Reporting of gender-related information in clinical trials of drug therapy for myocardial infarction

Abstract

Background: Concern has been expressed that women are not adequately represented in clinical trials evaluating treatments for medical conditions they commonly experience. This study was designed to assess the reporting of data on women in recently published trials of drug therapy for myocardial infarction, including those funded by an agency with a gender-related policy.

Methods: All randomized controlled trials and meta-analyses of drug therapies for myocardial infarction published in The New England Journal of Medicine, The Lancet, The Journal of the American Medical Association, the Annals of Internal Medicine and the British Medical Journal from January 1992 to December 1996 were evaluated. On preliminary review 102 articles met the inclusion criteria; these were reviewed in detail, and 59 were excluded. Two reviewers independently extracted gender-related information from the 43 articles; discrepancies were resolved by consensus.

Results: Women presented up to 48% of the trial participants (mean 24.1%). In the trials funded by an agency with a gender-related policy, only 16.8% of participants, on average, were women. Of the 43 articles in the sample, only 14 (32%) provided gender-related results. Funding from an agency with gender-related policy did not affect the reporting of gender-related information. Subgroup analyses were provided for 14 (32%) of the 43 trials, including 2 (29%) of 7 trials funded by an agency with a gender-related policy. Of the 12 trials that included interaction analyses (excluding the 2 trials in which secondary analyses were conducted specifically to identify differences between women and men), 7 (58%) conducted an interaction analysis to determine if women responded differently than men; for one of these the interaction analysis was for a secondary outcome measure (drug safety). Only 5 (12%) of the 43 articles mentioned the differences between men and women in the Discussion section; 2 of these were studies that used secondary analyses to examine sex differences. Of the 5, only 1 was funded by an agency with a gender-related policy.

Interpretation: Women were poorly represented in the randomized controlled trials in this sample, regardless of whether the trials were funded by an agency with a gender-related policy. Structured reporting of gender-related information for clinical trials may improve the quality of information available about women and therefore facilitate the application of research findings to the care of women.