Background: Annually 1.5 million Americans face housing insecurity, and compared to their domiciled counterparts are three times more likely to utilize the Emergency Department (ED). Community Based Participatory Research (CBPR) methods have been employed in underserved populations, but use in the ED has been limited. We employed CBPR with a primary goal of improved linkage to care, reduced ED recidivism, and improved homeless health care.
Methods: A needs analysis was performed using semi-structured individual interviews with participants experiencing homelessness as well as with stakeholders. Results were analyzed using principles of grounded theory. At the end of the interviews, respondents were invited to join the CBPR team. At CBPR team meetings, results from interviews were expounded upon and discussions on intervention development were conducted.
Results: Twenty-five stakeholders were interviewed including people experiencing housing insecurity, ED staff, inpatient staff, and community shelters and services. Three themes emerged from the interviews. First, the homeless population lack access to basic needs, thus management of medical needs must be managed alongside social ones. Second, specific challenges to address homeless needs in the ED include episodic care, inability to recognize housing insecurity, timely involvement of ancillary staff, and provider attitudes towards homeless patients affecting quality of care. Lastly, improved discharge planning and communication with outside resources is essential to improving homeless health and decreasing ED overutilization. A limitation of results is bias towards social networks influencing included stakeholders.
Conclusion: CBPR is a promising approach to address gaps in homeless health care as it provides a comprehensive view incorporating various critical perspectives. Key ED-based interventions addressing recidivism include improved identification of housing insecurity, reinforced relationships between ED and community resources, and better discharge planning.