Individual and clinical factors associated with patient acceptance of referrals to social services and community resources at a multi-purpose resource hub.

Emerging evidence suggests that bi-directional communication and referral pathways, when employed strategically, can lead to favorable health outcomes by connecting patients with complex, multi-faceted health and social needs to appropriate services and resources. However, despite these benefits, patient acceptance of referrals via these pathways remains suboptimal. In this study, we describe individual and clinical factors associated with patient acceptance of these referrals. We extracted individual-level demographic and clinical data for patients referred primarily from a large safety-net health system to a multi-purpose resource hub co-located on the campus of its largest hospital, for the period October 2019 to June 2023. Descriptive statistics, Chi-square analyses, and multinomial regression modeling were performed to examine these data. Of 1865 patients in the study sample, 54.2% accepted a referral, 27.4% were lost to follow-up, and 18.4% declined. Most patients who accepted referrals were female (67.1%), Latino (81.5%), and had hypertension and/or prediabetes or diabetes (84.1%). In modeling analyses, those who accepted referrals tended to be female, and were referred from primary care clinics; many were referred for multiple service/resource categories. We found associations between patient acceptance of referrals and gender and source of referral. Drawing upon these results as well as experience implementing these systems, we propose several practical strategies for increasing successful referrals, including identifying and addressing barriers for patients who declined or were lost to follow-up; using standardized screening tools to routinely assess for multi-faceted health and social needs; increasing provider awareness about the benefits and functioning of these pathways; and monitoring progress so mid-course adjustments can be made when necessary.