Evaluation of repositories for sharing individual-participant data from clinical studies (Banzi et al., 2019)

This article reviews how well existing data repositories can host and share individual participant data and related documents from clinical studies. It is intended to help, especially non‑commercial sponsors and investigators, choose suitable repositories when planning data sharing and writing data sharing plans.

The authors screened 55 candidate repositories and evaluated 25 that were accessible to non‑commercial researchers and able to host individual participant data. For each repository they assessed 34 items in four domains: general characteristics, data upload and storage, data management and access, and discoverability. From these they derived eight key suitability indicators: guidance for upload and storage, support for de‑identification, data quality controls, contracts, metadata, identifiers, flexibility of access and long‑term preservation.

The repositories span generic, multidisciplinary platforms, social science and health repositories, national or institutional services, and disease‑ or funder‑specific repositories. Most do not charge for data deposition or access; a few (for example Dryad, Figshare, Vivli) apply fees. None meets all eight indicators in full, but three generic repositories (Dryad, DRUM and EASY) fulfil all of them fully or partially.

Investigators can use the eight indicators as a concise checklist when selecting a repository, consult the article’s tables to shortlist candidates that fit their study’s needs, and use the indicators to frame follow‑up questions where repository information is incomplete. The authors conclude that the current landscape is fragmented, with important information often missing, and they call for more standardised evaluation criteria and clearer repository descriptions to support wider, higher‑quality data sharing.