Post-Year Mental Health Treatment Outcomes of Co-Response and Mobile Crisis Models Compared to Law Enforcement Responses

Background and purpose: Law enforcement are the default responders to mental health crises, and their presence can escalate incidents to arrests, excessive force, emergency department transports, and deaths. Mobile crisis and co-response teams are two emerging alternative response models, but few studies have directly compared their post-year patient-level outcomes against law enforcement responses. This study assessed post-year mental health treatment outcomes of mobile crisis and co-response teams compared to law enforcement responses across four jurisdictions. Researchers hypothesized that mobile crisis or co-response recipients would receive more outpatient mental health services, fewer emergency department visits, and fewer inpatient stays than comparable law enforcement-only groups in the year after a crisis.

Methods: Researchers collected crisis response service data from two mobile crisis teams (n=181) and two co-response teams (n=47) in Michigan on services rendered in 2021. Law enforcement reports were collected from the mobile crisis jurisdictions (n=80) and the co-response jurisdictions (n=92) in the same year. For each person who received a crisis service or law enforcement-only response, state Medicaid data provided outpatient, inpatient, and emergency department service history. Inverse probability of treatment weighting balanced the groups by race, sex, age, jurisdiction, and prior-year Medicaid mental health service history. Negative binomial regressions assessed post-year outpatient mental health and emergency department services. A logistic regression assessed psychiatric inpatient usage.

Results: Mobile crisis response was associated with a higher incidence of receiving outpatient mental health services than the respective law enforcement-only response in the post-crisis year (IRR=1.54; CI= 1.150, 2.082; p=.004). Co-response predicted no post-year outpatient differences compared to the respective law enforcement-only response. Mobile crisis response was associated with a lower incidence of post-year emergency department visit than law enforcement-only response (IRR=.477; CI=.277, .818; p=.007). Co-response predicted no post-year difference in incidence of emergency department visit compared to law enforcement-only response. Neither mobile crisis nor co-response was associated with lower odds of psychiatric inpatient service than the respective law enforcement responses.

Conclusions and Implications: This study is the first to compare post-year client-level outcomes between mobile crisis and law enforcement responses. The mobile crisis teams demonstrated favorable outcomes relative to local law enforcement, while the co-response teams did not. The significant difference in post-year emergency department visits suggests mobile crisis teams can prevent future crises, potentially through connecting people to mental health treatment that facilitates stabilization, or increasing familiarity with non-emergency department solutions for future crises. Researchers excluded cases that involved threats to others and balanced the groups on observable baseline characteristics but did not randomize treatment; the law enforcement group may have had a higher risk profile. Mobile crisis teams’ institutional proximity to outpatient services may ease post-crisis connections. Co-response clinicians’ embeddedness in law enforcement departments may limit involvement in post-crisis service coordination, though this study’s small sample size warrants cautious interpretation. Neither crisis response model showed reduced inpatient stays relative to law enforcement; mental health crisis services may prevent future unnecessary ED visits, but there may be a level of acuity too high for mental health crisis services to manage.