C A Silagy1, D Jewell. 1. Department of Public Health and Primary Care, University of Oxford.
Abstract
BACKGROUND: Randomized controlled trials are being used increasingly to evaluate the effectiveness of health care interventions, including those in primary care. AIM: A study was undertaken to document the topic areas addressed by randomized controlled trials conducted in primary care and the methodological characteristics of these, based on a historical review of all trials published in the British Journal of General Practice (formerly the Research Newsletter and Journal of the Royal College of General Practitioners) between 1953 and 1991 inclusive. METHOD: Trials were identified by systematically hand searching all back issues. RESULTS: A total of 90 randomized controlled trials were identified, 78% of which were undertaken in general practice. The trials covered a wide range of topics; 62% examined pharmacological interventions, 21% non-pharmacological interventions, and 16% an intervention related to provision of an aspect of health service. Among the randomized controlled trials published there was a significant trend towards more reports on non-pharmacological trials in recent years (P < 0.001). The quality of trials, judged by the extent to which bias was controlled, varied considerably. Trials of pharmacological interventions generally controlled for selection bias at entry and at assessment of outcome more effectively than trials of non-pharmacological interventions. CONCLUSION: Some imaginative solutions to the logistic difficulties of conducting randomized trials in general practice were noted. These may serve as an incentive to those undertaking such studies in the future.
BACKGROUND: Randomized controlled trials are being used increasingly to evaluate the effectiveness of health care interventions, including those in primary care. AIM: A study was undertaken to document the topic areas addressed by randomized controlled trials conducted in primary care and the methodological characteristics of these, based on a historical review of all trials published in the British Journal of General Practice (formerly the Research Newsletter and Journal of the Royal College of General Practitioners) between 1953 and 1991 inclusive. METHOD: Trials were identified by systematically hand searching all back issues. RESULTS: A total of 90 randomized controlled trials were identified, 78% of which were undertaken in general practice. The trials covered a wide range of topics; 62% examined pharmacological interventions, 21% non-pharmacological interventions, and 16% an intervention related to provision of an aspect of health service. Among the randomized controlled trials published there was a significant trend towards more reports on non-pharmacological trials in recent years (P < 0.001). The quality of trials, judged by the extent to which bias was controlled, varied considerably. Trials of pharmacological interventions generally controlled for selection bias at entry and at assessment of outcome more effectively than trials of non-pharmacological interventions. CONCLUSION: Some imaginative solutions to the logistic difficulties of conducting randomized trials in general practice were noted. These may serve as an incentive to those undertaking such studies in the future.