Associations between social risks and primary care utilization among medically complex veterans

Abstract

Importance: Social risks lead to poor health outcomes, especially for medically complex patients. Primary care may modify outcomes, but patients with social risks face barriers to accessing primary care. Objective: To study associations between social risks and primary care utilization among medically complex patients. Design: Prospective cohort study among respondents to a 2018 mailed survey collecting information on social risks, followed up to two years after survey completion. Setting: Veterans Affairs (VA) health care system, in which primary care is delivered through a Patient-Aligned Care Team including the primary care provider (PCP), registered nurse, clinical pharmacist, and social worker. Participants: Survey was mailed to a nationally representative sample of 10,000 VA primary care patients with high (≥75th percentile) one-year risk of hospitalization or death. Exposures: Survey-based measures of low social support, not having a family member/friend who gets involved in your health care, unemployment, transportation problem, food insecurity, medication insecurity, financial strain, low medical literacy, less than high school graduate, and ≥1 social risk. Main outcomes and measures: Electronic health record-based number of PCP encounters, number of primary care team encounters (PCP, nurse, clinical pharmacist, and social worker), and ≥1 vs. 0 social work encounters in the two years after survey completion. Negative binomial regression models were used for PCP and primary care team utilization. Logistic regression models were used for social work utilization. Results: Among 4,680 respondents, mean age was 70.3 (SD 11.5), 93.7% were male, 71.8% White non-Hispanic, and 15.8% Black non-Hispanic. In fully adjusted models, unemployment was associated with fewer PCP and primary care team encounters, and low medical literacy was associated with fewer primary care team encounters. Among those with ≥1 social risk, 18.4% had ≥1 social work encounter. Low social support, transportation problem, and low medical literacy were associated with higher odds of ≥1 social work encounter. Conclusions: There were minimal disparities in PCP and primary care team utilization among medically complex Veterans with and without social risks. However, social work use was low, despite its central role in addressing social risks. More work is needed to understand barriers to social work utilization.