Eligibility definitions and response rate accuracy in studies using routinely collected PROMs: a systematic review

Patient reported outcome measures (PROMs) are increasingly used to assess treatment outcomes and quality of care. This systematic review examined how eligible populations and response rates (RRs) were reported in published studies using routinely collected PROMs.

A systematic literature search in the PubMed, Scopus, Web of Science, Cinahl, PsycINFO and Cochrane library databases on Jan 15, 2024, and Feb 10, 2025, focused on studies employing routinely collected PROM data in research. Articles not related to a health care context were excluded. Definitions of eligibility were categorized as "all treated", "all recruited" or "other". RRs were obtained at baseline (BL) and the last follow-up (FU), and their accuracy was assessed according to the eligibility definitions.

From 3806 abstracts and 344 full-text articles, 92 articles (94 datasets) were included. Eligibility was defined as "all treated" in 26% of the studies and as "all recruited or "other" in 41% and 33%, respectively. Authors frequently fail to apply their own eligibility definitions in RR calculations, leading to incorrect RRs in 47% of datasets due to incomplete eligible populations or flawed denominator choices. Only 5% of studies clearly defined RRs. RR values varied widely. The systematic use of a defined eligible population was linked to a decline in average RRs over the FU.

These findings suggest that inaccurate eligibility definitions and denominator use may bias RRs, limiting the reliability of PROM data. Although PROM datasets are increasingly used, data quality remains underdiscussed. This review highlights the need to improve the reporting of eligibility and RRs while using routinely collected PROMs.

PROSPERO CRD42023494821.