To determine whether published cost-effectiveness and cost-benefit analyses have adhered to basic analytic principles.

The extent to which a clinician should believe and act on the results of subgroup analyses of data from randomized trials or meta-analyses is controversial. Guidelines are provided in this paper for making these decisions. The strength of inference regarding a proposed difference in treatment effect among subgroups is dependent on the magnitude of the difference, the statistical significance of the difference, whether the hypothesis preceded or followed the analysis, whether the subgroup analysis was one of a small number of hypotheses tested, whether the difference was suggested by comparisons within or between studies, the consistency of the difference, and the existence of indirect evidence that supports the difference. Application of these guidelines will assist clinicians in making decisions regarding whether to base a treatment decision on overall results or on the results of a subgroup analysis.

Clinical trials of cholesterol lowering in apparently healthy persons (primary prevention) and in persons with known coronary disease (secondary prevention) routinely show a decrease in the incidence of nonfatal and fatal coronary events. Primary prevention trials, however, generally have failed to show a beneficial effect of cholesterol lowering on total mortality, because of both low overall death rates and a disturbingly high number of deaths from noncoronary causes in the treatment groups. Studies of secondary prevention give more clear-cut evidence of benefit. The rate of death from coronary heart disease is extremely high in these studies, overwhelming the rates for other causes of death. Thus, a possible increase in noncoronary deaths is of much less concern in secondary prevention studies. A possible link between cholesterol lowering and noncoronary causes of death should be explored using dose-response analyses of existing data sets, in both individual studies and aggregate meta-analysis; such analyses should be similar to those that have been done to study the relation between cholesterol lowering and coronary events. These analyses would greatly assist us in developing safe, efficacious ways to prevent coronary heart disease.

Seven randomized studies during the past 5 years have evaluated or are evaluating the efficacy of warfarin or aspirin or both in decreasing the risk of embolic events in patients with nonrheumatic atrial fibrillation. By March 1990, two of the studies had been published, both of which showed a statistically significant decrease in embolic events in patients treated with warfarin and a low rate of major bleeding events. The investigators associated with the other ongoing studies were forced to consider how these results should affect the decision to recruit and continue follow-up of patients in their own studies. The Steering Committee of the Canadian Atrial Fibrillation Anticoagulation (CAFA) study thought the newly published results from other studies were valid, clinically important, and generalizable. The committee considered the following options for the CAFA study: continue patient recruitment as planned, provide the data available in CAFA to its External Safety and Efficacy Monitoring Committee for analysis to determine whether the CAFA data already showed a benefit of warfarin, stop patient recruitment but continue to follow patients in the group to which they were assigned, stop the trial immediately and perform a final analysis, and attempt to perform a meta-analysis of all data available from all trials. The Steering Committee of CAFA decided that the evidence of benefit with warfarin, from the two published studies, was sufficiently compelling as to stop recruitment into CAFA without any preliminary examination of the CAFA data.

To describe the relative risk for serious gastrointestinal complications due to non-aspirin nonsteroidal anti-inflammatory drug (NSAID) exposure among NSAID users as well as in selected subgroups.

There has been considerable uncertainty about the best way to prevent stroke in patients with nonvalvular atrial fibrillation. Recent studies have suggested that low-dose warfarin therapy, in addition to producing fewer bleeding complications, may be as effective as higher-dose therapy in preventing thromboembolic events. Four large, prospective, randomized trials have examined the risks and benefits of warfarin therapy for stroke prophylaxis in patients with nonvalvular atrial fibrillation. All four studies showed a substantially reduced incidence of stroke and a low incidence of significant bleeding in patients treated with warfarin. One of these studies also showed that aspirin reduced the incidence of stroke. The benefits associated with long-term low-dose warfarin therapy appear to exceed the risks for serious bleeding in most patients with atrial fibrillation. Aspirin may be a viable therapeutic option for patients who are unable to take warfarin or for those in subgroups at a low risk for stroke.

To assess the effects of anabolic-androgenic steroids on human muscle strength.

To review the recent literature on the efficacy of combined insulin and sulfonylurea therapy in patients with type 2 diabetes.

To provide a clinical approach to the diagnosis of pulmonary embolism.

To evaluate the effectiveness of oral corticosteroid therapy in patients with stable chronic obstructive pulmonary disease.

To assess the cost effectiveness of the current recommendation that persons who have had an adenomatous colon polyp removed have periodic colonoscopic surveillance at fixed and regular intervals.

To determine the effect of intravenous streptokinase on early mortality in patients with suspected acute anterior and acute inferior myocardial infarctions.

No prospective, randomized clinical trial comparing coronary artery bypass grafting, percutaneous transluminal coronary angioplasty, and conservative therapy has been reported. To address when revascularization is indicated, we constructed a decision analytic model. Our model incorporates procedure-related mortality and morbidity, coronary artery disease-related mortality, and the benefit of revascularization. We determined the quality-adjusted life expectancy and expected costs for each strategy. Our model suggests that angioplasty is a reasonable alternative to bypass surgery in patients with favorable lesions if angioplasty would provide a comparable degree of revascularization. Our model predicts that patients treated with angioplasty will have more revascularization procedures than will patients treated with bypass surgery but predicts that both treatments will cost the same over the typical patient's lifetime. In many patients with severe angina or documented ischemia, angioplasty is indicated for stenosis of a single artery. In patients with two vessel disease that is amenable to angioplasty, angioplasty may be a reasonable alternative to bypass surgery. Even in patients whose three vessel disease can be completely revascularized by angioplasty, bypass surgery, although relatively expensive, is slightly better than angioplasty. In patients with three vessel disease and comorbidities that increase operative risk, angioplasty may be a reasonable alternative to either bypass surgery or medical therapy.

To evaluate the use of thyroid function tests to monitor therapy in patients taking levothyroxine.

To determine whether corticosteroids affect short-term mortality from alcoholic hepatitis.

To review the evidence for efficacy of three contemporary treatments for intermittent claudication: pentoxifylline, exercise programs, and smoking cessation.

To critically assess original studies evaluating the role of preoperative pulmonary function testing in predicting postoperative outcomes.

To review the evidence on four questions about screening asymptomatic adults for arterial hypertension: Is hypertension a significant health problem? Is it detectable at an early, presymptomatic stage? Is treatment available and effective? Do the benefits of screening outweigh the costs and risks?