Synergistic effects of social determinants of health and race-ethnicity on 30-day all-cause readmission disparities: a retrospective cohort study.

OBJECTIVE

The objective of this study is to assess the effects of social determinants of health (SDOH) and race-ethnicity on readmission and to investigate the potential for geospatial clustering of patients with a greater burden of SDOH that could lead to a higher risk of readmission.

DESIGN

A retrospective study of inpatients at five hospitals within Henry Ford Health (HFH) in Detroit, Michigan from November 2015 to December 2018 was conducted.

SETTING

This study used an adult inpatient registry created based on HFH electronic health record data as the data source. A subset of the data elements in the registry was collected for data analyses that included readmission index, race-ethnicity, six SDOH variables and demographics and clinical-related variables.

PARTICIPANTS

The cohort was composed of 248 810 admission patient encounters with 156 353 unique adult patients between the study time period. Encounters were excluded if they did not qualify as an index admission for all payors based on the Centers for Medicare and Medicaid Service definition.

MAIN OUTCOME MEASURE

The primary outcome was 30-day all-cause readmission. This binary index was identified based on HFH internal data supplemented by external validated readmission data from the Michigan Health Information Network.

RESULTS

Race-ethnicity and all SDOH were significantly associated with readmission. The effect of depression on readmission was dependent on race-ethnicity, with Hispanic patients having the strongest effect in comparison to either African Americans or non-Hispanic whites. Spatial analysis identified ZIP codes in the City of Detroit, Michigan, as over-represented for individuals with multiple SDOH.

CONCLUSIONS

There is a complex relationship between SDOH and race-ethnicity that must be taken into consideration when providing healthcare services. Insights from this study, which pinpoint the most vulnerable patients, could be leveraged to further improve existing models to predict risk of 30-day readmission for individuals in future work.