The use of key social determinants of health variables in psychiatric research using routinely collected health data: a systematic analysis.

Abstract

Social determinants of health (SDoH) impact psychiatric conditions. Routinely collected health data are frequently used to evaluate important psychiatric clinical and health services outcomes. This study explored how key SDoH are used in psychiatric research employing routinely collected health data. A search was conducted in PubMed for English-language articles published in 2019 that used routinely collected health data to study psychiatric conditions. Studies (n = 19,513) were randomly ordered for title/abstract review; the first 150 meeting criteria progressed to full-text review. Three key SDoH categories were assessed: (1) gender and sex, (2) race and ethnicity, and (3) socioeconomic status. Within each category, data were extracted on how variables were included, defined, and used in study design and analysis. All studies (n = 103) reported on at least one of the key SDoH variables; 102 (99.0%) studies included a gender and/or sex variable, 30 (29.1%) included a race and/or ethnicity variable, and 55 (53.4%) included a socioeconomic status variable. No studies explicitly differentiated between gender and sex, and SDoH were often defined only as binary variables. SDoH were used to define the target population in 14 (13.6%) studies. Within analysis, SDoH were most often included as confounders (n = 65, 63.1%), exposures or predictors (n = 23, 22.3%), and effect modifiers (n = 14, 13.6%). Only 21 studies (20.4%) disaggregated results by SDoH and 7 (6.8%) considered intersections between SDoH. Results suggest improvements are needed in how key SDoH are used in routinely collected health data-based psychiatric research, to ensure relevance to diverse populations and improve equity-oriented research.