Contribution of pharmaceutical manufacturer data submissions in identifying relevant studies for systematic reviews

Article type
Authors
Thakurta S1, Peterson K1, McDonagh M1
1Medical Informatics and Clinical Epidemiology, Oregon Health and Science University, Portland, Oregon, United States
Abstract
Background: For systematic reviews it is critical to identify as complete a set of relevant studies as possible. Given the limitations of electronic searching, use of additional means for identifying relevant studies is desirable. To address this need, the Drug Effectiveness Review Project (DERP) routinely solicits published and unpublished data from pharmaceutical manufacturers. Incorporation of stakeholder input also provides DERP the opportunity to improve the transparency of the systematic review process. Objective: To compare volume and types of relevant studies identified via pharmaceutical manufacturer solicitation versus those from electronic searches. Method: By October 2010, data from pharmaceutical manufacturer submissions and electronic searches will be collected for approximately 30 DERP reports conducted since 2007. Data regarding characteristics and volume of literature will be extracted by one reviewer and checked by another. The proportion of final included studies obtained through the pharmaceutical manufacturer submission process will be identified. Characteristics of studies identified by each method will be explored (e.g. direct comparisons, subgroup analyses, publication status). Results: Preliminary analysis found that pharmaceutical manufacturer submissions identified at least one additional publication for 75% of DERP reviews completed in 2009. Mean total number of citations submitted by manufacturers was 263 [standard deviation (SD), 193]. Among those, the mean number of citations not already identified by electronic searches was 5 (SD, 7), or approximately 2%. Mean number of included studies from all sources is 67 (SD, 43). Hence, the overall contribution of manufacturer submissions towards the relevant studies is about 7%; with only 5% of those being direct comparison trials. Conclusions: Initial findings indicate that pharmaceutical manufacturer submissions contribute &10% of the final set of included studies. Although this proportion is small, these additional studies include highly relevant direct comparison studies. The number of studies screened to identify these few important studies was large.