Abstract 20973: The Impact of Independent Statistical Analysis on Reported Outcomes in Cardiovascular Clinical Trials
Background: The influence of independent statistical analysis on the outcomes of cardiovascular trials has not been well documented. While, industry sponsored cardiovascular trials are more likely to report positive findings compared with non-industry sponsored trials, the mechanism underlying this phenomena is unclear.
Methods: 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 self-reported disclosure of whether an independent statistical analysis was conducted. Subgroup analysis by funding source was also conducted.
Results: Of the 509 CV publications reviewed in which details of an analysis plan were provided, 277 (54%) reported having an independent statistical analysis. Of these, 249 were solely industry funded, 173 non-industry funded (NIH, non-NIH government, clinical research networks, and university organization), and 87 a combination of both. Studies with an independent statistical analysis did not differ is the reporting rate of positive findings as compared to those without independent analysis (p=ns). While, industry sponsored trials were less likely to conduct an independent analysis than the combination of jointly funded or non-industry funded trials (45.4% vs 63.1%, p=0.001), no difference in the reporting of positive findings was noted.
Conclusion: Although industry sponsored trials are less likely to have independent statistical analysis, this does not appear to impact the likelihood of reporting of positive findings.
- © 2010 by American Heart Association, Inc.