Impact of San Francisco's New Street crisis response Team on Service use among people experiencing homelessness with mental and substance use disorders: A mixed methods study protocol.

Mobile crisis services for people experiencing distress related to mental health or substance use are expanding rapidly across the US, yet there is little evidence to support these specific models of care. These new programs present a unique opportunity to expand the literature by utilizing implementation science methods to inform the future design of crisis systems. This mixed methods study will examine the effectiveness and acceptability of the Street Crisis Response Team (SCRT), a new 911-dispatched multidisciplinary mobile crisis intervention piloted in San Francisco, California. First, using quantitative data from electronic health records, we will conduct an interrupted time series analysis to quantitatively examine the impacts of the SCRT on people experiencing homelessness who utilized public behavioral health crisis services in San Francisco between November 2019 and August 2022, across four main outcomes within 30 days of the crisis episode: routine care utilization, crisis care reutilization, assessment for housing services, and jail entry. Second, to understand its impact on health equity, we will analyze racial and ethnic disparities in these outcomes prior to and after implementation of the SCRT. For the qualitative component, we will conduct semi-structured interviews with recipients of the SCRT's services to understand their experiences of the intervention and to identify how the SCRT influenced their health-related trajectories after the crisis encounter. Once complete, the quantitative and qualitative findings will be further analyzed in tandem to assist with more nuanced understanding of the effectiveness of the SCRT program. This evaluation of a novel mobile crisis response program will advance the field, while also providing a model for how real-world program implementation can be achieved in crisis service settings.