Abstract 20929: Risk Reduction Reporting in Cardiovascular Clinical Trials
Background: The validity and quality of data reporting is often a point of contention. In the past decade there has been an increased effort to ensure the quality of data reported, most notably the Consolidated Standards of Reporting Trials (CONSORT). These efforts include emphasizing the importance of reporting Absolute Risk Reduction (ARR) and Number Needed to Treat (NNT), as opposed to Relative Risk Reduction (RRR) alone.
Method: We conducted a systematic PubMed search for original consecutive cardiovascular (CV) articles in the New England Journal of Medicine (NEJM), Journal of the American Medical Association (JAMA), and Lancet between January 2000 and April 2008 as part of the Clinical Trials Reporting Database (CTRD). Articles were abstracted for: (1) measures of benefit: whether RRR, ARR, and/or NNT were reported; (2) study design: hypothesis type, intervention type, and placebo or active controls used; and (3) whether the primary outcome measure was positive.
Results: Of the 566 CV publications reviewed, 207 did not report RRR, ARR, or NNT, 212 only reported RRR, 29 only reported ARR, 3 only reported NNT, 115 reported RRR with ARR and/or NNT. Active controlled drug trials were more likely to report RRR alone (64.6% vs. 51.7%; p=0.064) than placebo controlled trials. In trials with a superiority hypothesis, active controlled trials more often only reported RRR as compared to placebo-controlled trials (66.1% vs. 43.4%, p=0.006). This finding remained significant with the omission of trials without positive findings, for which an ARR would not be expected (61.8% vs. 37.5%, p=0.022).
Conclusion: The reporting of ARR and NNT has been extremely low, particularly in active controlled trials. The practice of using RRR solely as a measure of benefit trials can be problematic particularly in active controlled where small absolute differences between treatment groups are expected.
- © 2010 by American Heart Association, Inc.