How Police Officers Assess for Mental Illnesses

Abstract

This study examined how police officers assess for mental illnesses and how those assessments vary by location. Researchers conducted semi-structured interviews with 15 officers working in two police districts in one city. Findings from the study indicate that officers make assessments based on information from dispatch, collateral contacts and behavioral observations on the scene. The study also found that neighborhood context shaped the assessment process due to variations in types of information that were available to officers in different locations. The findings indicate that there is a need to improve the quality of all three sources of information.

On January 6, 2017, San Francisco police officers responded to a neighbor’s complaint about Sean Moore, a man diagnosed with bipolar disorder and schizophrenia. The police interaction resulted in one officer falling down a flight of stairs, another officer having a broken nose, and Mr. Moore being shot in the abdomen and groin. Mr. Moore’s family and lawyers claimed that officers should have been able to recognize Mr. Moore’s mental health symptoms and handled the situation differently (Emslie, 2017). The question is how are officers supposed to identify someone as having a mental illness? This question is pertinent not only to Mr. Moore’s situation but also to all encounters between police and individuals with mental illnesses. The ability of police officers to correctly identify mental health symptoms could make a substantial difference in many aspects of the encounter, including the level of force, whether an individual is connected to an emergency mental health evaluation, and whether the individual is arrested (Engel & Silver, 2001; Johnson, 2011; Swartz & Lurigio, 2007; Watson et al., 2010). However, despite a growing body of literature on police encounters with persons with mental illnesses, little is known about how officers are making these assessments.

One of the factors that make police officers’ assessments for mental illnesses challenging is that signs of a mental illness can look similar to signs of substance use. Evidence of this challenge can be found in nearly every hospital emergency room where even experienced psychiatrists frequently confuse signs of mental illness with those of substance use (Claassen et al., 1997; Schiller, Schumway, & Batki, 2000; Shanzer, First, Dominguez, Hasin, & Caton, 2006). However, the ability to identify signs of mental illness in suspected offenders could be crucial to the approach officers take during police encounters and the way in which a situation is resolved. Engel and Silver’s (2001) research on police interactions found that police officers’ recognition of a mental illness reduced the likelihood that an individual would be arrested, whereas recognition of substance use increased the risk of arrest. Similarly, Watson and colleagues (2010) found that people with a mental illness were at increased risk for arrest if they were also using a substance at the time of police contact. Taken together, these studies indicate that police officers might not always be able to recognize the presence of a mental illness (Engel & Silver, 2001; Teplin, 1984; Watson et al., 2010).

In the absence of data on how police officers assess if a person has a mental illness, interventions have developed from the premise that the assessment process can be improved by teaching officers about the signs and symptoms of mental illnesses. While these trainings, such as those offered through Crisis Intervention Team (CIT) programs, have been shown to increase officers’ self-reported confidence and skills in recognizing signs of mental illness (Compton, Bahora, Watson, & Oliva, 2008; Compton et al., 2014), much remains unknown about the ways in which officers use the information from trainings to assess for the presence of mental illness during the chaos of a live policing encounter. Research that examines the processes used by police officers to make mental health assessments on the street is important because officers are taught that, even with formal training, they must ultimately rely on their own judgment and experience when they encounter complex and ambiguous situations (McNulty, 1994).

Research that has examined police officers’ decision-making processes during live encounters has suggested that over time officers develop “recipe rules” (Ericson, 1982, p.25), that is, working rules based on their experience regarding how they should behave in certain situations (Stroshine, Alpert, & Dunham, 2008). Reacting intuitively is encouraged in the police officer training academy and further reinforced through officers’ experience on the job (McNulty, 1994; Pinizzotto, Davis, & Miller, 2004). In her ethnography of a police academy training program, McNulty (1994) found that officers were taught to act quickly and decisively when assessing situations that forced them to rely on something she called “commonsense knowledge” (p.282). Developing an understanding of officers’ commonsense knowledge about mental illnesses can provide insight into what information officers are using when making assessments of mental illness and the processes they might be using to integrate knowledge from their training into their daily practices.

Because socialization plays a key role in the development of commonsense knowledge, critical insights can be gained by examining the social context of the locations in which officers’ interactions take place (Chappell & Lanza-Kaduce, 2010; Ford, 2003; Marion, 1998; McNulty, 1994). Because most police officers’ activities are geographically bound, the nature of public life within their district is likely to affect the recipe rules that officers develop and the ways in which the officers react to the situations and people they encounter (Klinger, 1997). Klinger proposed that in areas with high levels of deviance such as those characterized by the public presence of drug dealing, prostitution, violence, as well as dilapidated buildings and poor living conditions, officers find considerable overlap between offenders and victims. This type of social context can lead officers to see all members of the community as deviant and undeserving of help.

Given the complex nature of police officers’ assessments in the field, developing a better understanding of the cognitive processes used to assess for the presence of mental illnesses during encounters with suspects could provide insight into the challenges officers face when trying to detect if someone has a mental illness. The research presented here adds to this literature by examining the following questions: (a) What sources of information do police officers draw on when assessing for the presence of mental illness during encounters? and (b) In what ways, if any, does this assessment process vary by local context?

Methods

Research Design and Setting

The research presented here used qualitative interviews to explore the ways in which police officers made assessments of mental illnesses during encounters. This research drew from a larger study conducted in 2009 in two large U.S. cities (one in the Northeast and one in the Midwest) that examined police officers’ attitudes and beliefs about mental illness (reference to published study includes authors’ names). The study data were drawn from two police districts located in the Northeastern city.⁶ The first district (District A) was home to more than 25,000 people and included a large business district that drew a substantial number of commuters into the district each day. More than 60% of the district’s population was White, and slightly more than 20% of the population lived in poverty. In comparison, the 2nd district in this study (District B) had a population of nearly 46,000 people. This district was only 20% White with a poverty rate of 45%.

In addition to differences in the racial and income demographics of the residents, there were also major differences in the amount of crime in these districts. The year the interviews were conducted, there were four murders in District A, much lower than the city average of 11 murders per district. District B had 27 murders, the second highest number of all districts in the city. Four of those murder victims were under 18, making District B the district with the largest number of juvenile homicides. District B also had over 1,600 narcotics arrests, the highest number of any district, while District A only had 214 people arrested for drug possession or distribution. District A’s predominant social problem was homelessness as there were five major homeless shelters in the district and the downtown area was a boon for people panhandling.

Participants

Researchers obtained lists of officers from District A (n = 90) and District B (n =156) from the police department. Using randomly generated numbers, the research team randomly selected 10 officers from each district and then invited the selected officers to participate in the study. If a selected officer declined to participate, then a new officer was selected using the same method for random selection. Following the recommendation of Lincoln and Guba (1985) to conduct interviews to the point of redundancy (i.e., saturation), the research team planned to interview 10 officers from each district; however, saturation was reached earlier than anticipated. Therefore, the final sample size for this study was 15 participants (District A, n = 6; District B, n = 9).

Data Collection

The officers who agreed to participate were interviewed at their district station during their on-duty hours. Each officer completed an individual semi-structured interview that lasted between 30 and 60 minutes. These interviews used prototype methodology, which asks the interviewee to describe a concept in great detail (Skeem & Golding, 2001; Smith, 1991, 1993). In using this methodology, each officer was asked to take a few minutes to picture a typical call in which a person was suspected of having a mental disturbance, and then to describe the event in as much detail as possible. The officer was then asked to talk about other types of calls they have responded to within their district. At least two research team members were present for each interview, and all interviews were audio recorded and transcribed verbatim after the interview; the transcripts were uploaded to Atlas.ti for management during analysis.

Interview Data Analysis

The first author used thematic analysis methods (Braun & Clarke, 2006) to analyze officers’ descriptions of typical mental disturbance calls. Thematic analysis involves six stages: becoming familiar with the data, generating initial codes, searching for themes, reviewing themes, defining and naming themes, and producing the final report (Braun & Clarke, 2006). During the first stage, to develop familiarity with the data, the first author read the transcripts multiple times and listened to the audio recordings of the interviews (Ritchie & Spencer, 2002). During this process, interview segments relevant to this study’s research questions were identified. Then, during the second stage of analysis, the first author used a line-by-line open coding approach to develop an initial list of descriptive codes that represented officers’ references to a person’s mental health.

During the third stage of analysis (i.e., searching for themes), the descriptive codes developed in the previous stage were compared with one another and then grouped into broader categories that represented types of information that officers identified as relevant to their assessment of mental illness. Stages 4 and 5 involved defining each of the themes and recoding the data using those themes. During Stage 4, to challenge the first author’s developing understanding of the data and to ensure analytic rigor, additional strategies were incorporated, including peer debriefing with an experienced qualitative researcher from another university who was familiar with this research area (Padgett, 1998). In addition, negative case analysis techniques were used to challenge the first author’s developing analytic framework; analysis continued until information from these cases were integrated into the final analytic framework (Creswell & Miller, 2000; Lincoln & Guba, 1985). The final stage of thematic analysis involved generating a report of the findings (Braun & Clarke, 2006), which is presented in the next section. To protect the officers’ confidentiality, the research team assigned each officer a number (ranging from 1 to 15); District A officers are represented by numbers 1 to 6, and District B officers are represented by numbers 7 to 15.

To ensure the authenticity and rigor of the analyses, additional techniques were incorporated into the analytic process, including developing a clear audit trail by using coding and analytic memos that explained and documented coding decisions and the meaning of specific codes that emerged during the analysis (Pandit, 1996). Additionally, the first author wrote theoretical memos regarding emerging themes, relationships in the data, and questions that required further coding or consideration during peer debriefing sessions (Strauss & Corbin, 1994).

Results

Thematic analysis identified several sources of information that officers reported using to assess the presence of mental illnesses as part of their response to calls. These sources of information can be grouped into three themes or categories: (a) information from police dispatch, (b) information from collateral contacts during the police encounter, and (c) the behavioral observations made by the officer during the encounter. A detailed discussion of each of these information sources is presented in the following section, including a brief definition of the category, a description of how the officer used each source of information in their assessment of mental illness, and how the officers’ use of each source of information varied by geographic location (i.e., police district).

Information from Police Dispatch

Officers participating in this study reported that information from police dispatch was usually conveyed to them by police radio before the officer arrived at the scene. Information provided by police dispatch included details related to the nature of the problem, the location of the incident, and the description of those involved. When officers were asked about the first thing that came to mind when they received a mental disturbance call, three of the six officers in District A indicated their initial thought was that the individual might be homeless. Officer 4 said, “I guess initially you might think it’s like a homeless male or female acting up or doesn’t want to be inside of a place; or whatever it could be. You don’t know. That’s what happens in this particular district.” None of the District A officers mentioned initial thoughts of substance use as the cause of the mental disturbance call. Conversely, none of the nine District B officers indicated homelessness was their initial thought when they received a mental disturbance call; however, three of the nine mentioned that the subject person of the call might not have a mental illness, but rather, might be on drugs. For example, when Officer 11 said, “It could be a rash of things. It could be they’re not— they’re not mental disturbed; they’re high.” In a separate interview, Officer 13 responded,

What comes to mind, I would probably say, anything that deals with drugs, narcotics. In this area, it’s [substance use] prevalent… I wouldn’t say that you that you would have the same kind of disorders like in [downtown] as you would have in the [District B] in this area.

Officers in both districts reported using police dispatch information regarding the physical location of the suspect at the time of the call to help discern whether the call could involve a person with mental illness. One way officers reported using the physical location was to determine whether the address was associated with a mental health facility. For example, 2 of 6 officers in District A referred to a particular mental health facility in their discussion of how they used dispatch information in the assessment process saying, “We get calls in [neighborhood]. There’s a building there that has a lot of 302s [involuntary commitments] phoned in [from that location](Officer 2).”

Another way that officers reported using the physical location of the subject was to identify whether the call emanated from a residential address associated with a person who had a mental illness. The comment from Officer 6 was typical of the responses to this question:

Okay, I had a person with a gun call and a bunch of other officers went because they knew the address. I’m new to this district so I didn’t know. So, I was the first one there and I go in and they said, “Be careful, there’s a known 302 [involuntary commitment] there.”

In District B, which had more residential property than District A, five of nine officers reported that might identify a call as involving a person with a mental illness based on the address. These five officers referenced what were known as frequent callers: people with mental illnesses who called to report bizarre crimes. Officer 7 described one such call that involved an address associated with an individual with mental illness:

So, just responding to a house for a female 302 caller who would state that the neighbor in the adjoining home is spraying chemicals through her walls that are burning her skin and breaking in through her back door and placing bell peppers in her refrigerator and taking out tomatoes and replacing them with peppers, that I guess help irritate her skin some more. The neighbor [female 302 caller] is constantly [accusing the next door neighbor of] stealing documents. The neighbor [female 302 caller] has forced her [the next door neighbor]to have to change her locks several times, and the individual [female 302 caller], because of the number of chemicals that are allegedly being sprayed through the walls, she smears her face; like completely covers her face with Vaseline where it looks like her face looks like a wax candle, basically.

Officer 13 provided a similar account of a frequent caller with an address well known to officers as associated with a person who has a mental illness:

Well, there’s one lady up in the Northwest part that probably will call maybe once a day because of loud music. We get there, and there’s never any loud music. It’s kind of sad because she lives by herself, she’s like 80 years old, and there are no kids or nothing. So we’re like “ok.”

The other type of dispatch information that officers reported using in their assessment of mental illness involved reports that a person might have a mental illness, be suicidal, or be the subject of an involuntary commitment. For example, Officer 6 described a call dispatched as a suicide, “…a woman must have called the Suicide Hotline and they called 911.” Officer 12 relayed a story about a man who was the subject of an involuntary commitment:

The last call I had was a male with approved paperwork. I don’t remember his name. They said he was suicidal and possibly violent. After we got there, a very large male sitting on his front porch. His mother, an older lady probably in her eighties, she had the paperwork…

Because much of District A was a business area with high concentrations of homeless persons, the people who called the police often did not have direct knowledge of the person’s mental health diagnosis. The officers did not report that callers were intentionally misleading the dispatchers but instead recognized that these callers often had limited information. For example, Officer 4 said, “Usually when somebody is just talking to themselves people call. The citizen or victim was concerned, so they just call. You get there and sometimes they’re not [still talking to themselves]. So it’s like 50/50.” District B officers reported that calls involving people with mental illness typically originated from friends, family, or neighbors who had more intimate knowledge of a person’s condition and background. Several District B officers expressed a belief that some callers intentionally provided dispatch with inaccurate information or exaggerated the severity of a person’s symptoms. For example, Officer 12 recounted a situation in which a parent called to report that her child was having a mental health problem:

Like the mom calls, and he’s out there smashing all the furniture. You get in and there’s no paperwork or anything like that. They [parents] know that their kids are doing drugs, [but] they’re not going to say they know, no matter who they’re calling...

Although dispatch sometimes provided useful information, at other times the dispatch information was limited or even misleading. Therefore, because officers could not rely on information from dispatch to identify whether a situation involved a person with a mental illness, the officers frequently needed other sources of information when trying to assess whether mental illness was present in the situation.

Collateral information.

Collateral information was defined as the information conveyed to police officers by bystanders, friends, family members, or neighbors during the time the officers were on the scene of the call. Officers reported that collateral information was particularly helpful in providing background information regarding either the person who was the subject of the call or the situation the officers were called to address. Across districts, five of the six officers in District A and eights of the nine officers in District B discussed the role that collateral information played in their process of assessing for mental illness during calls. Officers reported that the collateral contacts who provided collateral information (e.g., bystanders, family, friends, or neighbors) were frequently, but not always, the same people who had made the initial call and provided information to dispatch. Nevertheless, even if the collateral contact was the same person as the initial caller, the interaction with collateral contacts at the scene often provided an opportunity for officers to obtain more information about the individual. Officer 5 discussed the importance of obtaining information from collateral contacts, saying, “You have to ask the appropriate questions, obviously. If there’s family around, what kinds of issues do they have; that kind of stuff, to try to get to the root of what’s causing this behavior.” When Officer 10 was asked how he knew a man had a mental illness, he provided the following response:

Yeah, they told me about it on the porch, and that he needed his medicine and all that good stuff. After I told him [that] I wanted to take him to get his medicine—like I said, just talking to him, “What’s your problem? What’s going on? What’s bothering you?” It brought him down.

The quote from Officer 10 shows the connection between the information officers receive from collateral contacts at the scene and the officer’s behavioral assessment. Based on the information provided by the family, this officer knew what types of questions to ask to find out more about the suspect’s mental health issues. Officers not only reported that they wanted to obtain as much information as possible from collateral contacts but also reported feeling frustrated when others assumed the officers did not need that information. In talking about an interaction with a homeless outreach program in which the officer was transporting a person for a psychiatric evaluation, Officer 2 commented:

I mean, it’s a job and we don’t mind, but sometimes it’s just the way the job is perceived by them [the public], is just like, “ Take ‘em. What’s wrong with them?” We want to know, too. You know, they kind of call, they call [the psychiatric crisis center]. We always have to notify [the psychiatric crisis center] before we take a body over there. So they’re kind of warned, but they also want to know what’s going on with the person. Sometimes we don’t know, we’re like, “We just picked this guy up, so we’ve got to bring him over there.”

Officers reported that occasionally neighbors, who were not the initial callers, were able to provide helpful collateral information. For example, when discussing a frequent caller, Officer 7 stated, “So because of her antics [the frequent caller], she would, in turn, bother the neighbor; so, we go and talk to the neighbor to kind of get an idea from her what actually has been going on with her next door [the frequent caller].” Similarly, Officer 8 reported that neighbors would help by providing collateral information:

You can go to somebody’s house in the neighborhood; They say, like, “He’s crazy, he’s got mental problems, he’s taking medication.” Because everybody knows. Now if it’s just like I just walk down the street and I see somebody on the street flipping out, I mean, I wouldn’t know. (Officer 8)

This quote from Officer 8 clearly indicated that the officer likely would have had a difficult time identifying why a suspect was behaving bizarrely if not for the information provided by family and other people in the neighborhood. In contrast, officers in District A usually did not have the benefit of neighbors with collateral information. In cases of calls from mental health facilities and homeless outreach programs, the officers could obtain additional information on the scene. Officer 1 explained, “They [the collateral contacts] usually are like, I guess, the case managers, and they’ll explain the situation and what kind of treatment they [the suspect] needs.” However, many of the calls in District A came from concerned citizens reporting the bizarre behavior of someone on the street; frequently these suspects were no longer on the scene when officers arrived. Officers in those situations were almost solely reliant on their own assessments of the situation to discern whether mental illness was involved.

In other cases, officers might have to rely on their behavioral assessments if they felt as if one or more collateral contact might be providing inaccurate information or intentionally omitting important details. Notably, none of the District A officers talked about collateral contacts intentionally providing inaccurate information. However, several officers in District B reported feeling that collateral contacts might intentionally mislead police officers, especially in situations involving youth. For example in Officer 11 said,

Usually, if I get a call, I’ll say [a call from a] mom or whoever, if they say they’re [the suspect is displaying behavior that they’re [the caller] worried about—’cause a lot of people misinterpret 302. You know what I mean? They’ll say a 302 because he’s [the suspect] violent and he does this and that. I’m like, “Do they do drugs? Maybe they get high?” There are a lot of factors that can be involved in that [behavior display].

In this example, Officer 11 did not address whether the collateral contact was intentionally excluding information about drug use. In other situations, particularly in District B, officers reported believing that collateral contacts were intentionally misleading the police officers. For example, Officer 14 reported a belief that some parents would try to use the mental health system to deal with a child they could not control:

Some of them, the parents, just want to get rid of the kids. “No, I can’t deal with this. [Family court] doesn’t want to take them.” So, they want to get them [the child] 302d, and it’s like, “Okay, but he’s not crazy or she’s not crazy.” There’s nothing wrong with them [the child] …. Matter of fact, we had one last week. The girl is having problems with the mom and the mom’s boyfriend. She’s basically acting out and mom is insisting, “No, she’s bipolar.” The girl is like, “I never took medication for being bipolar. I’m not bipolar. No doctor has ever literally told me that I’m bipolar. It’s just my mom; she doesn’t want to deal with me anymore.” (Officer 14)

Similarly, Officer 15 echoed a similar situation and sentiment: “A lot of times we’ll have individuals who want a 302 petition for their children just because their children are out of control or having behavioral issues that aren’t necessarily mental illness.”

District B officers also provided several examples of situations in which the officer responding to a call reported believing that collateral contacts were intentionally providing misleading information about an adult. For example, Officer 8 said,

We show up and it’s like, I can’t 302 this person because they’re not showing it [signs of mental illness], but sometimes people do get upset, which is normal, and then you mistake that for them being emotionally disturbed. You’ll think they’re being crazy when actually they’re just upset because this person [the caller] is lying on them, and then you end up taking them [the suspect] and 302ing them, so this person who called the police initially wants that person out so they can keep their money. That’s a lot of cases.

In addition, District B officers reported believing that collateral contacts might provide inaccurate information if they were worried that the suspect would get in trouble for drug use. Officer 15 said,

If there’s some sort of substance abuse, a lot of times people will try to cover the fact that it’s [the disturbed behavior is related to] substance abuse, and that’s a lot of time why they call it in as a 302 call or because they’re trying to hide the fact that this person’s having a bad trip on PCP [phencyclidine].

Behavioral assessments.

Officers reported that information from collateral contacts and other sources could be incomplete, inaccurate, or misleading. To address this problem, officers reported relying heavily on their own behavioral observations of individuals during face-to-face interactions. While on the scene of the call, officers indicated that they used their own behavioral observations of the suspect and others to inform their assessment for mental illnesses. These observations enabled officers to identify that an incident likely involved someone with a mental illness even if that information was not included in the call dispatch. For example, Officer 2 recounted a call initially dispatched as a robbery:

They said a guy had taken a soda from one of the hotdog vendors at (address)…When the hotdog guy came out to approach him, he [the suspect] assaulted the guy. So now that’s put it up to a robbery instead of a mere theft. They’re both arrestable offenses, but suddenly we find the guy. I see the guy on [address] and we stopped him. He was rambling on and he said he was – it’s probably not the correct name, but he gave me something to the effect of Kubla Khan or Shaka Zulu. He said his name was Shaka Zulu. I couldn’t understand half of what he was saying. I just remember him saying, Shaka Zulu. He had all kinds of cigarette butts and trash in his pockets. He was very unkempt. He smelled bad. So he’s been on the street. I don’t know when the last time this guy got a shower was but he was still involved in an incident and the hotdog vendor was the complainant. He [the vendor] IDs him as the gentleman who assaulted him. At that point, I’m not going to arrest this guy [the suspect] because he’s not aware of what it is he’s doing, or why he was doing it, or the fact that he was even doing something wrong, because he didn’t run.

Officer 9 experienced a similar situation when dispatched to a domestic disturbance:

The lady was throwing stuff out the window [like a] a microwave, banging the door, or whatever. So, when I knock on the door, a little Asian kid who looked like he was handicapped because he didn’t talk and he had no clothes on. So, now you open the door and first you look inside. You don’t know what’s… who you’re… you’ve got to be careful first, and then you hear water running from the kitchen. You look inside and you see the TV on the floor, you see plants all over. Now you still look and make sure nobody’s in the kitchen, and then when I look upstairs, I see the guy’s mom with no clothes on; banging the door, opening and closing the door. [She says,] “Sit down, have a seat. I’ll be right down. Sit down, have a seat. I’ll be right down.” Then you figure the little kid with no clothes, the house is ransacked; you try to bring the lady… first, you call a supervisor and then you try to bring her down. You sit her down, but she still can’t talk to you. You know what that was: a 302.

In both incidents, officers walked into the situation with no indication that the call involved someone with a mental illness. In the absence of information from dispatch or collateral information regarding the likelihood of mental illness, these officers were forced to rely on their environmental and behavioral observations to assess for the presence of mental illness.

In recounting their encounters with people with mental illnesses, every officer in the study used the term not right when describing behaviors indicative of an altered mental status. Overall, the officers did not describe the behavior using diagnostic terms except when a collateral contact provided the terms for a specific mental diagnosis. In addition to the term not right, officers tended to use vague terms to describe a person’s mental status, including that the person was “absolutely gone” or had “left the reservation.” For example, Officer 8 described his assessment of a suspect’s mental status:

I couldn’t describe his mental status like in a medical term, but he’s absolutely gone. I mean, there’s no rational thinking. It sounds like it’s like a 10-year-old kid talking with the things that he’s making up, you know, the things that he’s envisioning or whatever.

Moreover, officers described a wide range of behaviors to illustrate what they meant by a person is not right. Officers in District A primarily focused on behaviors that created public disturbances:

Numerous times I’ve come across someone who may be lying in the street over a grate, not clothed, walking in and out of traffic, mumbling to themselves, cursing—you know, profanity at no one or anyone; chasing people, punching people, someone who is not in their right frame of mind and a danger to themselves or others. (Officer 5)

In this example, the majority of the behaviors that Officer 5 described involved the suspect responding verbally or physically to some type of internal stimuli. The officer was not referring to the cognitive processing such as hearing voices or experiencing delusional beliefs. The District A officers did not describe encounters that involved suspects with visual or auditory hallucinations. Although the District A officers referred to people talking to themselves, the officers did not state an explicit connection between people talking to themselves and the likelihood that the person was experiencing hallucinations. Overall, officers in District A tended to focus on external symptoms of mental illness, but not necessarily evaluating for internal symptoms. Only two District A officers described encounters in which an individual was experiencing delusional beliefs. Officer 2 described a woman with paranoid beliefs:

She was ready to just do whatever she had to do to, you know, for herself, I mean, she thought that we were doing wrong to her. She thought that we were there to hurt her, instead of being there to help her.

Officer 6 also described an encounter with a woman who had delusional beliefs: “She was paranoid schizophrenic or whatever and thinks her neighbors are breaking in, trying to rob her, and people [are] trying to kill her; but clearly not the case.”

In contrast, District B officers referred to a wide range of symptoms when describing their encounters with people who had a mental illness. For example, these officers referred to impaired cognitive processes such as “hearing voices” and “seeing things.” These terms, as opposed to more clinical language such as hallucinations, show how officers are using their commonsense knowledge to assess individuals. These officers gathered this information from collateral contacts and from the people experiencing the symptoms. Officer 12 described the process he used to inquire about an individual’s mental health symptoms:

We talked to him and one of the good things, which I never thought to ask before, which I got with the CIT training, which is you ask if the person is hearing voices, ask them what they’re [the voices] saying. I would’ve never thought to come at that at an angle like that prior to that [training]. This gentleman, if he wanted to fight, we were going to have a fight on our hands, because he was a big boy. I asked him, I said, “Are you hearing voices?” He said, “Yeah, I am.” I said, “What are they telling you?” “They’re telling me to punch you in the face and then kill everyone here.” I said, “We don’t want that.” But because of that, I was able to understand that the male obviously has something going on in his head that he can’t control, so we wanted to treat him with extreme caution, which we did.

In the situation Officer 12 described, it was unclear whether the officer had received information about the individual’s mental status from dispatch or once on the scene from a collateral contact. Therefore, the questions and behavioral observations were likely not for the purpose of detecting the presence of a mental illness, but rather to uncover the nature of the mental health symptoms. As the officer explained, he was able to use the information gathered at the scene in deciding how to interact with the individual.

Similarly, Officer 11 described the way in which he used questions to assess an individual’s mental status:

People who have mental illness can come with a rash of different things. They don’t look mentally ill; they don’t always act mentally ill. They talk and look normal, but sometimes you get to talk to them or sometimes (inaudible), “Do you take medication for anything?” They’ll tell you yes or no, or “I haven’t been taking my medicine.” It depends. You have to basically talk to them to see what the situation is about and then you will kind of get the gist of, yeah, this person has a couple of mental issues, you deal with it that way.

District B officers also described callers who reported experiencing strange beliefs such as believing they were Jesus or that they were FBI agents. A common delusion that was referenced by several officers involved callers who believed they were the victim of a crime. Based on the officers’ prior experience with these individuals or due to the bizarre nature of the stories, the officers recognized that the individual might be experiencing delusional beliefs. The officers did not use the term delusions but instead referred to delusional thinking under the general label of a person is not right. For example, Officer 7 described a woman who he has encountered frequently because of her mental illness:

She really hasn’t been right since and has kind of gone off the reservation a bit. She calls regularly… She really just says like stupid things, like goofy things; nothing that really raises concern that you think that she would harm herself or harm somebody else.

When officers in both districts were asked how they knew if a person had a mental illness, they typically had difficulty articulating a response. An officer in District A responded by saying, “Sometimes they have like—I don’t want to sound stereotypical—but they have like, like a look to them.” Officer 6 said, “You can tell right away that they all have the same look and that stuff.” The responses of these officers suggest officers rely on some type of commonsense element to the recognition of mental illness. Although officers might have difficulty in articulating this commonsense aspect or the process they use to recognize the presence of mental illness, the officers’ responses indicate these officers believe that they know mental illness when they see it.

Discussion

The results of this research provide a framework for the sources of information or commonsense knowledge that officers use in their interactions with citizens to identify mental illness. This framework can be used to develop strategies for augmenting formal training programs for police officers such as creating direct links between training programs and the information sources that officers use to assess for mental illness in their routine interactions. Although some research has examined dispatch data and officer assessments in an attempt to identify police encounters that involved mental illnesses (Ritter, Teller, Marcussen, Munetz, & Teasdale, 2010), our findings identified a third crucial source of information: collateral information provided by collateral contacts. Moreover, this study illustrates how officers integrate information from the various sources in their assessments in specific situations. Additionally, findings from this study demonstrate how the geographic location of officers can shape the availability and interpretation of all three sources of information.

This research found that in predominantly commercial areas, such as District A, information provided by police dispatch might play a relatively smaller role than other information sources in helping officers determine whether a situation involves someone with a mental illness. This diminished role of dispatch information is principally because many callers reporting a public disturbance do not have knowledge of the individual’s background. In districts with a larger proportion of residential areas in which family or neighbors are more likely to have some knowledge of an individual’s mental illness, police departments should consider training emergency communicators to elicit information about mental illness from the caller, and then ensuring that information is relayed to officers. Emergency communicators or 911 operators could be trained on specific questions to ask callers and instructed in a protocol for sending the most relevant information.

Findings from this study indicate that a district location might also influence the type of collateral information that collateral contacts share with officers as well as the ways in which officers interpret collateral information. For example, officer interactions with collateral contacts in District B might have been influenced by that community’s high level of drug trafficking, as evidenced by the assumptions of some officers that mental health calls were most likely about drug use. Additionally, some officers believed that if a behavior was caused by drug use and not a mental illness, a collateral contact might intentionally provide inaccurate information.

When a collateral contact is a service provider, one of the barriers to sharing information is that the Health Insurance Portability and Accountability Act (HIPAA) of 1996, amended to include a Privacy Rule in 2002 further modified in 2003, restricts providers to sharing only the minimum necessary information (United States, 2004). State privacy rules also apply and may be stricter than federal regulations. Officers in the study did not mention HIPAA or other privacy statutes as a barrier to obtaining information, but it is possible that privacy statutes were one of the reasons that homeless outreach programs and mental health and drug and alcohol agencies might only share minimal details about the background of individuals that they are calling about. In case where collateral contacts are neighbors, friends and family members, improving police-community relations could lead to greater sharing of accurate information. Findings from this study indicate that officers are aided in their assessment when people are willing to provide background information about individuals who may have a mental illness. If people within a community trust police officers, then the residents might be more likely to share accurate information. Likewise, when officers develop trust in people living in the district, then the officers might be more likely to trust the information they receive about an individual’s mental health or behavioral background. Increased background information could substantially assist officers in their assessment process, and thereby, improve the ways in which officers interact with individuals with mental illness. Additionally, police training can include content to help officers ask collateral contacts the type of questions that will produce the information the officers need to better assess a situation.

The officers’ conceptualization of mental illnesses as being characterized by bizarre behavior suggests that training programs should emphasize the wide range of signs and symptoms experienced by people with mental illnesses. Because police officers should not be asked to do the job of mental health clinicians, the identification of the specific mental illness is not as important as recognizing signs and symptoms of a range of diagnoses. Teaching officers to ask questions that assess for internal symptoms could enhance officers’ ability to recognize when people are experiencing cognitive or emotional issues. Providing officers more opportunities to actively engage in assessments of the wider range of signs and symptoms of mental illnesses, could also help officers to better identify symptoms of mental illness during their interactions and to more effectively interpret information from dispatch and collateral contacts.

Additionally, training programs on interacting with people with mental illnesses, need to include a component on how to assess and interact with a person using drugs and alcohol. More than half of those with a mental illness who end up in the criminal justice system also have a co-occurring substance abuse disorder (Peters, Wexler, & Lurigio, 2015). Therefore, it is important for officers to be trained in how to recognize the presence of both substance use and mental illness. Ideally this training would teach officers about the types of questions that can be used to assess for mental illnesses and substance abuse so that they can recognize a fuller range of factors that might be impacting a person’s behavior.

The research presented was drawn from one exploratory study of officers’ accounts of their experiences with people they identified as having a mental illness. Therefore, this study has several limitations that are important to consider when interpreting the findings. First, the study did not allow us to examine encounters in which officers failed to recognize the presence of mental illness in an encounter with a person with mental illness. Conversely, the data did not confirm that the study participants had accurately recognized an individual’s mental illness in the encounters they described. Future research involving direct observations of officers could allow for a comprehensive analysis of factors that facilitate or hinder an officers’ ability to recognize when they are interacting with a person with mental illness. Additionally, because this research relied on retrospective personal accounts, officers’ recall of the encounters might not have been accurate. However, this research provides valuable insight into the types of information that officers rely on when making assessments about the presence of mental illness in their daily interactions with citizens and advances the understanding of the processes officers use in making mental health assessments.

The framework we have presented is based on the experiences of officers in one police department. The information available and the ways in which that information is interpreted is likely to differ not only in rural and suburban areas but also by geographic regions (Klinger, 1997). Therefore, further research is needed in other settings.

Conclusion

Understanding some of the factors that influence the processes used by police officers to assess for mental illnesses is a crucial component of enhancing interventions that enable officers to refer people in mental health crises to qualified clinicians for in-depth evaluations. Study findings suggest that behavioral health professionals, in conjunction with police departments, must focus on not only improving police officers’ assessment skills but also improving the quality of information available to the officers to inform their assessment. In addition, both police and behavioral health professionals should focus on developing a better understanding of how geographic context shapes officers’ assessments of mental illness conducted as part of their daily interactions with citizens.