Diverting Persons Experiencing a Behavioral Health Crisis from the Criminal-Legal System at Intercept 0: An Examination of Emergency Dispatch and Law Enforcement Calls for Service Codes

Background and Purpose: The Sequential Intercept Model (SIM) maps the criminal-legal system as a continuum with attention to various “intercepts” where persons with mental health (MH) issues can be diverted from this system (SAMHSA, 2020). The SIM consists of five intervention points, with Intercept 0 (Community Services) and Intercept 1 (Law Enforcement; LE) being critically important for preventing further enmeshment into the criminal-legal system (courts, jails, prisons, etc.). Promising practices for MH diversion exists at Intercept 1, with the development of models of interventions such as specialized outreach (Chapman et al., 2020), but limited investigations of dispatch exist. This study examines how dispatch and LE code calls for service when the crisis event is MH related, specifically how the codes vary between these agencies and what demographic and community level factors impact coding.

Methods: All LE call reports primarily or secondarily related to MH calls for service (N=1,538) between January 1, 2014 and December 30, 2016 from a Midwestern county were examined. Specifically, dispatch classification codes on these reports were collapsed into four categories: Mental Health, Crime-Related, Community-Related, and Other. Descriptive statistics were used to determine the frequency and types of classification codes that were a positive match between dispatch and LE. Bivariate analysis was then conducted to determine what factors, if any, where indicators for a more successful match. Variables examined included age, gender, race, drug use at time of call, and community poverty status of the individual in crisis.

Results: Of the 1,538 LE call report classification codes that were examined, 44.9% (n=691) were found to be a match between LE and emergency dispatch and 55.1% (n=847) were not. Of the 691 matched calls, 92.8% (n=641) were coded as MH and the remaining were coded as community-related (4.2%, n=29), crime-related (1.4%, n=10), and other (1.6%, n=11) calls. Of all calls, dispatch coded 934 as MH related, of which 68.6% (n=641) matched successfully with LE, while 31.4% (n=293) did not. Further, while age, gender, race, and drug use at time of call were not found to significantly indicate a match; the community poverty rate (per 100,000 residents) was found to be significantly lower for matched versus unmatched calls (8.97 vs. 10.08, t(1519)=-2.20, p<.05).

Conclusions and Implications: These results suggest that emergency dispatchers are able to identify MH crisis more often and accurately then other crisis calls. However, these results suggest a bias towards communities of poverty, since MH related calls originating in lower poverty areas were more likely to be miscoded by dispatch as crime-related. While improvements at Intercept 1 to MH crisis calls are promising, they have largely occurred after dispatch has deployed law enforcement. Intervening at dispatch, including education on how MH disproportionately effects those living in poverty (Funke et al., 2012), could result in greater improvements for persons experiencing a MH crisis and their families and could alleviate law enforcement from responding to such calls for service which are best addressed by other social systems, such as community mental health.