Challenges in identifying published registration studies for FDA-approved pharmaceuticals

presents a challenge. Product approvals are based on results of registration studies. We tested an approach to identify registration studies using clinical trial data in prescribing information (PI). Research design and methods: PIs were reviewed for a sampling of products granted FDA approval during 2011-2013. Registration study data were assessed in the clinical trial sections of PIs, and PubMed was searched using the generic drug name and “phase 3,” the latter as a surrogate for registration. Results: 29 products were identified.* Some discrete trial or aggregate-only information was provided in 32% of PIs; 68% provided details by trial, the latter mostly with single registration trials. No PI included ClinicalTrials.gov identifiers or published data citations. Results of cross-matching a random sampling (n=9) of PIs showed no correlation between numbers of studies included in PIs, number of PubMed hits from the above search, and number of papers with phase 3 in the title. More detailed searching, eg, with patient numbers and drug regimens, yielded better yet still inconsistent results and caused a large increase in time spent to identify publications. Conclusions: Identifying published registration studies appears to be an exercise in back guessing based on various parameters. A standardized system is needed whereby registration study data can be easily tied to published studies.