Methods: Participants (n=437) are homeless adults, aged 40 years or older, transitioning to PSH in Los Angeles. Interviews were conducted at baseline (prior to move in) using a computerized assisted personal interview (CAPI) and social network interview (SNI) delivered through an iPad. CAPI questions assessed predisposing factors (i.e. race, gender, veteran status, time homeless) and need factors (i.e. chronic physical and mental health conditions, including substance use). Enabling factors drew from the SNI and assessed availability and source of social support (i.e. relative, street peer, emotional support, instrumental support) in respondents’ networks. Univariable analyses were tested with each predisposing, enabling and need factor with four dichotomous outcomes: appropriate medical use (outpatient doctor visits), appropriate mental health use (outpatient therapy visits), and high cost medical and psychiatric hospital use in the past three months. All factors significant (p<.05) at the bivariate level were included in the final multivariable models.
Results: With regard to appropriate service use, the odds of using outpatient medical services were greater for participants who reported the presence of a person who provides emotional support in their social network (CI=1.07, 2.99). Having more chronic physical health conditions also increased the likelihood of outpatient medical visits in the past three months (CI=1.11, 1.54). Participants who identified as women (CI=1.061, 2.878), reported past three month contact with a relative (CI=1.01, 3.01), were diagnosed with depression (CI=2.19, 6.34), PTSD (1.63, 4.55), or a psychotic disorder (CI=1.09, 3.04) all had increased odds of using appropriate mental health treatment. Reporting higher rates of diagnosed chronic physical health conditions (CI=1.32, 2.026) was the sole predictor of inpatient medical use in the final model. Being a veteran (CI=1.01, 3.40), having a diagnosis of depression (CI=1.01, 4.35) and using hard drugs (CI=1.43, 5.87) significantly increased the likelihood of using an inpatient psychiatric facility.
Conclusion: While need factors consistently contributed to service use across all outcomes, these findings indicate a variation and complexity of coinciding predisposing and enabling factors related to increased appropriate and high cost service use. Several social support indicators contributed to appropriate service use, suggesting the characteristics of one’s social networks may also be a critical component in enhancing linkages and preventing high cost utilization.