery or conservative therapy. At 1 year of follow-up, they found no change in nocturnal oxygen desaturation but some improvement in the mean score of a visual analog scale of daytime sleepiness and in the mean self-report of out-of-bed sleeping episodes. Unfortunately these subjective outcome measures have not been validated and are subject to bias. This, in addition to lack of intention-to-treat analysis, renders these results uninterpretable.

The other trial compared UPPP with dental appliances in late 1999 and found dental appliances superior to surgery for improving apnea scores.⁷

Current recommendations in favor of surgical therapy are based solely on case series, poorly designed trials, expert opinion or pathophysiologic reasoning.¹ Although such literature is voluminous, the most optimistic review estimates that less than 50% of patients with obstructive sleep apnea will benefit from surgery.⁸ The nonrandomized design of such trials may overestimate treatment effects by more than 40%.⁹

Given how little is known about its effectiveness, compounded by potential iatrogenesis, surgery should be considered the last resort for patients with obstructive sleep apnea. Doctors should discuss CPAP (the most effective therapy) and condition-specific dental appliances (the best tolerated therapy) with their patients until more is known about the efficacy of surgery. All patients should be counseled regarding healthy lifestyle modification.