Skip to main content
NCBI home page
Oxford University Press logoLink to Oxford University Press
. 2021 Feb 28;15(3):1948–1962. doi: 10.1093/police/paab010

The Impact of Crisis Intervention Team Response, Dispatch Coding, and Location on the Outcomes of Police Encounters with Individuals with Mental Illnesses in Chicago

Amy C Watson 1,✉, Linda K Owens 2, Jennifer Wood 3, Michael T Compton 4
PMCID: PMC8507917  PMID: 34659453

Abstract

The Crisis Intervention Team (CIT) model has been implemented in over 3,000 communities across the USA. Research to date has shown beneficial results in terms of officers’ knowledge, attitudes, self-efficacy, stigma, and force preferences. This study aimed to broaden the lens on the implementation context of CIT to examine whether factors in the environment and response process affect how calls are resolved. This study focused on several factors—CIT response, call location, and upstream decisions to pre-identify calls as mental health-related—that may impact call outcomes. Our findings suggest that CIT response, dispatch coding, and the places where calls originate play a role in shaping outcomes. More research is needed to unpack the effects of this wider CIT implementation environment.

Introduction

Law enforcement officers have long played a role in responding to mental health crisis (Bittner, 1967). Over the past 50 years, they have increasingly become the de facto mental health crisis responders, and accordingly, the criminal legal system has developed into a de facto mental health system. This half century long process of shifting responsibility from the health and social services sectors to the law enforcement sector has undermined the nation’s commitment to community safety and wellbeing. For people with mental illnesses involved in police encounters, unmet physical and behavioural health needs (Lorvick et al., 2018), and experiences of stigma (Watson and Angell, 2013) can be exacerbated through deeper legal system involvement. Additionally, people with serious mental illnesses (SMIs) are over overrepresented among police involved shooting deaths (Fuller et al., 2015; Lowery et al., 2015).

Recognizing the need to reduce the harms of police and legal system involvement for people with SMIs, leaders in government and policing have led efforts to improve law enforcement’s capabilities to provide effective responses and redirect individuals to the mental health system whenever possible. In the USA, the most well-known strategy to do this is the Crisis Intervention Team (CIT) model. CIT was first developed in Memphis, Tennessee, over 30 years ago and has been implemented in more than 3,000 communities, including several outside of the USA (University of Memphis, n.d.).

The research to date provides strong support for CIT training’s effectiveness for improving officers’ knowledge, attitudes, self-efficacy, stigma, and use of force preferences (Watson et al., 2017). Several studies also indicate that CIT implementation increases linkages to mental health services and reduces use of force with more agitated and resistant call subjects (Morabito et al., 2012; Compton et al., 2014b; Kubiak et al., 2017). Notwithstanding these positive outcomes, there remains a need to further examine the effects of CIT training and other factors, some that may be impacted by CIT program implementation, on the decisions that officers make in resolving mental health-related calls. Therefore, our study set out to examine whether contextual features of the decision-making environment, such as the processes at the point of call-taking and dispatch to pre-identify calls as mental health-related, may play a role in affecting call outcomes (i.e., arrest, linkage to services, and informal resolutions).

This study takes advantage of a novel dataset that allows for an analysis of the relationships between call outcomes, call locations (e.g., private home, street, business, and hospital), pre-identification of mental health-related calls, and CIT officer response, all nested within police districts. This study represents and arm of a larger study aimed at examining the impact of CIT on call outcomes across all 22 Chicago Police districts. Our findings suggest that there are aspects of the wider CIT implementation environment that play a role in shaping police call-handling decisions, including the nature of the places where police encounters occur and prior decisions at the call-taking stage that frame calls as mental-health related.

Police contacts with individuals with mental illnesses

While definitive data on the proportion of police contacts with the public that are related to mental illness is lacking, there is evidence to suggest that individuals with SMI are at high risk of police contact and subsequent processing into and through the criminal legal system. A recent review of studies on police encounters with people with SMIs estimated that in the USA, 6% of police calls for service involve people with mental illnesses, 25% of people with mental illnesses had been arrested at some point in their lifetime, and 29% had police involved in their pathway to care (Livingston, 2016). Early work by Teplin (1984) indicated that in police encounters, the presence of mental illness increased the likelihood a suspect would be arrested. Research conducted several decades later in different cities and using varied methods found that the presence of mental illness decreased the likelihood of arrest (Engel and Silver, 2001). Regardless of whether mental illness increases or decreases the likelihood of arrest in the immediate encounter, people with SMI are over-represented in jail and prison populations, with prevalence estimates of 14.5% for male and 31% for female jail inmates (Steadman et al., 2009), compared with prevalence rates of 3.3% and 5.7% for the general population aged 18 years and older (Substance Abuse and Mental Health Services Administration, 2018).

Persons with SMI are also over-represented among individuals shot and killed by police officers, with estimates that at least one in four involve a person with SMI (Fuller et al., 2015; Lowrey, 2015). An analysis that used the Washington Post database on fatal police shootings as a starting point suggests that fatal shootings involving individuals with mental illnesses were more likely to have been initiated by calls to 911 by family or friends and less likely to have been police initiated than fatal shootings of individuals without mental illnesses (Frankham, 2018). Additionally, individuals with mental illnesses were both less likely to have a projectile weapon and less likely to have attacked police than those without mental illnesses, suggesting other factors played a role in the use of deadly force.

Thus, individuals with mental illnesses are at risk for contact with police, and often the outcomes of these contacts are not ideal, and sometimes tragic. While some persons with mental illnesses may be appropriately arrested by criminal legal system standards, there is a general consensus that jails are not ideal settings in which to provide acute psychiatric treatment, and many individuals may be better served by mental health service responses in their community, assuming such services are available. Thus, communities have sought strategies to address this issue and improve police responses while diverting individuals with mental illnesses away from the criminal legal system when possible (Watson et al., 2019).

The CIT model

The CIT model emerged in the aftermath of a shooting of a black man experiencing a mental health crisis by a white police officer in Memphis, Tennessee, in 1987 (CIT Memphis, n.d.). Following the tragedy, the Mayor of Memphis enlisted representatives from the local chapter of the National Alliance on Mental Illness (NAMI), two universities, community mental health service providers, hospital administrators, local clergy, and the Memphis Police Department to develop a strategy to better respond to people experiencing mental health crises. While CIT is often thought of as a police training programme, the approach the group created includes multiple elements, which are detailed in a 2007 paper (Dupont et al., 2007) and a manual published by CIT International (Usher et al., 2019). Here, we briefly describe three of the main components: (1) collaboration, (2) designated psychiatric emergency drop off, and (3) CIT training.

Collaboration

A foundational component of the model is the collaboration between police, mental health services, advocates, and people with the lived experience of mental illness. In fact, this is what the ‘Team’ in CIT refers to. This collaboration is key to delivering CIT training as well as to the ongoing operation of the CIT programme. CIT programmes typically have a CIT Advisory Council made up of representatives from police, mental health, advocacy, and peer organizations that meets regularly, first to develop the local programme, and then to monitor ongoing operations.

Designated psychiatric emergency drop-off

CIT programmes identify a designated facility (or facilities) where officers can bring a person in need of emergency psychiatric assessment. Memorandums of Understanding support an efficient process where officer transports are given priority so that officers can get back to their patrol duties quickly with a guarantee that the person will be assessed. CIT programmes in communities that do not have receiving facilities work to develop options to address this need.

CIT training is a 40-h training week in which participants receive didactic, experiential, and role-play content provided by police training staff and content experts from the community. The didactic content includes signs and symptoms of mental illnesses, developmental disabilities, and co-occurring substance-use disorders; suicide risk assessment and intervention; mental health crisis response policy and procedures; de-escalation skills; and community resources. Trainees participate in experiential exercises, have opportunities to hear from people with lived experience of mental illnesses and their family members, and go on visits to local mental health provider agencies where they interact with staff and clients. Finally, significant time is dedicated to role-play exercises where participants practice de-escalation skills. According to the model, CIT training is for officers who volunteer to take on this role and are vetted as appropriate for it. While they remain patrol officers, when mental health-related calls occur, they take on a specialist role as a CIT officer. Fully implemented CIT programmes provide training to emergency communications personnel (911 call takers and dispatchers) so they can identify potential mental health-related calls and dispatch CIT officers to them.

According to the University of Memphis CIT Center website, there are 2,645 agency and 351 regional CIT programmes in the USA (University of Memphis, n.d.). This is within the context of approximately 18,000 law enforcement agencies (Banks et al., 2016). Given the lack of centralization of law enforcement information in the USA, it is currently not possible to determine the true number of agencies implementing CIT or the number of police officers that are CIT trained. Additionally, there is significant variation in CIT implementation; thus, it is not clear the extent to which agencies that have CIT programmes are fully implementing with fidelity to the model’s core elements.

Research on the CIT model

While to date, there has not been a randomized controlled trial of CIT, there is a growing body of evidence supporting its effectiveness. There is strong evidence that supports CIT’s effectiveness for improving officer-level outcomes in terms of knowledge, attitudes, self-efficacy, force preferences, de-escalation skills, and referral decisions (Compton et al., 2014a; Kubiak et al., 2017). Numerous studies also indicate that CIT is effective for increasing linkages to psychiatric care (transports and referrals) to resolve mental health-related encounters (Teller et al., 2006; Watson et al., 2010; Compton et al., 2014b; Kubiak et al., 2017). These effects are strongest when officers self-select into the specialist role (Compton et al., 2017) and in areas with greater availability of mental health services (Watson et al., 2011). The evidence of CIT’s impact on safety outcomes is more limited. In our prior study of CIT in Chicago, we found that CIT officers used less force with more resistant subjects than their non-CIT counterparts (Morabito et al., 2012). Likewise, the evidence of CIT’s impact on injuries is very limited given that injuries are rare. Findings related to arrest outcomes are mixed, with some studies indicating reductions of arrests of persons with mental illnesses and others finding no effect (Teller et al., 2006; Kerr et al., 2010; Compton et al., 2014b).

Qualitative research on CIT supports that CIT-trained officers take a more nuanced and compassionate approach to responding to mental health crisis situations than their non-CIT peers (Hanafi et al., 2008; Canada et al., 2012). Additionally, studies utilizing focus groups and interviews with officers and key informants suggest that CIT implementation supports better understanding and collaboration across police and mental health agencies (Skubby et al., 2013; Kubiak et al., 2017).

Emergency communications and call coding

Neusteter et al. (2019) point out the important role of the 911 system in policing. Indeed, emergency communications agencies are key partners in fully implemented CIT programmes. According to the model, call takers are trained to identify mental health-related calls and dispatch them to CIT-trained officers with information about the mental health nature of the call (Usher et al., 2019). While not focused on response to mental health-related calls, several studies demonstrated that interactions between callers and call-takers shape the definition of the situation conveyed to officers (Whalen and Zimmerman, 1990; Garcia and Parmer, 1999; Gillooly, 2020). Furthermore, information provided by dispatchers to officers can shape their judgements about the legitimacy of calls (Moskos, 2007). In a recent article dissecting the 911 call related to the arrest by the Cambridge Police Department (MA) of Henry Louis Gates, Jr in 2009, a Black Harvard Professor returning to his home following a trip abroad, Gillooly (2020) illustrates how information relayed to officers by dispatchers may prime officers and contribute (in this case negatively), to call outcomes. Lum et al. (2020) speculate that ‘it may be the case that the amount, level, quality, and type of information the dispatcher provides to officers about the location or individual the officer is about to encounter may influence the eventual outcome (p. 17).’

Location of calls

Mental health-related police encounters occur in people’s homes, government facilities, public transportation, businesses, parks, and on the street (Green, 1997; Wood et al., 2017). Sociologist Bittner (1967), one of the first researchers to examine police interactions with people with mental illnesses, noted the influence of situational factors, including the setting (public or private) on how officers resolved encounters, finding that officers were more likely to take formal action when encounters occurred in public settings. The nature of the encounter, stakeholders present, safety, and legal factors vary based on the space in which police interactions take place.

Research questions

The primary goal of this study was to examine the impact of CIT response on outcomes of mental health-related police encounters. Based on prior CIT research, we expect that CIT response will increase linkages to mental health care, and potentially reduce arrests and informal/resolve on scene outcomes (Watson et al., 2017). We also sought to consider elements of the broader CIT implementation context. Emergency communicators are important stakeholders and partners in the CIT programme model, and the emerging literature suggests they may play a role in influencing officer response (Lum et al., 2020). We will examine the impact of dispatch coding that pre-identifies an encounter as mental health related on outcomes. Finally, the setting in which an encounter occurs is a contextual factor that may shape officer decision making important ways (Bittner, 1967). Thus, we include encounter location type in our models. Hence, our research questions were:

  1. Is CIT response associated with the outcomes of mental health-related police encounters?

  2. Are dispatch codes pre-identifying a call as mental health related associated with outcomes of mental health-related encounters?

  3. Is the encounter location associated with outcomes of mental health-related police encounters?

Methods

The setting

The setting for this study consisted of all 22 Chicago police districts. The city of Chicago covers an area of 228.5 square miles (367.7 km2) with a population of approximately 2.9 million residents. It is divided into 22 police districts, each containing 9–15 beats.

At the time of the study, Chicago’s CIT programme had been operating for almost a decade. Approximately 18% of the 13,000 Chicago Police Department (CPD) sworn personnel were CIT trained. The Department had Memorandums of Understanding with 13 hospital emergency departments (EDs) spread throughout the city that served as designated drop-off sites for officers to bring individuals needing emergency psychiatric assessment, either voluntarily or involuntarily. One of the EDs was in a free-standing psychiatric hospital, the remaining were located in general medical EDs.

In the years immediately preceding this research, the programme’s infrastructure had been reduced as new Superintendents focused on different priorities. The CIT Advisory Committee, comprised of representatives from partnering city, mental health, and advocacy agencies, was no longer active. The central unit at CPD responsible for the programme—working with NAMI Chicago and other partners to provide CIT trainings and coordinate the programme Citywide—had shrunk from being a 12-member unit to a 5-member unit. Also, at the time of the study, the current emergency communication centre staff did not use a specific script to identify mental health-related calls, and had not been trained to identify potential mental health-related calls well-suited for CIT-trained officers, nor were they consistently able to identify officers who were CIT trained. Thus, both CIT-trained and non-CIT officers were regularly responding to mental health-related calls.

In the larger environment of the CIT programme, mental health funding at the state level had been cut by a third over three consecutive fiscal years beginning in 2009. At the same time, the number of visits to EDs for people in crisis increased by almost one-fifth (Thresholds, 2013). In 2012, half of the 12 City of Chicago-operated mental health clinics were closed, and from 2015 to 2017, community services were reduced due to the failure of the state to resolve a budget impasse. Late or nonpayment of State bills forced agencies serving vulnerable populations to limit their services, or for some, close altogether (Mendoza, 2018).

The data

Data reported here are from a large, 5-year multi-method study of CPD’s CIT programme and responses to situations involving persons with mental illnesses. The overall study aimed to examine immediate and longer-term outcomes of mental health-related encounters in the context of the CIT programme. All study activities were approved by the Principal Investigator’s University’s Institutional Review Board. Here, we utilize data from officers’ reports on 428 mental health-related calls. A previous paper provided descriptive statistics on the nature of these calls (Watson and Wood, 2017). This report examines the impact of CIT (versus non-CIT) response, pre-identification as a potential mental health call, officer race and gender, and call location type on immediate call outcomes.

Data were collected on police encounters with persons perceived to have a mental illness and/or experiencing a mental health crisis, based on the dispatch code, responding officers’ observations, or information gathered at the scene. From July 2013 to September 2016, members of the research team visited all 22 Chicago police districts and the public transit unit. Over a 6-week period in each district, team members addressed roll call briefings on each watch six times to provide information on the study and to request that officers participate if they respond to a mental health-related call during their watch. Team members met officers as they came in from their watch to follow-up and complete a researcher-administered survey if they had responded to a mental health-related call with an adult. The officer survey incorporated items from surveys developed and tested in prior CIT studies conducted by two of the authors (Watson et al., 2011; Compton et al., 2014b) and included questions about the nature of the call, location, subject behaviour and risks noted, information gathered, strategies used, call disposition, and officer and subject demographics. The survey was reviewed and revised based on feedback from law enforcement and mental health content experts familiar with policing and mental health services in Chicago, and piloted with several officers before being used to collect study data to ensure questions and response options were relevant and clear.

Officers completed the survey for each mental health call they identified, but only completed the officer demographic questions once. Because an individual officer could respond to multiple calls, the 428 calls in the data do not correspond to 428 officers but are spread across 294 officers. The majority of officers (84.1% overall, 83.6% of CIT officers, 85.7% of non-CIT officers) reported on a single call—97% of officers reported on three or fewer calls. Given this distribution of encounters across officers, we are not able to account for nesting within officers (where it occurs) in our analysis.

For the analysis reported here, the unit of analysis is the encounter, rather than officer, with CIT response treated as a characteristic of the encounter. At the time of data collection, dispatch personnel had not been trained to identify potential mental health-related calls, nor did they know which officers were CIT trained. Thus, there was no systematic assignment of pre-identified mental health-related calls to CIT-trained officers. While clearly a weakness of the CIT programme’s implementation, this allowed us to make comparisons between calls that received a CIT response and those that did not, without the confounding effect of more obvious mental health-related calls for service being dispatched intentionally to CIT-trained officers. Additionally, in some cases, multiple officers responded to a call. If one was a CIT-trained officer acting in their CIT role, that is the officer we interviewed.

The 428 calls are distributed across the 22 districts and the transit unit, with a range of 2–33 and a mean of 18.6 (SD 7.8) call per district/unit. Our primary independent variable is CIT response. Additional independent variables were selected for inclusion in the model based on theoretical importance and bivariate analyses. Variables significantly associated with call outcomes were included in the model. Call subject demographics were examined but were not associated with call outcomes.

Dependent variable

The dependent variable is the outcome of the call, or how the officer resolved the encounter. The interview instrument included a series of questions asking the officer about the actions he or she took to resolve the call. Options included arrest (26, 6.1%), transport to an ED (184, 43.0%), refer to services (32, 7.5%), or resolve on the scene/no formal action (186, 43.5%). Of the 26 arrests, 4 were for outstanding warrants, the remaining 22 were for charges ranging from loitering and public drinking to aggravated assault. To simplify the analysis, the outcomes were collapsed into three categories—arrest, linkage to services (transport to an/ED or refer to services), and informal resolution (resolve on scene/no formal action).

Independent variables

CIT response

This variable is coded 1 if the primary responding officer had completed CIT training (134, 31.3%) and 0 (294, 68.7%) if he or she had not.

Call location

The original variable measuring the location of the call included seven categories that were collapsed into four categories: private home or housing (147, 34.4%); street or park (133, 31.1%); business, government facility, police station, or transit station (128, 29.9%); and hospital or mental health provider (20, 4.7%).

Pre-identified mental health call

Calls that were preidentified and dispatched as ‘Mental Disturbance’ or ‘Suicide Threat/Attempt’ were coded as 1 (109, 25.5%); all other calls were coded as 0 (320, 74.5%).

Officer race and gender

Among the 418 calls with valid officer race data, 306 (73.4%) were White or Caucasian, 81 (19.4%) African American, and 30 (7.2%) Asian. For analysis purposes, this variable was coded as White = 0, nonwhite = 1. The variable measuring gender is coded 1 if the officer is male and 0 if female. A total of 421 officers had valid gender data—351 (83.4%) were male and 70 (16.6%) were female.

Analytic strategy

Because of the nested nature of the data—calls within police districts—this study employs hierarchical linear modelling (HLM), which takes into account the random effect of police district. For the models presented here, there are no level-2 variables in the analysis—that is, no district level measures are included. The analysis includes only individual level predictors and a random intercept. The random intercept captures the variation at the district level at average levels of all independent variables. Separate models were run for each outcome: arrest, linkage to services, and resolve on scene/no formal action.

Null models

To assess the appropriateness of HLM analyses, we ran empty logit models using STATA’s melogit procedure. These three models included only the dichotomous outcomes and a random intercept. The results are presented in Table 1. The likelihood ratio (LR) test compares the goodness of fit of a hierarchical model to a regular logit regression. If the difference is significant, it indicates there is variation in the outcome at the district level. The ICC shows the percent of the variance attributable to the police district. Two of the LR tests are statistically significant and the other is just over the 0.05 convention. The two outcomes with significant LR tests have low ICCs. So, while there is some variation in outcome at the district level, the district does not account for much of the total variance. The ICC is higher in the arrest outcome (0.140), but the LR test is not significant. However, with only 26 arrests in the dataset, it is possible the model is underpowered.

Table 1:

Random intercept null models

Constant District level variance LR test ICC
Arrest −2.94 0.536 2.41 (P = 0.06) 0.140
Linkage to services 0.037 0.127 3.44 (P = 0.032) 0.037
Informal resolution −0.292 0.120 3.35 (P = 0.034) 0.035
Open in a new tab

While the analysis suggests a multilevel model may not be necessary, Bliese et al. (2018) make a convincing case for using a multilevel model approach even when ICC values are low and statistically non-significant. Police culture is heterogeneous and may vary across district or operational environment (Paoline, 2013). Furthermore, areas included in each district vary in terms of community demographics, geography, extent of crime, and availability of resources. Failing to account for nonindependence in the data could lead to increased risk of Type I error. Thus, the analyses presented here are all conducted within a multi-level framework.

For all three outcome variables, we ran a multilevel logistic regression with call location, mental health-related call designation, CIT response, officer race, and officer gender as predictors and district-level random intercepts. Because race and gender were not significant for any models, they are not reported.

Results

Arrest

The only significant predictor of the call resulting in arrest rather than other outcomes is whether or not the call was pre-identified as a mental health call (Table 2). The odds of the call subject being arrested are much lower if the call was preidentified as a mental health call than if it was not. The ICC indicates that 14% of the variance is due to district effects. The odds ratio of hospital/mental health provider location is 1 because no calls that resulted in arrest originated from these locations.

Table 2:

Predictors of arrest

Odds ratio Std. error z P > |z|
CIT 0.471 0.280 −1.27 0.205
Location (private home/public housing is reference category)
 Street/park 0.551 0.379 −0.87 0.386
 Business/gov/police/transit 2.20 1.19 1.47 0.142
 Healthcare provider 1 na na na
Mental health call 0.088 0.092 −2.32 0.02
(Constant) 0.067 0.023 −7.91 0.000
Random effects parameters
 District 0.533 0.561
 ICC 0.139 0.126
Open in a new tab

LR test versus logistic model: chibar2(01) = 1.84 Prob ≥ chibar2 = 0.088.

Wald χ2(4) = 12.65, P = 0.013.

Linkage to services

CIT response increases the odds that a call will result in linkage to mental health services (OR = 1.66, Table 3). Additionally, compared with calls that originate in a private home/housing, calls that originate in a street or park—or in a business, government office, police station, or on public transit—are less likely to result in linkage to services. While the odds of linkage to service are much higher for calls originating from hospitals and mental health providers, this effect is not statistically significant. Calls identified as mental health-related calls at the outset were more likely to result in linkage to service than those that were not (OR = 2.83). The ICC is 0.044, indicating that little of the remaining variance is due to random district effects.

Table 3:

Predictors of linkage to services

Odds ratio Std. error z P > |z|
CIT 1.66 0.40 2.11 0.034
Location (private home/public housing is reference category)
 Street/park 0.380 0.104 −3.52 0.00
 Business/gov/police/transit 0.409 0.115 −3.17 0.002
 Healthcare provider 2.749 1.869 1.49 0.137
Mental Health call 2.830 0.761 3.87 0.000
(Constant) 0.847 0.282 0.7 0.487
Random effects parameters
 District 0.150 0.134
 ICC 0.044 0.037
Open in a new tab

LR test versus logistic model: chibar2(01) = 2.45 Prob ≥ chibar2 = 0.059.

Wald χ2(5) = 48.9, P = 0.000.

Informal resolution

Table 4 provides results from the analysis of informal resolution (resolve on scene/no formal action) as a call outcome. Location of call and pre-identified mental health call are both statistically significant, but CIT response is not. Calls that originate from a street or park—or from a business, government building, police station, or transit—are more likely to have an informal resolution than calls that originate from a home. Calls that originate from a hospital or mental health provider are less likely to be resolved informally, but this result is not significant. Calls pre-identified as mental health calls are less likely to be resolved informally than those that are not (OR = 0.471).

Table 4:

Predictors of informal resolution

Odds ratio Std. error z P > |z|
CIT 0.689 0.163 −1.58 0.115
Location (private home/public housing is reference category)
 Street/park 2.89 0.781 3.92 0.000
 Business/gov/police/transit 1.86 0.518 2.25 0.024
 Healthcare provider 0.469 0.314 −1.13 0.258
MH call 0.476 0.127 −2.79 0.005
(Constant) 0.604 0.143 −2.13 0.033
Random effects parameters
 District 0.086 0.098
 ICC 0.025 0.028
Open in a new tab

LR test versus logistic model: chibar2(01) = 1.29, Prob ≥ chibar2 = 0.128.

Wald χ2(5) = 37.38, P = 0.000.

Summary of findings

Research Question 1

The primary variable of interest—CIT response—has a significant effect on a call resulting in linkage to services. A call that receives a CIT response is 1.7 times more likely to result in linkage to services than a call that does not receive a CIT response. CIT response did not have a significant effect on the outcomes of arrest or informal resolution.

Research Question 2

Whether or not the call was pre-identified as a mental health call is the only variable to affect all three outcomes. Pre-identified mental health calls are almost three times (OR 2.83) more likely to result in linkage to services and less likely to result in arrest (OR 0.09) or informal resolution (OR 0.48).

Research Question 3

Call location affects the likelihood of linkage to services and informal resolution, with calls from streets or parks—or businesses, government buildings, police stations, or public transit—being less likely to result in linkage to services and more likely to lead to informal resolution compared with calls occurring in homes. Call location did not affect the likelihood of arrest.

Discussion

Consistent with prior research on CIT, including our prior work in Chicago that was not able to account for potential district effects (Watson et al., 2011), our findings suggest that CIT response increases the use of transports and other linkages to mental health services to resolve mental health-related calls. For the most part, this linkage was a transport to an ED that may or may not have resulted in effective connections to needed ongoing community mental health care. There is growing acknowledgement in the field and literature that EDs are not ideal places for people experiencing mental health crisis, nor are they necessarily effective for linking people to needed care (Broadbent et al., 2014; Wise-Harris et al., 2017). A theme that emerged in qualitative interviews with officers conducted as part of the larger study and described elsewhere (Watson and Wood, 2017; Wood et al., 2020) indicated their frustrations with EDs as their primary option. Officers suggested a need for alternatives to the ED and consistent and assured follow-up by mental health providers. Thus, while our data suggest that CIT response can increase linkages to care, research that examines the impact of the availability of linkage options on how calls are resolved and longer-term outcomes for persons with mental illnesses is needed. Since data collection for the study ended, several crisis stabilization centres have opened and pilot follow-up and linkage programmes have been implemented in several areas of the city.

The finding that the location where calls take place matters for two of our outcomes (and may have been for arrest if that outcome had been adequately powered) connects to a broader literature on place-based criminology that examines various aspects of where crime and other social problems cluster and what it is about places that shape the behaviours within them (Weisburd et al., 2009; Hibdon and Groff, 2014). Recently, criminologists have broadened their inquiries to examine potential overlaps between crime, health disparities, and disadvantage (Weisburd and White, 2019). Lum (2011) also argues that officer decision-making should be understood from a place-based perspective. Officers may respond to ‘environmental cues’ (p. 57) in a given location, such as demographic make-up, or socioeconomic status, or known violence in a place. ‘Such cues’, she writes, ‘provide officers and others with a quick understanding of the types of people within a location, which turn may condition their responses to incidents at those places’ (p. 657). More needs to be understood about the place-based dimensions of police decision-making in specific relation to incidents with apparent mental health components. Certain places like parks and transportation hubs may be drivers of particular types of calls for police service, including mental health-related calls, because they may function as ‘hot spots of vulnerability’ (Wood and Beierschmitt, 2014) for ecological reasons that need to be better understood. Future research should explore the precise environmental cues picked up (and acted upon) by officers in the locations of mental health-related incidents.

Relatedly, more empirical studies are needed to better understand what people are requesting or demanding when they call the police for service. In public places like streets or parks, it could be that people simply want individuals causing a perceived ‘threat’ to be ‘moved along’ by police. Encounters occurring on streets and in parks may be more likely to be ‘on view’ rather than called into 911. In this case, there may not be a prior decision maker or request for a specific outcome, providing the officer with more discretion in resolving the encounter. These are more ‘gray zone areas’ (Wood et al., 2017) of decision making by officers where informal resolutions may be deemed more appropriate.

In private places, like homes or clinics, people calling for help likely have more familiarity with the health vulnerabilities of the subject in question. A mental health provider in a mental health setting will only call police when there is a serious crisis, and they identify mental illness symptomology. Similarly, a family member may call 911 to request police assistance with a loved one who is acutely symptomatic and disruptive. They may indicate to the call taker that their loved one is experiencing a mental health crisis and request that the officer transport the person to the hospital.

When callers in private places ask for help, it is therefore plausible that they provide specific mental health-related information when they speak to a call-taker, which then in turn may influence how calls are dispatched and what officers know about a person, their vulnerability, and their family or social situation when they arrive to a scene. Whether or not a call is pre-identified as a mental health call may be related to the ability of the call taker to recognize a potential mental health issue and ask specific questions. While the 911 Center personnel had not received CIT-specific training at the time of data collection in Chicago (they subsequently have received this training), this component of CIT model implementation could have a significant impact on improving call outcomes. Officers who receive this information enroute to a call may be sensitized and approach the situation with a different mindset and strategy.

A tragic example of 911 call takers failing to recognize a mental health crisis occurred on December 26, 2015. Nineteen-year-old Quintonio LeGrier called 911 several times indicating someone was trying to kill him but was hung up on by two call takers when he refused to give his full name. His father called and indicated his son was armed with a baseball bat and trying to break into his bedroom. The call taker asked about drug use, but not about mental health issues. When officers arrived, they fatally shot LeGrier and his 55-year-old neighbour Bettie Jones. It was later learned the LeGrier, a college sophomore home on his winter break, had been behaving erratically in the months prior and had come to the attention of campus police on several occasions. According to newspaper accounts, in one instance, he was hospitalized for a psychiatric evaluation. It is impossible to know whether the outcome would have been different if the call had been dispatched as a mental disturbance call (Chicago’s dispatch code). However, it is critical that researchers pay more attention to these initial decision-making processes involved in enlisting the police in the first place, starting with citizen call-for-service requests/demands and how those requests/demands are articulated to call-takers, including the information provided and how it is ultimately passed on to responding officers. The roles of 911 callers, call-takers, and dispatchers in shaping the larger process of police response merits much greater attention, as Gillooly (2020) recently argued. Overall, researchers should strive toward a better understanding of factors shaping this longer ‘decision pathway’ (Lum, 2011) beginning with a focal event or incident and leading up to an officer’s choice of call resolution.

Limitations

The study findings should be considered within the context of several limitations. First, as discussed, the implementation of CIT in Chicago was not robust outside of the provision of CIT training during the data collection period. Thus, while we still found a significant impact of CIT response on mental health linkage outcomes, the study cannot claim to be a test of a high-fidelity implementation of the CIT model. Second, given that we could not force officers to participate and report every mental health-related call they responded to during the study period, our sample of mental health calls cannot be assumed to be representative of all mental health calls that occurred. We considered utilizing CPD administrative call data to examine CIT effectiveness. However, at the time of data collection, mental health-related calls were not consistently coded as such or identifiable in the CPD data systems. Significant improvements have recently been made that will allow for future studies utilizing CPD call data. Third, with only 26 calls resolved with arrest in the sample, we were underpowered to fully examine the impact of CIT response and other factors on this outcome. While not statistically significant, the larger amount of variance attributable to police district would be of interest to explore in future studies. Despite these limitations, our findings offer additional evidence of the effectiveness of CIT for increasing officers’ attempts to link people to care and suggest that training of 911 personnel could further support the effectiveness of CIT implementation.

Conclusion

This study identified factors that may shape the outcomes of mental health-related encounters with police. Whether or not calls were identified during the call-taking and dispatch process as mental health related had a significant effect on how calls were resolved. In particular, officers were more likely to link people to services, including EDs, if they knew in advance the calls had a mental health component. This finding supplements the existing knowledge that CIT-trained officers are more likely than non-CIT officers to effect interventions that address people’s mental health needs. Call location also affects outcome resolution. Calls from streets or parks, businesses, government buildings, police stations, and public transit are less likely to result in linkage to services. Conversely, calls originating from homes are more likely to result in linkage to services. Taken together, these findings on the place-based origins of calls are unsurprising theoretically, as one assumes that situational demands of callers vary by place, circumstance, and knowledge of an individual’s vulnerabilities. That being said, further research should aim to unpack the upstream chain of decision-making processes associated with enlisting the help of the police, including what is demanded of officers and how such demands are articulated to call-takers and dispatchers, and ultimately to responding officers.

For encounters that officers approach ‘on view’ in the course of routine patrol, the dynamics of their decision-making are still not fully understood. However, a growing body of qualitative research suggests that officers yearn for more mental health linkage options (greater access to behavioural health and social services) beyond transport to EDs, especially when subjects display chronic vulnerabilities, but not necessarily acute symptomology. In conclusion, it is important to advance a fuller account of decision-making in the wider environment of CIT implementation, starting with the decisions of callers, call-takers, dispatchers, and police. Understanding this wider context of police interventions will foster more precision in what we know about the impact of CIT programmes.

Disclosure. During the funding period for this project, A.C.W. provided consultation services to the Chicago Police Department on projects unrelated to the study reported on in this article.

 

A.C.W., Ph.D., is a Professor at the Helen Bader School of Social Welfare, University of Wisconsin–Milwaukee. Her research focuses police involvement in responding to mental health crisis and strategies to reduce police and criminal legal system involvement for people living with mental illnesses.

L.K.O., Ph.D., is currently a Science Writer for Carle Foundation Hospital, Champaign, IL, USA. In her previous position, she was the Assistant Director for Sampling and Analysis at the University of Illinois Survey Research Laboratory, where she provided sampling and statistical support to investigators in a wide range of disciplines.

J.W., Ph.D., is a Professor in the Department of Criminal Justice at Temple University. Her research focuses on the intersections between policing and public health, including the roles of officers in intervening with people experiencing health vulnerabilities such as mental illnesses.

M.T.C., M.D., M.P.H., is a Professor in the Department of Psychiatry, Division of Behavioral Health Services and Policy Research, Vagelos College of Physicians and Surgeons. He is also a Research Psychiatrist at the New York State Psychiatric Institute.

This work was supported by the National Institute of Mental Health R01MH096744. The contents of this article are solely the responsibility of the authors and do not necessarily represent the official views of the National Institutes of Health or the National Institute of Mental Health. This data were provided by and belongs to the Chicago Police Department. Any further use of this data must be approved by the Chicago Police Department. Points of view or opinions contained within this article are those of the author and do not necessarily represent the official position or policies of the Chicago Police Department.

References

  1. Banks D., Hendrix J., Hickman M., Kyckelhahn T. (2016). National sources of law enforcement employment data. Bureau of Justice Statistics. NCJ 249681.
  2. Bliese P., Maltarich M., Hendricks J. (2018). ‘ Back to Basics with Mixed-Effects Models: Nine Take-Away Points.’ Journal of Business and Psychology 33(1): 1–23. [Google Scholar]
  3. Broadbent M., Moxham L., Dwyer T. (2014). ‘ Implications of the Emergency Department Triage Environment on Triage Practice for Clients with a Mental Illness at Triage in an Australian Context.’ Australasian Emergency Nursing Journal 17(1): 23–29. [DOI] [PubMed] [Google Scholar]
  4. Canada K., Angell B., Watson A. (2012). ‘ Intervening at the Entry Point: Differences in How CIT Trained and Non-CIT Trained Officers Describe Responding to Mental Health-Related Calls.’ Community Mental Health Journal 48(6): 746–755. [DOI] [PMC free article] [PubMed] [Google Scholar]
  5. Compton M. T., Bakeman R., Broussard B., et al. (2014a). ‘ The Police-Based Crisis Intervention Team (CIT) Model: I. Effects on Officers’ Knowledge, Attitudes, and Skills.’ Psychiatric Services 65(4): 517–522. PMID: 24382628 [DOI] [PMC free article] [PubMed] [Google Scholar]
  6. Compton M., Bakeman R., Broussard B., et al. (2014b). ‘ The Police-Based Crisis Intervention Team (CIT) Model: II. Effects on Level of Force and Resolution, Referral, and Arrest.’ Psychiatric Services 65(4): 523–529. [DOI] [PubMed] [Google Scholar]
  7. Compton M. T., Bakeman R., Broussard B., D’Orio B., Watson A. C. (2017). ‘ Police Officers’ Volunteering for (Rather than Being Assigned to) Crisis Intervention Team (CIT) Training: Evidence for a Beneficial Self-Selection Effect.’ Behavioral Sciences & the Law 35(5–6): 470–479. [DOI] [PMC free article] [PubMed] [Google Scholar]
  8. Dupont R., Cochran S., Pillsbury S. (2007). Crisis Intervention Team Core Elements. Memphis, TN: University of Memphis. [Google Scholar]
  9. Engel R., Silver E. (2001). ‘ Policing Mentally Disordered Suspects: A Reexamination of the Criminalization Hypothesis.’ Criminology (Beverly Hills ) 39(2): 225–252. [Google Scholar]
  10. Frankham E. (2018). ‘ Mental Illness Affects Police Fatal Shootings.’ Contexts 17(2): 70–72. [Google Scholar]
  11. Fuller D. A., Lamb H. R., Biasotti M., Snook J. (2015). Overlooked and Undercounted. Treatment Advocacy Center, Arlington, VA. http://www.tacreports.org/storage/documents/overlooked-in-the-undercounted.pdf.
  12. Garcia A. C., Parmer P. A. (1999). ‘ Misplaced Mistrust: The Collaborative Construction of Doubt in 911.’ Emergency Calls. Symbolic Interaction 22(4): 297–324. [Google Scholar]
  13. Gillooly J. W. (2020). ‘How 911 Callers and Call-Takers Impact Police Encounters with the Public: The Case of the Henry Louis Gates Jr. Arrest.’ Criminology and Public Policy (doi: 10.1111/1745-9133.12508).
  14. Green T. (1997). ‘ Police as Frontline Mental Health Workers: The Decision to Arrest or Refer to Mental Health Agencies.’ International Journal of Law and Psychiatry 20(4): 469–486. [DOI] [PubMed] [Google Scholar]
  15. Hanafi S., Bahora M., Demir B., Compton M. (2008). ‘ Incorporating Crisis Intervention Team (CIT) Knowledge and Skills into the Daily Work of Police Officers: A Focus Group Study.’ Community Mental Health Journal 44(6): 427–432. [DOI] [PubMed] [Google Scholar]
  16. Hibdon J., Groff E. R. (2014). ‘ What You Find Depends on Where You Look: Using Emergency Medical Services Call Data to Target Illicit Drug Use Hot Spots.’ Journal of Contemporary Criminal Justice 30(2): 169–185. [Google Scholar]
  17. Kerr A. N., Morabito M. S., Watson A. C. (2010). ‘ Police Encounters, Mental Illness and Injury: An Exploratory Study.’ Journal of Police Crisis Negotiations 10(1–2): 116–132. [DOI] [PMC free article] [PubMed] [Google Scholar]
  18. Kubiak S., Comartin E., Milanovic E., et al. (2017). ‘ Countywide Implementation of Crisis Intervention Teams: Multiple Methods, Measures and Sustained Outcomes.’ Behavioral Sciences & the Law 35(5–6): 456–469. [DOI] [PubMed] [Google Scholar]
  19. Livingston J. D. (2016). ‘ Contact between Police and People with Mental Disorders: A Review of Rates.’ Psychiatric Services 67(8): 850–857. [DOI] [PubMed] [Google Scholar]
  20. Lorvick J., Comfort M., Kral A. H., Lambdin B. G. (2018). ‘ Exploring Lifetime Accumulation of Criminal Justice Involvement and Associated Health and Social Outcomes in a Community-Based Sample of Women Who Use Drugs.’ Journal of Urban Health 95(4): 584–593. [DOI] [PMC free article] [PubMed] [Google Scholar]
  21. Lowery W., Kindy K., Alexander K. L. (2015). Distraught people, deadly results. The Washington Post. http://www.washingtonpost.com/sf/investigative/2015/06/30/distraught-people-deadly-results/.
  22. Lum C. (2011). ‘ The Influence of Places on Police Decision Pathways: From Call for Service to Arrest.’ Justice Quarterly 28(4): 631–665. [Google Scholar]
  23. Lum C., Koper C. S., Stoltz M., et al. (2020). ‘ Constrained Gatekeepers of the Criminal Justice Footprint: A Systemic Social Observation Study of 9-1-1 Calltakers and Dispatchers.’ Justice Quarterly. 10.1080/07418825.2020.1834604. [DOI] [Google Scholar]
  24. Mendoza S. A. (2018). Consequences of Illinois’ 2015–2017 Budget Impasse and Fiscal Outlook. Office of the State Comptroller. https://illinoiscomptroller.gov/financial-data/find-a-report/special-fiscal/consequences-of-illinois-2015-2017-budget-impasse-and-fiscal-outlook/ (accessed 29 November 2019).
  25. Morabito M. S., Kerr A. N., Watson A., et al. (2012). ‘ Crisis Intervention Teams and People with Mental Illness: Exploring the Factors That Influence the Use of Force.’ Crime & Delinquency 58(1): 57–77. [Google Scholar]
  26. Moskos P. (2007). ‘ 911 and the Failure of Police Rapid Response.’ Law Enforcement Executive Forum 7(4): 137–149. [Google Scholar]
  27. Neusteter S. R., Mapolski M., Khogali M., O’Toole M. (2019). The 911 call processing system: a review of the literature as it relates to policing. Vera Institute of Justice.
  28. Paoline E. A. III. (2003). ‘ Taking Stock: Toward a Richer Understanding of Police Culture.’ Journal of Criminal Justice 31(3): 199–214. [Google Scholar]
  29. Steadman H. J., Osher F. C., Robbins P. C., Case B., Samuels S. (2009). ‘ Prevalence of Serious Mental Illness among Jail Inmates.’ Psychiatric Services 60(6): 761–765. [DOI] [PubMed] [Google Scholar]
  30. Skubby D., Bonfine N., Novisky M., Munetz M., Ritter C. (2013). ‘ Crisis Intervention Team (CIT) Programs in Rural Communities: A Focus Group Study.’ Community Mental Health Journal 49(6): 756–764. [DOI] [PubMed] [Google Scholar]
  31. Substance Abuse and Mental Health Services Administration. (2018). Key Substance Use and Mental Health Indicators in the United States: Results from the 2017 National Survey on Drug Use and Health (HHS Publication No. SMA 18-5068, NSUDH Series H-53). Rockville, MD: Center for Behavioral Health Statistics and Quality, Substance Abuse and Mental Health Services Administration. https://www.samhsa.gov/data/sites/default/files/cbhsq-reports/NSDUHFFR2017/NSDUHFFR2017.pdf. [Google Scholar]
  32. Teller J., Munetz M., Gil K., Ritter C. (2006). ‘ Crisis Intervention Team Training for Police Officers Responding to Mental Disturbance Calls.’ Psychiatric Services 57(2): 232–237. [DOI] [PubMed] [Google Scholar]
  33. Teplin L. (1984). ‘ Criminalizing Mental Disorder: The Comparative Arrest Rate of the Mentally Ill.’ American Psychologist 39(7): 794–803. [DOI] [PubMed] [Google Scholar]
  34. Thresholds. (2013). ‘The Path Forward: Investing in the Illinois Community Mental Health System Improving Lives, Saving Money’, Thresholds. http://www.thresholds.org/wp-content/uploads/2013/11/Path-Forward_Investing-in-Illinois-Community-MenUshertal-Health_Final.pdf (accessed 19 February 2015).
  35. University of Memphis. (n.d.). Local and Regional Programs. http://cit.memphis.edu/ (accessed 12 August 2020).
  36. Usher L., Watson A. C., Bruno R., et al. (2019). Crisis Intervention Team (CIT) Programs: A Best Practice Guide for Transforming Community Responses to Mental Health Crises. Memphis: CIT International. citinternational.org/bestpracticeguide. [Google Scholar]
  37. Watson A., Angell B. (2013). ‘ The Role of Stigma and Uncertainty in Moderating the Effect of Procedural Justice on Cooperation and Resistance in Police Encounters with Persons with Mental Illnesses.’ Psychology Public Policy Law 19(1): 30–39. [DOI] [PMC free article] [PubMed] [Google Scholar]
  38. Watson A. C., Compton M. T., Draine J. N. (2017). ‘ The Crisis Intervention Team (CIT) Model: An Evidence‐Based Policing Practice? United States.’ Behavioral Science & the Law 35(5–6): 431–441. [DOI] [PubMed] [Google Scholar]
  39. Watson A., Compton M. T., Pope L. G. (2019). Crisis Response Service for People with Mental Illnesses or Intellectual and Developmental Disabilities: A Review of the Literature on Police-Based and Other First Response Models. New York: Vera Institute of Justice. [Google Scholar]
  40. Watson A. C., Ottati V. C., Draine J. N., Morabito M. (2011). ‘ CIT in Context: The Impact of Mental Health Resource Availability and District Saturation on Call Outcomes.’ International Journal of Law and Psychiatry 34(4): 287–294. [DOI] [PMC free article] [PubMed] [Google Scholar]
  41. Watson A. C., Ottati V. C., Morabito M., et al. (2010). ‘ Outcomes of Police Contacts with Persons with Mental Illness: The Impact of CIT.’ Administration and Policy in Mental Health and Mental Health Services Research 37(4): 302–317. [DOI] [PubMed] [Google Scholar]
  42. Watson A. C., Wood J. D. (2017). ‘ Everyday Police Work during Mental Health Encounters: A Study of Call Resolutions in Chicago and Their Implications for Diversion.’ Behavioral Sciences and the Law 35(5–6): 442–455. [DOI] [PMC free article] [PubMed] [Google Scholar]
  43. Weisburd D., White C. (2019). ‘ Hot Spots of Crime Are Not Just Hot Spots of Crime: Examining Health Outcomes at Street Segments.’ Journal of Contemporary Criminal Justice 35(2): 142–160. [DOI] [PMC free article] [PubMed] [Google Scholar]
  44. Weisburd D., Bernasco W., Bruinsma G. (2009). Putting Crime in Its Place: Units of Analysis in Geographic Criminology. Springer-Verlag. New York [Google Scholar]
  45. Whalen M. R., Zimmerman D. H. (1990). ‘ Describing Trouble: Practical Epistemology in Citizen Calls to the Police.’ Language in Society 19(4): 465–492. [Google Scholar]
  46. Wise-Harris D., Pauly D., Kahan D., et al. (2017). ‘ Hospital Was the Only Option’: Experiences of Frequent Emergency Department Users in Mental Health’. Administration and Policy in Mental Health 44(3): 405–412. [DOI] [PubMed] [Google Scholar]
  47. Wood J., Beierschmitt L. (2014). ‘ Beyond Police Crisis Intervention: Moving “Upstream” to Manage Cases and Places of Behavioral Health Vulnerability.’ International Journal of Law and Psychiatry 37(5): 439–447. [DOI] [PMC free article] [PubMed] [Google Scholar]
  48. Wood J. D., Watson A. C., Barber C. (2020). ‘ What Can We Expect of Police in the Face of Deficient Mental Health Systems? Qualitative Insights from Chicago Police Officers.’ Journal of Psychiatric and Mental Health Nursing 28(1): 28–42. [DOI] [PubMed] [Google Scholar]
  49. Wood J. D., Watson A. C., Fulambarker A. J. (2017). ‘ The “Gray Zone” of Police Work during Mental Health Encounters.’ Police Quarterly 20(1): 81–105. [DOI] [PMC free article] [PubMed] [Google Scholar]

Articles from Policing (Oxford, England) are provided here courtesy of Oxford University Press

RESOURCES