Vice President Medical Affairs, Pharmaspectra IQVIA Parsippany, United States
Background: Only 37.3% (95% CI, 35.3% to 39.3%) of abstracts presented at scientific conferences reach peer-reviewed publication. Randomized-controlled trials show greater estimated likelihood of reaching peer-reviewed publication (68.7%) compared to other study designs (44.9%) (Scherer, et al. Cochrane Database of Systematic Reviews 2018, Issue 11). Insights from well-designed sub-analyses and post-hoc analyses (SAPHAs) presented at scientific conferences are valuable for epidemiologists and life sciences professionals to refine target study populations, characterize drug use and treatment outcomes, and more. Such evidence therefore may remain unrealized by those using publication databases, such as PubMed, for research planning.
Objectives: Quantify loss of scientific evidence from SAPHAs between scientific conference presentations (captured in a conference publication database, CPD) and publications (captured in PubMed database) using a therapeutic area with a high degree of epidemiologic evidence as a test case.
Methods: We selected hepatitis C (HC) as an area for identification from the scientific literature. A standard Boolean search strategy identifying SAPHAs for HC was developed and used to query CPD for a 5-year time period, Jan 2016 through Dec 2020, with this upper-bound date set to allow for a lag-period of at least two years to achieve journal publication. For each of the conference presentations from CPD, the best matching PubMed journal abstract was obtained using Similarity Search Retrieval, a statistical, linguistic approach for the automated generation of search expressions that retrieves relevant documents ranked by similarity to a model set of documents. Matched pairs of presentations and publications were reviewed to confirm concordance and assess journal publication rate.
Results: For the 5 year study period, 97 SAPHA conference abstracts were identified for HC. A total of only 38 (39.2%) were identified in the scientific literature. This included analyses which reached publication as stand-alone or as part of meta-analyses or integrated analyses.
Conclusions: In the area of HC, potentially useful SAPHAs had been presented at the conference level but were not found in PubMed. Although limited to HC, this finding suggests that health/research professionals are limited to a subset of scientific evidence, potentially impacting important research and decision making. Further research is necessary to determine the significance of unpublished SAPHAs on a larger scale and the reasons why they may or may not be accepted or submitted for publication. Studies in additional therapeutic areas using the above methodology are ongoing.
