Association between author conflicts of interest and industry‐sponsorship with the favorability of outcomes of systematic reviews focusing on treatments of erectile dysfunction

Abstract

Background: Authors' conflicts of interest and industry sponsorship have been shown to influence study outcomes. Objective: We aimed to determine whether author conflicts of interest and industry sponsorship influenced the nature of results and conclusions of systematic reviews focusing on treatment interventions for erectile dysfunction. Materials and methods: We searched PubMed and Embase for systematic reviews and meta-analyses focusing on erectile dysfunction treatments published between September 1, 2016, and June 2, 2020. Authors' conflicts of interest were collected from the systematic reviews' disclosure statements. These disclosures were verified using the information provided by the Open Payments, Dollars for Profs, Google Patents, and US Patent and Trademark Office databases and from previously published disclosure statements.Results: Our study included 24 systematic reviews authored by 138 authors. Nineteen authors (13.8%) were found to have conflicts of interest (disclosed, undisclosed, or both). No authors completely disclosed all conflicts. Nine reviews (37.5%) contained at least one author with conflicts of interest; of which eight reported narrative results favoring the treatment group, and seven reported conclusions favoring the treatment group. Of the 15 (62.5%) reviews without a conflicted author, 11 reported results favoring the treatment group, and 12 reported conclusions favoring the treatment group. Discussion:The results and conclusions of systematic reviews for erectile dysfunction treatments did not appear to be influenced by authors who reported conflicts of interest. However, our search algorithm relied on the US-based Open Payments database and a large percentage of reviews in our study were produced by authors with international affiliations. Our study results underscore the difficulties in conducting such analyses.