How do we measure data sharing in the biomedical sciences? A measurement systematic review of biomedical data sharing-related knowledge, attitudes and practices across stakeholder groups, data types and geographies

TL;DR

This paper reviews how data sharing is measured in biomedical research, finding that most surveys lack rigorous development and are limited in usefulness for comparing attitudes across groups and regions.

Contribution

The first systematic comparison of survey instruments for measuring data-sharing knowledge, attitudes, and practices in biomedical research.

Findings

Most surveys originated from high-income countries and were used only once.

Fewer than one-third of surveys reported pilot testing, and only three provided measurement properties.

The review highlights the need for better survey development to support cross-country and cross-disciplinary comparisons.

Abstract

Enabling the reuse of participant-level health data is central to advancing public health and clinical practice. Measuring knowledge, attitudes and practices (KAP) related to data sharing is essential for understanding how stakeholders perceive data reuse and where further investment is needed. We conducted a measurement systematic review to identify and describe the development, scope and measurement properties of quantitative surveys assessing data-sharing-related KAP in biomedical research.

Systematic review using the COnsensus-based Standards for the selection of health status Measurement INstruments (COSMIN) approach.

Ovid (MEDLINE), EMBASE, CINAHL, PsycINFO and HaPI were searched for relevant surveys from 1 January 2000 to 7 April 2021. The Ovid (MEDLINE) search was updated on 30 May 2022 and 15 April 2024.

Quantitative surveys measuring knowledge, attitudes, behaviours or practices related to sharing or reusing participant-level health data were included.

Two independent reviewers screened studies, extracted data and, where possible, applied the COSMIN Risk of Bias checklist to assess survey measurement properties. We summarised survey scope, target populations, data types, development and measurement properties narratively. Due to substantial heterogeneity, survey findings were not compared across studies.

We screened 3684 title-abstracts, reviewed 104 full texts and extracted data from 72 publications representing 60 independent surveys. Most surveys originated from high-income countries and were used only once. Fewer than one-third reported pilot testing. Only six surveys provided sufficient information to apply COSMIN, and only three reported measurement properties, indicating low certainty in the available evidence.

This is the first systematic comparison of the development and measurement properties of quantitative survey instruments assessing data-reuse KAP. Most surveys lacked rigorous development and reporting, limiting their utility for comparing KAP related to data sharing across stakeholders and settings. The review findings will inform the creation of a cross-country, cross-disciplinary question bank to support future tool development.

CRD42021243926.

Figures