Health Outcomes of Interest in Observational Data: Issues in Identifying Definitions in the Literature

ObjectivesTo describe the literature search strategies and results to inform development of health outcomes of interest (HOI) definitions studied within the Observational Medical Outcomes Partnership (OMOP) research program. These HOIs represent the safety and benefit endpoints that may be associated with the use of particular medications.MethodsTwo organizations implemented search strategies to identify articles that described the validation of various HOI definition algorithms and produced systematic reviews of the published literature for the OMOP HOIs; reviews of 5 HOIs were available from each organization. Search results were summarized descriptively.ResultsBased on the differences in publications identified by each organization, neither search strategy captured all the relevant literature across 5 HOIs. A composite search strategy developed by the OMOP research team failed to effectively capture all relevant publications as well. Among publications in observational databases, a large proportion did not report actual codes used to identify HOIs in administrative claims databases, nor did they provide substantive detail of any validation studies.ConclusionNo single search strategy or literature database captured all relevant articles efficiently. The lack of specific search terms hampered identification of relevant articles. Due to limited details about positive predictive value of various HOI definitions, information culled from available literature was insufficient to identify one best definition for each HOI. Researchers seeking to capture and understand prior published work that defined relevant outcomes of interest need to be aware of the shortcomings of this approach and the lack of detail of validation studies available in published literature.