Potentially Identifying Variables Reported in 100 Qualitative Health Research Articles: Implications for Data Sharing and Secondary Analysis

Abstract

There is a growing trend in social science research to share qualitative data in a repository for others to access. However, some researchers are reticent to share qualitative data. One major concern is how to de-identify qualitative data while maintaining adequate contextual detail to allow secondary users to meaningfully interpret de-identified data. To help inform discussions regarding qualitative data sharing, we reviewed 100 qualitative health science studies to determine what potentially identifying variables (PIVs) are reported in the published literature. There are relatively few PIVs reported in each qualitative study; the majority of studies (n=64) reported two or fewer PIVs. The most commonly reported PIVs were profession, sex or gender, and age. Our findings can help guide de-identification efforts going forward as presumably the PIVs that are most commonly reported provide essential contextual details that will also be needed by secondary users, while PIVs that are rarely reported may not provide essential contextual information for interpretation of data. This suggests it is possible to share qualitative data that are both de-identified and useful for secondary analysis. As data are shared, we recommend researchers mask study sites, as these may uniquely increase the chance of re-identification.

Keywords

• qualitative data sharing
• health sciences
• secondary analysis
• potentially identifying variables
• de-identification