This report presents the findings of a mixed-methods randomized field experiment evaluating the Serving A Vulnerable Entity (SAVE) initiative operated by the Southeastern Pennsylvania Transportation Authority (SEPTA) Transit Police. The initiative sought to determine whether adding a contracted civilian outreach specialist to a police outreach team improved the likelihood that individuals experiencing homelessness, mental illness, addiction, or other vulnerability conditions would be transported to an appropriate treatment or shelter facility. The study responds to national interest in alternative crisis response models and the increasing intersection of policing and public health within urban transit systems.
Across a 12-month period (June 2022–June 2023), SAVE officers conducted 158 treatment conversations involving 165 individuals in or around central Philadelphia transit stations. Shifts were randomly assigned to either a police-only condition or a police-plus-outreach-specialist condition. Treatment conversations were extended, need-focused discussions that included identifying appropriate facilities and offering transportation. Two primary outcome measures were examined: (1) “agree to transport,” indicating willingness to accept services, and (2) “treatment initiation,” defined as successful conveyance to a treatment or shelter facility. Complementing the quantitative experiment, the research team conducted 150 hours of field observation and semi-structured interviews with SAVE officers, outreach specialists, and police leadership to explore implementation processes, contextualize outcomes, and assess program fidelity.
Results showed that individuals were modestly more likely to be transported when engaged by a police-specialist team (41 percent) than by police officers alone (32 percent), representing a relative increase of 29 percent. However, this difference did not reach statistical significance, due in part to staffing-related implementation challenges that produced an imbalance in treatment exposure (specialist-supported conversations with 61 people vs. police-only conversations with 104 people), and in part due to the overall low n of the study. Qualitative findings illuminated factors influencing outcomes, including pervasive client resistance, the narrow and unpredictable window in which individuals are willing to accept help, and the importance of relationship-building through repeated contacts. Stakeholders also emphasized substantial variation in specialist readiness, difficulties with hiring and retention, the physically demanding nature of the transit environment, and the need for clear co-response protocols, safety procedures, and consistent training.
The evaluation demonstrates that co-response in a transit environment is feasible and can produce modest operational benefits, but that outcomes depend heavily on specialist quality, program stability, and the practical realities of street-level outreach. Findings suggest that well-trained police officers can independently achieve treatment initiation at rates nearly comparable to specialists. Agencies considering similar programs should prioritize specialist recruitment, durable staffing models, and robust preparation for both partners in co-response teams. Further research is warranted to explore long-term engagement outcomes, cost-benefit implications, and alternative deployment strategies within transit and urban settings.
(Publisher abstract provided.)
Downloads
Similar Publications
- Treating Drug-Abusing Women Prisoners: An Outcomes Evaluation of the Forever Free Program
- Strain of Change: Voices of Los Angeles Police Officers
- A Computational Study on the Atmospheric Fate of Carbon-Centered Radicals from the 3-Methyl-2-butene-1-thiol + •OH Reaction: Mechanistic Insights and Atmospheric ImplicationsArticle link copied!