Social Risk at Individual vs Neighborhood Levels and Health Care Use in Medicaid Enrollees.

Importance: Health-related social risks are increasingly recognized as important contributors to health. Compared with individual screening, neighborhood measures are potentially a lower cost, scalable strategy for identifying social risk. Objective: To inform health resource planning and social risk screening strategies by comparing self-reported vs neighborhood-level social risk with inpatient, emergency department (ED), and outpatient care. Design, Setting, and Participants: A cross-sectional analysis of self-reported social risks measured during Medicaid enrollment and neighborhood-level social risk in relation to health care use was conducted. Members of Kaiser Permanente Northern California, a large integrated health care delivery system, who completed the Medicaid Integrated Outcomes Questionnaire from January 1, 2018, to February 29, 2020, were included. Analysis took place from January 8 to November 29, 2024. Exposure: Neighborhood-level social risk (living in the least-resourced Neighborhood Deprivation Index quartile) and self-reported social risk (indicating a need or wanting help with finances, food, housing, or transportation domains). Main Outcomes and Measures: Hospital and ED admissions, primary care, specialty care, mental health, and social work visits in the year prior to questionnaire completion. Multivariable negative binomial regression models were analyzed for each type of health care use, controlling for demographic characteristics and several health conditions (eg, asthma, hypertension, and chronic pain). Results: Among 13 527 respondents (8631 [63.8%] female; 5289 [39.1%] aged 25-44 years; 2846 [21.0%] Asian, 1986 [14.7%] Black or African American, 3040 [22.5%] Hispanic, 4602 [34.0%] White, and 1053 [7.8%] other race or ethnicity), 33.8% in the most-resourced neighborhood reported at least 1 social risk vs 40.1% in the least-resourced quartile (P < .001). Individual- and neighborhood-level measures were each associated with ED visits (marginal effect estimate for both measures: 0.23; 95% CI, 0.17-0.29). Neither measure was associated with hospital admissions. Individual risk was associated with greater use of all outpatient services (ranging from primary care visit marginal effect estimate: 0.22; 95% CI, 0.13-0.31 to mental health visit marginal effect estimate: 1.21; 95% CI, 0.67-1.75). Neighborhood-level risk was not associated with most outpatient visits and was negatively associated with mental health visits. Conclusions and Relevance: In this cross-sectional study, associations were found for hospital and ED use but not outpatient visits, especially mental health visits. These findings suggest that individual social risk screening appears to provide distinct information compared with neighborhood social risk.