(153) Data Source Heterogeneity in Multi-Data Base Pharmacoepidemiologic Studies: An ISPE-sponsored Scoping Review

Publication Number: 484

Background: Interpreting the findings of multi-data base studies requires a clear understanding of the context and purpose of data collection for each participating data source.
There is currently no guidance for how heterogeneity between data sources (HDS) should be reported and leveraged upon. The DIVERSE scoping review on this topic is sponsored by ISPE.

Objectives: To report on the selection for a scoping review to identify literature describing recommendations, or tools to describe, report on or leverage HDS for pharmacoepidemiologic research.

Methods: We identified relevant literature published up to end 2021. We first selected a set of core papers, based on the expertise of the co-authors. We then executed three steps: 1) a snowball search of the reference lists of the core papers, 2) a PubMed search and 3) a web-based search for grey literature. Documents were reviewed in a standardised process by independent pairs of reviewers; first title and abstract (TIAB), then full text review (FTR) if TIAB was of interest. We excluded documents if they met any of the following criteria: 1) described only clinical trials and not observational (or routinely collected) data sources, 2) reported only statistical methods to handle heterogeneity in results (e.g. meta-analysis), 3) applied pharmacoepidemiologic studies that were not focussed on methodology or providing guidance, 4) papers otherwise out of scope (e.g. pre-clinical studies). We included documents if they reported at least one of the following: 1) relevant descriptions of multiple data sources, 2) tools to report on HDS, 3) strategies to use HDS to improve evidence, 4) guidance on reporting HDS. The initial core papers were automatically included in the scoping review.

Results: We selected 23 core documents and retrieved 687 documents (total = 710). After TIAB, 186 documents (27.1% of 687) were included for FTR. After FTR, 50 (26.9% of 186) were included in the scoping review by the independent pair of assessors. Including the core papers, a total of 73 documents entered the scoping review, 42 (57.5%) included relevant descriptions of multiple data sources, 42 (57.5%) included tools to report on HDS, 52 (71.2%) included strategies to use HDS to improve evidence, and 32 (43.8%) included guidance on reporting on HDS.

Conclusions: Among the reviewed papers, a minority included guidance on reporting diversity, while a higher number included descriptions, tools, and strategies. This set of papers will undergo further content analysis, to extract common themes and identify knowledge gaps. This scoping review will eventually inform the development of a guidance paper.