Search

Service on the Data Monitoring Committee of the CPEP (Calcium for Pre-eclampsia Prevention) has led us to four conclusions about clinical trials which we should like to present to this gathering of biostatisticians for their reactions: (i) meta-analyses of the pertinent published trials of the same therapy should always be undertaken before the start of a new trial, and the results examined to help determine the design of a new trial or determine if a trial should be undertaken at all; (ii) assuming that a decision is made to go ahead, the results of the past trials should be used in sizing the new one; (iii) in the course of the new one, regardless of the size estimates, stopping early should be considered if the trends conform to the results of the meta-analysis; and (iv) heterogeneity of patients entering clinical trials is desirable and should be specifically studied, and it should never be concluded that an average outcome is applicable to all future patients.

The findings in clinical trials of antiarrhythmic drug efficacy and safety are frequently difficult to compare, since study design often has an important effect on trial outcome. To explore this problem further, we compared three designs--randomized control, nonrandomized control, and uncontrolled--collectively enrolling 2415 patients in 21 trials reporting on the role of quinidine in the prevention of chronic atrial fibrillation. The proportion of patients remaining in sinus rhythm at 3, 6, and 12 months after cardioversion was calculated by means of Kaplan-Meier techniques, and the data were pooled for each trial design. For the randomized control trials the difference in the absolute percentage of patients remaining in sinus rhythm in the quinidine and control groups was 24% at each of the three follow-up intervals. Contrary to findings in the randomized control trials, the magnitude of the treatment benefit in nonrandomized trials was smaller and declined markedly over time. The percentage of patients remaining in sinus rhythm in the uncontrolled trials was intermediate to the percentages in the other two trial designs. When the data from all three trial designs were pooled, the crude mortality rate was 2.0% in quinidine-treated patients and 0.6% in control patients. Sudden cardiac death or ventricular fibrillation was the cause of death in 13 of 19 patients for whom the cause of death was known, highlighting the potential risk of quinidine-induced proarrhythmia. Although quinidine is effective in maintaining sinus rhythm, estimates of the treatment effect vary among trial types.(ABSTRACT TRUNCATED AT 250 WORDS)

Seventy-one "negative" randomized control trials were re-examined to determine if the investigators had studied large enough samples to give a high probability (greater than 0.90) of detecting a 25 per cent and 50 per cent therapeutic improvement in the response. Sixty-seven of the trials had a greater than 10 per cent risk of missing a true 25 per cent therapeutic improvement, and with the same risk, 50 of the trials could have missed a 50 per cent improvement. Estimates of 90 per cent confidence intervals for the true improvement in each trial showed that in 57 of these "negative" trials, a potential 25 per cent improvement was possible, and 34 of the trials showed a potential 50 per cent improvement. Many of the therapies labeled as "no different from control" in trials using inadequate samples have not received a fair test. Concern for the probability of missing an important therapeutic improvement because of small sample sizes deserves more attention in the planning of clinical trials.

A new type of research, termed meta-analysis, attempts to analyze and combine the results of previous reports. We found 86 meta-analyses of reports of randomized controlled trials in the English-language literature. We evaluated the quality of these meta-analyses, using a scoring method that considered 23 items in six major areas--study design, combinability, control of bias, statistical analysis, sensitivity analysis, and application of results. Only 24 meta-analyses (28 percent) addressed all six areas, 31 (36 percent) addressed five, 25 (29 percent) addressed four, 5 (6 percent) addressed three, and 1 (1 percent) addressed two. Of the 23 individual items, between 1 and 14 were addressed satisfactorily (mean +/- SD, 7.7 +/- 2.7). We conclude that an urgent need exists for improved methods in literature searching, quality evaluation of trials, and synthesizing of the results.

The relative accuracy of randomized control trials (RCTs) and historical control trials (HCTs) in determining effective therapies has not been compared since there is no external verification of efficacy. We reviewed six therapies studied by both methods. Most HCTs concluded therapy was better than control, but few RCTs agreed. We calculated sensitivity and specificity for each type of trial by combining published results with all possible combinations of effectiveness. The sensitivity of HCTs was 0.80 to 1.00 (mean, 0.90) and specificity was 0.0 to 0.27 (mean, 0.11). The sensitivity of RCTs was 0.0 to 0.27 (mean, 0.12) and specificity was 0.67 to 1.00 (mean, 0.88). Defects of RCTs are more easily corrected than those of HCTs. Readers should consider trial design and the probability of errors when deciding how much credence to give to a clinical trial.

Controlled clinical trials of the treatment of acute myocardial infarction offer a unique opportunity for the study of the potential influence on outcome of bias in treatment assignment. A group of 145 papers was divided into those in which the randomization process was blinded (57 papers), those in which it may have been unblinded (45 papers), and those in which the controls were selected by a nonrandom process (43 papers). At least one prognostic variable was maldistributed (P less than 0.05) in 14.0 per cent of the blinded-randomization studies, in 26.7 per cent of the unblinded-randomization studies, and in 58.1 per cent of the nonrandomized studies. Differences in case-fatality rates between treatment and control groups (P less than 0.05) were found in 8.8 per cent of the blinded-randomization studies, 24.4 per cent of the unblinded-randomization studies, and 58.1 per cent of the nonrandomized studies. These data emphasize the importance of keeping those who recruit patients for clinical trials from suspecting which treatment will be assigned to the patient under consideration.