Results availability from clinical research of female urinary incontinence

Aims To ascertain possible publication bias by examining the completeness and publication of studies on nonsurgical treatments for female urinary incontinence (UI). Methods We analyzed information about studies from several trial registries through May 2010. We searched Medline using study registration identifiers to determine publication status, and compared percentages of completed and published studies by study and subject characteristics. Results Among the 166 closed studies, we found 120 completed (73%), 12 terminated (7%), 3 withdrawn (2%), and 4 (2%) that provided no reasons for noncompletion. Only 17% of closed registered studies (28/166 studies) were published in peer-reviewed journals; publication that did occur was an average of 2.2 years after study completion. The proportion of studies published did not increase over time. Studies sponsored by industry were published less often than those funded by NIH (OR=0.04; 95%CI: 0.004-0.38). Drug studies were published less often than studies of other interventions (OR=0.22; 95%CI: 0.05-0.96). Of the 166 closed studies, 7 (4%) posted results on the ClinicalTrials.gov website. Of Phases III and IV trials, 7% and 3% posted results, respectively. Conclusions The absence of results from a substantial proportion of conducted studies suggests that treatment decisions for women with UI are based on selected rather than comprehensive evidence of benefits and harms. Regulatory policy for clinical research should guarantee availability of the outcomes for the public, clinicians, and policymakers.