Send a cop, not a social worker
Police do a fine job of helping the mentally ill
Years ago — well before policing was being “reimagined” — I was ranting to my then-car partner about all the stupid mental health calls we had to go to. “Why do we have to go to every 911 call from a crazy person? Every call from a suicidal subject?” I told him police should be out of the mental health business altogether. We should have “mental health EMTs” I told him, and they should deal with this crap instead.
He told me to please shut up because we had work to do.
Dealing with mental illness as a cop is frustrating. I recall one occasion when I, as a rookie 23-year-old cop, was called to a scene where a transgender woman was threatening to jump off a building. Fortunately, she came down on her own just as we arrived and nothing tragic happened. I tried to idly make conversation with her while we waited for the ambulance and realized I had absolutely no fucking idea what I was doing. I had no idea what this person was going through, and no idea what to say to her. “It’ll all be okay?” Seems cliche.
So even a lot of cops hate doing this work, and I was one of them. But last year, as the “defund police” movement was gaining traction, I had a conversation with a few officers from my old department. The prevailing belief at the time was that the city was going to cut the police budget and officers were going to be laid off, possibly including the officers I was talking to. I asked my former colleagues what they thought about a new proposal to stop sending police to non-emergency calls for service which involved low-level mental health crises. “One less thing you’ll have to do when you’re all critically understaffed,” I joked.
To my surprise, they were steadfastly against the idea. “This is what we do, this is the job,” one of them told me. “We’re good at this, we know what to do, and if we don’t help these people, nobody else will.” I pointed out that canceling mental health calls was about to be a necessity from a resource perspective. If cops were really laid off, there was no way he and the remaining officers would be able to handle the volume of calls for service. “Probably,” he said, “but we’ll do the best we can. We’ll help one person, and then go on to the next one.”
Maybe it’s because policing is such a cynical profession, but something about the earnestness of that broke me a little bit.1 This cop was looking at one of two futures; one where he was unemployed and another in which he faced a Sisyphean task every day. But he knew he could help people who needed it, and he refused to quit. So I’ve spent a lot of time since that conversation reading and thinking about this problem.
I’ve come to realize that he was right, and I was wrong. Police exist to solve problems and serve the public, and that necessarily entails helping people who have mental health problems. Contrary to the prevailing narrative, the data shows that police are quite good at helping the mentally ill without using physical force. And there’s no evidence that social workers alone can handle these situations as well as cops do.
Mental health calls are as old as the job
There is a pernicious myth out there, popular both among police and their progressive critics, that cops only recently began dealing with mentally ill people. At some time in the past, the story goes, all the services which would have helped the mentally ill were defunded and the whole thing fell on the police. But at best, this is a half-truth. Even before the de-institutionalization movement gained steam in the 1960s, police were routinely asked to help the mentally ill.
Consider this account from the New York Times, published 115 years ago on December 3rd, 1906. The story is about a woman named Mary Nevison, found wandering the hall of a hotel without shoes by a bellboy. Nevison told the bellboy to get her some stationary and proceeded to write out a bunch of gibberish.
Nevison then locked herself in a hotel room and refused to exit for twenty-four hours. In response, “a policeman appeared from the Tenderloin Station” and tried to assist the hotel’s staff in convincing Nevison to open the door and go to the hospital.
Or consider this report about a Boston suicide hotline, published in the Times on February 8th, 1959. I’ve cut out most of the article to save space, but take a look at the paragraph on the right. How did the suicide hotline get help for a suicidal caller in 1959? They called the police, who showed up and took him to the hospital.
Fast-forward to 1970, and my favorite academic work about policing: Egon Bittner’s “The Functions of the Police in Modern Society.” Bittner observed, quite correctly in my view, that policing is not really about law enforcement at all:
Many puzzling aspects of police work fall into place when one ceases to look at it as principally concerned with law enforcement and crime control… It makes much more sense to say that the police are nothing else than a mechanism for the distribution of situationally justified force in society.
Bittner didn’t mean “situationally justified force” in the sense of cops pummelling people. He meant that cops use coercion to solve problems people ask them to solve, whether or not that problem is a criminal one. And people called the cops about all manner of non-criminal social problems:
In a tenement, patrolmen were met by a public health nurse who took them through an abysmally deteriorated apartment inhabited by four young children in the care of an elderly woman. The babysitter resisted the nurse's earlier attempts to remove the children. The patrolmen packed the children in the squad car and took them to Juvenile Hall, over the continuing protests of the elderly woman.
In a downtown residential hotel, patrolmen found two ambulance attendants trying to persuade a man, who according to all accounts was desperately ill, to go to the hospital. After some talk, they helped the attendants in carrying the protesting patient to the ambulance and sent them off.
So cops responding to suicide, social problems, and mental illness is nothing new. It’s been part of police work for over a century. I could tell you stories about policing in the 2010s that are functionally identical to any of these decades-old anecdotes.
Police already de-escalate dangerous situations
The prevailing opinion among lawyers, academics, the media, and even police chiefs at the moment is that police should not be asked to address mental health issues because there is simply too much risk that the cops will use violence. When a mental health incident does end in violence, especially deadly violence, many in the police reform crowd often insist that cops should have “deescalated” the situation.
Sometimes, when discussing a use of force incident involving a mentally ill or drug-addicted person, I have asked a police critic: “What were the cops supposed to do?” And they always say: “De-escalate.” It’s like people think that cops all have a Magic Deescalation Button they could push but refuse because they are bloodthirsty maniacs who enjoy violence. The real problem is that while de-escalation is a great idea, it’s hard to execute in reality. What, specifically, do you do and say? When do you turn to force? To get a sense of the challenge, I suggest this video:
I think this situation and the fact that Seattle’s civilian oversight body attempted to discipline this officer for “failing to de-escalate” illustrates the challenge of de-escalation in the real world. When a man is walking down the street waving an ice ax, what is the safest way to stop him? They tried talking to him. They gave him distance and time to comply. An officer tackled him and nobody got hurt.
I can see the argument from the oversight board’s side — if the guy had swung the ax at the cop as he ran up, this could end in a police shooting. On the other hand, how long do you let the guy wanted for robbery walk around downtown near bystanders waving an ice ax before you use some level of force to stop him? What if he takes a hostage and then someone gets shot? That’s exactly what happened in this case. So every choice means taking a calculated risk, and in the moment, it’s hard to know which is best with any degree of certainty.
Perhaps psychologists, psychiatrists, and social workers really do have the secret sauce for resolving these situations, but I doubt it. I’ve completed Crisis Intervention Team Training for cops taught by mental health experts. That training focused almost entirely on types of mental illness and social service programs but was pretty light on exactly what you say to schizophrenia-meth-guy so he’ll put the knife down. I assume that if the social workers really did know the magic words anybody can use to bring a dangerous confrontation to an end, they’d be sharing.
So how do you get good at talking to and safely detaining delusional people in an uncontrolled, dangerous environment? You practice, because this kind of thing is an art form. And that’s why cops are so good at this — they do it all the time, while social workers and other mental health professionals almost always work in hospitals, group homes, or other controlled environments. When a suicidal person is holding a gun to their own head or a schizophrenic is trying to stab imaginary demons with a knife, the expert you want isn’t a psychiatrist — it’s a hostage negotiator.
If you look at the FBI’s data about hostage negotiation cases above, you’ll see some remarkable stuff: Roughly 96% of hostage negotiator callouts are actually for “emotionally driven situations with no substantive demands” rather than traditional hostage situations like a bank robbery. Of these situations, 72% involved a weapon; 37% involved a handgun. A majority were resolved with negotiation, and the hostages were almost never injured.
My point here is that treating people with a mental illness and safely detaining those same people in uncontrolled environments are two related but different skill sets. If you sat on my couch and asked me to help you recover from depression, I would not do a very good job. But if you asked me to use a combination of verbal persuasion, teamwork, and less-lethal weapons to safely detain a mentally ill guy brandishing an ice pick on a public sidewalk? I’m gonna be better than any social worker, because I have been trained on using time, distance, and cover to safely resolve that situation.
This is why for police, successful de-escalation of mental health situations is the rule — not the exception. Despite the frequently repeated claim that police often use force against the mentally ill, most police departments are remarkably good at resolving mental health crises without using force. For example:
D.C.’s Metropolitan Police receive 90 mental health calls every day, for an annual total of over 32,000 calls. In 2020, DC police used physical force against mentally ill people a grand total of 243 times — meaning 99.2% of calls involving mental illness were resolved without physical force.
In 2019, the San Francisco Police Department received over 50,000 calls about mentally ill people and requests to check on people in distress. SFPD used force on mentally ill individuals just 51 times, meaning 99.9% of these incidents were resolved without any use of force.
Over the last four years, the Seattle Police Department has responded to over 36,000 calls involving individuals in a mental health crisis. Over 98% of these incidents were handled without any use of force.
Can social workers do better than a 99.9% success rate? Can they do it while also protecting the public and without placing themselves in danger? Maybe. But 99.9% is pretty damn good. Personally, I think if it ain’t broke, don’t fix it.
What’s the alternative?
You might be reading this and thinking: “Well Graham, that’s fine, but I’m not talking about sending social workers into potentially violent or criminal situations.” But if you aren’t thinking that, plenty of other people are. In Chicago, for example, they are starting to send “mental health clinicians” to 911 calls. Apparently the success of the program “will be measured against the cases of Laquan McDonald and Quintonio LeGrier.” Both of those individuals were mentally ill and killed by police while brandishing weapons — McDonald a knife, LeGrier a baseball bat. Will the city send social workers to calls about people brandishing weapons?
In Washington State, cops have already started walking away from calls about suicide. Mental health professionals evaluating patients in the field previously were accompanied by police under the Marty Smith law — named after a crisis responder who was stabbed to death by a patient he was trying to commit. But in the wake of a new law passed last summer, numerous police departments in Washington have determined they are no longer allowed to forcibly detain someone for mental health reasons — and are now refusing to do so, even when the detention is requested by a judge or mental health professional.
What happens to suicidal, seriously ill, and gravely disabled people when both police and mental health professionals stop responding?
A second problem arises from the fact that the line between criminal investigation and crisis intervention is often unclear. Professor Jerry Ratcliffe, a criminologist at Temple University, studied calls to police in Philadelphia last year and found that “some medical or public health activity initially masquerades as crime or other policing work and some events eventually determined to be police/crime activity can initially appear to be public health related. About 20% of activity in this area does not appear predictable from the initial call type as handled by police dispatch.”
I suspect Professor Ratcliffe’s findings are actually a bit of a lowball. San Francisco police report that 42% of the mental health detentions they conducted in 2020 “originated in no-crisis calls for service such as a physical fight, person with a knife call, vandalism, [or] assaults.”
Finally, when it comes to suicidal persons, disorderly conduct, and non-violent crime, there is remarkably little evidence supporting the idea that social workers can be replaced with police. The most frequently touted police alternative is a program in Eugene (OR) called Crisis Assistance Helping Out On The Streets (aka CAHOOTS). CAHOOTS has become so popular that other cities are spending hundreds of thousands of dollars on similar programs and the original CAHOOTS is now selling “consulting services.” But the evidence that CAHOOTS or similar programs can actually replace police is remarkably weak:
CAHOOTS’ contract with the City of Eugene requires employees to summon a police officer anytime “a dangerous situation appears to exist.”
Of the calls CAHOOTS handled with no police assistance, basically none were reported crimes and less than 10% involved suicidal, disorderly, or intoxicated subjects. One-fourth of them were just people asking for a ride:
The other most frequent calls for CAHOOTS include situations where the “caller specifically requests CAHOOTS” and “non-emergency service requests from the public” including “counseling”, “injury evaluation after a person declined to be evaluated by a medic”, and “providing general services.”
When CAHOOTS is dispatched to calls involving a report of criminal activity, they frequently call for police backup—for example, when CAHOOTS is sent to criminal trespass calls, police backup is requested roughly one-third of the time.
To be clear, I have nothing against CAHOOTS. I’m fine with mobile crisis teams and having social workers available on-call 24/7. But these services complement a police response — they aren’t a replacement for it.
The root of the problem
So it’s not true that police are gratuitously using violence against mentally ill people, and it’s not true that police only recently started dealing with the mentally ill. So why does this keep ending up on the news? Because deinstitutionalization has put many, many more dangerously ill people out on the street than ever before.
A while ago, I wrote about the death of Daniel Prude. Prude was high on PCP and delusional. Rochester police safely detained him and took him to the hospital. Just four hours later, the hospital cut him loose. Prude then stripped off his clothing and started destroying property in the middle of the night. When police detained him a second time, he died of excited delirium. The cops got it right the first time, and if he’d been kept in the hospital just a little while longer, he would still be alive.
This is, and always has been, the real mental health crisis facing police departments in the United States. In San Francisco, one individual has been detained by police for mental health reasons an astonishing 23 times in just one year — and returned to the street every single time.
Under these circumstances, it really doesn’t matter how good at de-escalation and crisis intervention your cops are. San Francisco police can — and do! — resolve 99.9% of mental health calls without violence. But when essentially every seriously ill person safely detained is immediately released, a tragedy is inevitable. And sending a social worker isn’t going to change that.
Comments Archive (6/30/22):
A friend who is a social worker is perfectly willing to answer mental health calls, but she insists on a police escort. She works in a County office sorting accused criminals, and she has a lively understanding of how dangerous these people can be. I don't want to see social workers, however well trained, sent alone into these situations. Police officers are armed for a reason.
A very well-argued and researched article. I'm a social worker who's run in disability advocacy/deinstitutionalization circles for a few years now, and there's lots here to mull over, so thank you very much for that.
What are your thoughts on having social workers employed directly by police departments -- not to respond to calls, but more for victim advocacy, mental health support for officers, etc.? I'm thinking an approach like the one discussed here. Do you believe this is a more appropriate way for social workers to collaborate with police departments, and do you foresee any drawbacks to this approach?
I am all in favor of having social workers work with police departments. I think this is what makes CAHOOTS work. It’s really, really helpful as police to be able to make the scene safe and then hand off social work (this person needs housing, medication, counseling) etc. to take over in that regard.
Support for officers, I’m less sure. I think if police departments want to support officer mental health they need to hire outside therapists who don’t answer to the agency command.
Retired LEO. My experience mirrors the author. I also agree the biggest problem for the police is the follow up by the mental health system. Over the years, there were too many individuals that I had taken in for mental health evaluation (5150 W&I CA) that were just released back to the streets in a few hours just to continue with their ongoing mental crises. Eventually these same people would end up in jail after committing a minor theft or physical assault.
The mental health professionals are needed on the back end, not the front end or the issue.
The numbers of mental health calls you quote for three cities are stunningly high. Do you have the number of total calls to the police for these cities? Just trying to put these numbers in context. Do mental health calls represent 30% of total calls or 3%? Just curious.
Couldn't find data for SF. DC, as of 2015, about half a million calls a year (probably gone up a fair amount since then). Seattle in 2019, 460,000. So probably about 5-10%.
About the 99%+ success rate and startling small percentage of things going sideways, Neil DG Tyson made a counterpoint that we should be outraged the success rate isn't even higher, especially when cops/unions almost always investigate themselves and even cover things up. He points out that airplanes crash once out of 4 million flights but it's such a scandal when a plane does crash. On top of that, there's an independent body (FAA) that investigates crashes. Similarly, teachers who are anonymously accused of sexual harassment can face the loss of their careers and an investigation where they're guilty until proved innocent. Police are trusted with far more than the public's education. Given the weight on responsibility placed on police, should we be striving for that standard and outraged that we aren't?
A study was coincidentally released yesterday on NYC's pilot B-HEARD program, showing that "Mental Health Response Teams Yield Better Outcomes Than Police In NYC": In 95% of cases, people accepted care from the B-HEARD team, compared with 82% for traditional 911 response teams, which include police. 50% of people treated by B-HEARD were transported to the hospital for more care, a far lower number than the 82% who are transported to the hospital with traditional 911 response. How might we be misreading the results? Do you think there's any bias or other dimensions we should consider?
Second, the calls being sent to this BHEARD program - like the calls being sent to CAHOOTS - are the lowest acuity calls. It’s not surprising the people calling are more likely to accept help, since they’re the people with the fewest behavioral issues. This is not a random sample we are comparing. Also, B-HEARD handled a grand total of about 100 calls - in a part of the NYC which gets over 8000 mental health calls a year. Can it scale? I dunno, maybe. But that’s not a lot of data at all.
I really hate the “airplane crash” analogy. I hinted at this when I discussed the success rate, but police using force is not a “failure” in the way an airplane crash is a failure. 100% of airplane crashes are bad, that’s not what airplanes are supposed to do. But police are supposed to use force. That’s the whole point of having police. So, the video above shows an example of one of the 2% of times that Seattle police used “force” on a mentally ill person, but that’s hardly a failure. It would have been a failure to let a guy who had just committed a robbery escape, or hurt some third person. Even a use of deadly force by police is not automatically a failure. If a mentally ill person is shooting up your kid’s school, you want police to shoot that guy. Of course, a social worker will never use force at all, so they’ll always use less force than even the best police officers, but the trade off is they might not get someone help who needs it and they might not be able to intervene with a person who poses a risk to others. So maybe you can eliminate that last 0.01% where force gets used, but at what cost?
The same point can be made about taking people to the hospital. The B-HEARD call “sample” excluded people the dispatcher thought would need to be involuntarily hospitalized, NYPD’s didn’t. But again, taking someone to the hospital could be good or bad. If they’re seriously ill and need to go, they should go! Especially if the alternative is leaving them in a public place where they’re going to harm themselves, someone else, or generally engage in disorderly activity.
I don't see the special skills cops have that let's them most handle these crises. And the CAHOOTS thing can be read as, a non-police solution knows how to call for cop backup when necessary. Why not partition labor so that cops can do things that only cops can do, and other responsibilities are more distributed?
A reply from a retired officer:
"I don't see the special skills cops have that let's them most handle these crises."
99% of patrol level calls for service are some kind of verbal conflict resolution. Police officers get a lot of hands on experience everyday dealing with conflict resolution, 24/7 in all environments and weather.
A ride along with any medium to large police department may better highlight the skills that cops develop during the regular course of their shift.
I am a licensed therapist that is embedded with our police department. I have been on the job for three years and worked in Child Protective Services for 8 years before that. I can tell you without any doubt that I would never go to a scene without police backup. I will not be in charge of the call until the officers have cleared the scene for the safety of the consumer/client/citizen, myself and them. I am not trained in the use of force nor do I carry a weapon. I have absolute faith in my officers to keep me safe and allow me to help my clients. If I were asked to go to a scene without an officer I would refuse.
A third reply:
Social workers can learn these skills on the job, but they need time to do it, probably years to build up an institutional framework of this new role that would sit between the cops and desk-based (for lack of a better term) social workers. And Graham's point is that if over 99% of mental health calls to cops are resolved non-violently, is it a good idea to slash police funding immediately and divert resources to this new institution that would be re-learning a lot of the skills police already have?
To me the takeaway seems to be that cops get mental health calls resolved correctly more often than is popularly perceived, and that while there's certainly room to develop stuff like CAHOOTS and experiment with ways to increase specialization of roles in how we maintain public order/enforce the law, the evidence doesn't seem to support a massive immediate change in the police's role in resolving emergency mental health crises.
These guys weren’t rookies either.