Measures

Outcome measures were extracted from discrete EHR data fields and included two binary measures indicating: 1) whether the patient was ever screened for FRS during the study period (repeat screening was not considered); and 2) among those screened, whether the patient had a documented self-reported need in any FRS domain.

Race and ethnicity^28,29 are data fields in the EHR that are intended to be based on patient self-report. They were used to create the main independent variable which consisted of seven groups: three in which the patient did not self-identify with Hispanic ethnicity and reported White, Black, or other race (non-Hispanic White, non-Hispanic Black, and non-Hispanic other race); three in which patients identified as Hispanic and with a race (Hispanic White, Hispanic Black, Hispanic other race); and one for which there was no data for either the Hispanic indicator or categorical race reported as Race / Ethnicity Unknown. ―Other race‖ reflects the grouping of racial / ethnic categories with smaller samples as captured in the OCHIN database for the purposes of this analysis.²⁸ Groups classified as Other race identified with either American Indian / Alaska Native, Asian, Native Hawaiian / Pacific Islander, or multiple races.

To account for potential confounding, patient-level variables were included as covariates for preferred language (English, Spanish, other), sex as documented in the EHR, age group (age 18–39, 40–64, 65+ years), insurance type at last encounter (private, public, uninsured), last recorded federal poverty level (>200% FPL, <=200% FPL, not documented), total number of visits in the study period, and presence of a documented cardiometabolic disease in the problem list (diabetes mellitus, hypertension, dyslipidemia, or obesity), per International Classification of Disease 9 & 10 codes (Appendix). Last, an indicator variable was included noting whether the patient’s clinic had received social risk screening implementation support in a prior study.