Standing before a classroom of police officers, Lt. Mark Poisson of the Wethersfield, Connecticut Police Department cues up video of a young man talking about the night he tried to get Poisson to kill him. “Seth” (a pseudonym), who was 19 at the time and attending college in New Jersey, had already attempted suicide twice. He’d never been in trouble with the law but had spent years crippled by depression, and he was searching for the best way to die. Eventually, he decided the surest method was a gun. But he didn’t own one; neither did his parents.

That’s when it came to him: Police have guns.

The plan was simple: Get in the car and drive like crazy. Eventually, a cop would pull him over, and he’d do something threatening to get the cop to shoot him. On the night of March 22, 1997, Seth dropped his girlfriend off after a college formal, popped the bottle of cheap champagne knocking around in the back seat and started driving. Fast. He tore through New Jersey and into his home state of Connecticut. It was about 3 a.m., though, and he went miles without seeing a police car. So he drove to the parking lot of the Wallingford Police Department and leaned on his horn, trying to get their attention. It worked. A chase ensued with squad cars from several cities either following Seth or attempting to cut him off. Finally, Seth rammed into a police cruiser and veered off the road. Poisson, who’d been following him, screeched to a halt and prepared for a foot chase.

But instead of running, Seth got out and started coming toward Poisson, brandishing the champagne bottle as a weapon. “I’m gonna kill you!” he screamed. “You better shoot me, or I’ll kill you!”

Poisson, one of two officers on the scene, drew his gun and started to back away. He had been threatened before; two years earlier a man he’d caught stealing from vending machines came at him with a crowbar. But this was different. Poisson saw something in Seth’s face that said he wasn’t going to drop the bottle. Poisson kept backing up, kept telling Seth to drop it, but Seth kept coming at him, screaming “I’ll kill you if you don’t shoot me!” Finally, when Seth had backed Poisson into the middle of the highway, where any car coming up the hill might wipe him out, he pulled the trigger. The bullet hit Seth in the abdomen and lodged near his spine.

The officers cuffed Seth and turned him over. Poisson knelt down beside the bleeding boy, who looked at him and said, “Shoot me in the head.”

Poisson switches off the video and motions for a woman sitting in the back of the room to come address the officers. “My name is Louise Pyers,” says the woman. “And Seth is my son.” When Seth regained consciousness after surgery in the hospital, she says, he looked at her and said, “Tell that officer I’m sorry, Mom.”

In her quest to understand what happened to her son, Pyers dove into research and soon found out that there wasn’t much to grab onto. Determined to bring focus to the issue, Pyers formed the Connecticut Alliance to Benefit Law Enforcement, recruited some sympathetic officers and set about convincing police chiefs all over the state that they needed to pay more attention to the devastating phenomenon.

It hasn’t been an easy task. There are few circumstances more terrifying for a police officer than facing a person with nothing to lose. In the past few years, suicide by cop has become not so much a problem to be solved, but a catchy phrase that has taken on a life of its own, with the media and police spokespeople floating it as a possible explanation for a wide range of officer-involved shootings.

In 2003, Poisson and Pyers began giving a four-hour “suicide by cop” presentation to select Connecticut police officers. The men and women they’re speaking to today are members of police departments across the state who have applied and been selected to take part in a 40-hour training program that will qualify them to be members of their department’s Crisis Intervention Team. Developed in Memphis, Tenn., in the early 1990s, CIT — which includes officer training as well as tight police cooperation with community hospitals, mental health workers and advocacy groups — is considered the gold standard for preparing police to effectively respond to community members with mental illness, including people bent on committing suicide by cop.

But as Pyers discovered in her research, understanding of suicide by cop is far from complete. Partly, this is because only a handful of researchers have attempted to quantify how many of the 350 to 400 people cops kill each year actually wanted to die. Depending on whose definition of suicide by cop you accept, somewhere between 35 and 120 people use the police as instruments of their own destruction every year in the U.S. But partly because of this lack of consensus, hundreds of shootings are called “suicide by cop” in the media and occasionally by law enforcement agencies looking to justify a questionable civilian death.

Only 10 percent of the nation’s approximately 15,000 police departments offer crisis intervention programs, and according to Ron Honberg, the legal adviser for the National Alliance on Mental Illness, it is often a tragic shooting — like Seth’s — that serves as the catalyst for bringing the program to a community. “Police officers are frequently not amenable to being told there’s a better way of doing things than what they’re doing,” says Honberg, who has worked on crisis intervention for two decades. “It’s an entrenched culture that doesn’t change easily. The attitude is, if there’s no fire, why throw water?”

But many in law enforcement believe there is, indeed, a fire. Nothing strains community-police relations like a fatal shooting, and with many police use-of-force incidents resulting in lawsuits — not to mention psychological trauma for the officer involved — advocates for crisis training believe the time has come for police organizations to take the problem seriously. “There is still no uniform definition of suicide by cop,” says Anthony Pinizzotto, a former FBI forensic psychologist who has done extensive research on the phenomenon. “There are no criteria by which to judge whether it’s suicide by cop, and there are no strategies to offer law enforcement to say this is how you go about investigating a suicide by cop.”

Yet.

According to many researchers of the topic, the term “suicide by cop” was coined in the early 1980s by Karl Harris, a former psychologist with the Los Angeles County Coroner’s office, but it doesn’t appear in academic literature until a decade later, and as time passed, the number of shootings called “suicide by cop” crept up. In 1996, Rick Parent, a former member of the Delta Police Department in British Columbia, found that between 10 and 15 percent of the 58 police shootings he examined could be considered premeditated suicides. The next year, researchers using data from the Los Angeles Sheriff’s Department looked at 437 officer-involved shootings between 1987 and 1997 and found that 13 percent of fatal shootings were suicide by cop situations. They also noted that, in the last year of the sample, the number jumped to 27 percent. In 1998, three criminal justice professors published an article in the FBI Law Enforcement Bulletin concluding that of the 240 police shootings they looked at, 16 percent had “probable or possible suicidal motivation,” but that 46 percent “contained some evidence of probable or possible suicidal motivation.”

This wide range illustrates some of the challenges involved in studying suicide by cop. In March 2009, Kris Mohandie, a former Los Angeles Police Department psychologist who regularly consulted with the department’s hostage team, made news when he published a study in theJournal of Forensic Sciences reporting that 36 percent of the 707 police shootings he studied were suicide by cop. Media outlets reported Mohandie’s results as fact, but some in the field raised their eyebrows.

Sean Joe, a professor of psychiatry at the University of Michigan who studies suicide and participated in a University of Virginia conference on suicide by cop in 2008, told me he thinks Mohandie’s criteria for a suicide by cop is too broad. “Just because a subject displays aggression toward the police, doesn’t necessarily mean he has the intent to die,” he says.

In his research, Mohandie broke the shootings he considered to be suicide by cop into two categories: planned and spontaneous. Seth’s suicide attempt would have fallen into the first category, which Mohandie found made up just 17 percent of the incidents he deemed suicide by cop.

The second category is more controversial. Mohandie’s research found that 81 percent of the suicides by cop he identified were incidents in which the subject — often someone feeling hopeless or self-destructive — did not decide he actually wanted to die until the police became involved. Such a definition, other researchers say, leaves open the possibility that police could have done something to escalate or even provoke the subject’s reaction — and still have the shooting called a “suicide.”

“If you get a true suicide by cop, there is no negotiating with them,” says Detective Sgt. Don Hull, who has been a hostage negotiator with the Oklahoma City Police Department for more than 20 years. “I get a gut feeling and realize, ‘This guy is going to make us shoot him.’” In his two decades in law enforcement, Hull says he’s seen fewer than five “true” suicides by cop.

James Drylie, a former New Jersey police officer who teaches criminal justice at Kean University in New Jersey, conducted extensive research for his book Copicide: Concepts, Cases and Controversies of Suicide by Cop, and explicitly left out the cases of people with a severe mental illness. He did not believe they possessed full understanding of the fact that their aggressive actions would cause their death. Mohandie disagrees with this assessment, citing the fact that people with severe mental illness, such as schizophrenia, are at high risk for suicide and suicide attempts.

Drylie developed three criteria for determining if an officer-involved shooting was a suicide by cop: The subject had to voluntarily enter into a confrontation with police; to communicate suicidal intent, either through words or gestures (like putting a gun to his head); and to act in a threatening manner toward police.

Using these criteria, Drylie spent the summer of 2005 going through the investigation files of 63 cases that officers at the FBI’s National Academy had chosen to represent suicide by cop. Only 46 percent fit his criteria, which told him that most police officers don’t really understand the phenomenon. “I truly believe that if police knew more about these situations, we would be able to mitigate the number of fatal shooting deaths,” Drylie says.

But understanding requires data.

National data on police shootings is remarkably thin. Although the FBI publishes a detailed annual report about the circumstances surrounding the deaths of police officers killed in the line of duty (an average about 50 per year since 2002), it does not routinely collect identifying information about the hundreds of people police kill. The best data available seems to be a 2001 special report by the Bureau of Justice Statistics that looked at the 8,578 people killed by police between 1976 and 1998. The report found that 98 percent were male, 55 percent were white, and 53 percent were between ages of 18 and 30. But the report does not say whether the people killed had criminal records, whether they were armed (and if so with what) or whether they had a diagnosed mental illness.

According to Parent, who researched suicide by cop in British Columbia, this lack of data has made it easier for police departments to excuse questionable shootings by saying the subject was suicidal.

Anthony Pinizzotto has been trying to change this situation. After Congress passed the Hate Crimes Statistics Act of 1990, then-Attorney General Richard Thornburgh asked Pinizzotto, who was working with the FBI’s Uniform Crime Report division, to create a way to investigate and record whether a crime was hate- or bias-related. A forensic psychologist, Pinizzotto began looking at the research on hate crimes and quickly realized that, as with suicide by cop, there was no agreed upon definition for hate crimes. So his team began meeting with departments around the country, asking about their investigative strategies and eventually developed a training guide for hate crime data collection. The next step was spreading the word to the more than 15,000 police agencies across the country. The agencies weren’t terribly welcoming at first.

“Their attitude was, ‘We have criteria for a crime, and if someone offends those rules we arrest them and let the judge and jury worry about motivation,’” Pinizzotto says. “I remember cops asking me, ‘You want us to do that now?’ The answer was yes — but we were prepared to provide them with the best information we could on how to judge that motivation.”

When Pinizzotto turned his attention to suicide by cop in the late 1990s, he used his work with hate crimes as a model. In 2005, he and his research partner, former FBI behavioral scientistEdward Davis, published a definition: Suicide by cop is “an act motivated in whole or in part by the offender’s desire to commit suicide that results in a justifiable homicide by a law enforcement officer.” They also published a simple protocol for who should investigate whether a shooting was a suicide by cop.

But Congress had given no mandate to create a definition or collect data on such incidents. To get people moving, Pinizzotto and Davis helped organize a small conference on the phenomenon in December 2008; among those in attendance was former attorney general Edwin Meese III, who became interested in suicide by cop through his work with the Law Enforcement Legal Defense Fund. In June 2009, Meese asked the FBI’s Advisory Policy Board to adopt Pinizzotto’s definition and require the bureau to begin collecting data as part of its uniform crime report. The board declined.

Audrey Honig, the chief psychologist for the Los Angeles County Sheriff’s Department, sees the problem as one of inertia. “Law enforcement are so busy fighting fires that they don’t get out in front to prevent the fire,” she says. “When you show them something new, sometimes their initial reaction is just to jerk back and say no. It’s a political thing that needs to be changed.” In 2006, Honig published a paper on suicide by cop arguing that law enforcement agencies should pay attention to the phenomenon to help mitigate their liability in police shootings.

Pinizzotto and Davis say they plan to take another run at the FBI and will continue their own research on suicide by cop. They also hope the International Association of Chiefs of Police might get on board. But when I spoke with IACP President Mark Marshall, he told me that, while he’s not opposed to collecting data on suicide by cop, it’s not high on the association’s radar. Someone, he told me, “would have to take the lead” — indicating that it wouldn’t be he.

Most people who have studied the phenomenon will tell you that, typically, suicide-by-cop scenarios fall into two categories: the “fleeing felon” who tries to escape police and, once cornered, decides he’s going to go out in a blaze of glory; and the “emotionally disturbed person,” who, like Seth, is looking for a way out of the pain of either mental illness or some kind of life failure.

The police encounter “emotionally disturbed persons” so regularly that, in cop lingo, they are called “EDPs.” Whether it’s a domestic call (a man with a history of depression has become violent because his ex won’t take him back), a workplace incident (an employee locks herself in a bathroom with a letter opener after being let go), or a schizophrenic homeless man screaming obscenities at shoppers at the local Dollar Store, police are often the first responders to problems involving our nation’s mentally ill. Situations involving the emotionally disturbed are volatile and can quickly spiral out of control, but most American police officers receive little specialized training on dealing with them.

When police do not properly understand the mentally ill — whether suicidal or not — the result can be dangerous to everyone. Officers who shoot citizens are very often haunted by the event. When the subject is suicidal, the anguish is exacerbated, because the officer can feel manipulated. According to Mohandie, about one-third of officers suffer severe post-traumatic stress disorder or depression after killing a suicidal civilian; some are so distraught they end up leaving the profession. It’s one thing to kill a “bad guy,” quite another to kill, as Louise Pyers puts it, “a sad guy.”

Despite the benefit to officers, training for crisis intervention teams and similar specialized groups is, by and large, offered only in departments with police chiefs who believe in it. According to Honburg, the director of policy and legal affairs for the NAMI, a lot of agencies are confused about crisis intervention training. “They say, we already do two or four hours on the mentally ill,” he says. “They don’t understand that CIT is about changing the culture — teaching them ways to respond that are safer for them and civilians.”

And though not a magic bullet, in some places where crisis intervention teams bloom, there is progress. Oklahoma City is one of those places.

Paul Hight was a Catholic priest who suffered from paranoid schizophrenia. On the afternoon of Dec. 14, 2000, police came to his Oklahoma City apartment after neighbors called saying he was banging on doors and ranting. Hight confronted the officers with a kitchen knife; they shot him dead.

After the shooting, Paul’s brother Joe Hight wanted answers. Couldn’t the police have done something differently? Hight, an editor at The Oklahoman newspaper in Oklahoma City, arranged meetings with the chief of police and the head of the local chapter of National Alliance on Mental Illness.

“I realized these people weren’t talking to each other,” Hight says. Like Pyers, he formed a nonprofit to bring the groups together and helped support the creation of a crisis intervention team in Oklahoma City. Today, the 600-officer department has 117 certified crisis intervention team members. Capt. Bob Nash says they aim to keep 20 to 25 percent of all officers certified, so that on every shift in every area, there is always a crisis intervention officer available when police encounter what Sgt. Keith Simonds calls “a sick consumer.”

Simonds co-taught a five-day crisis intervention course I attended in August 2008. The curriculum included everything from detailed information on the most current psychotropic drugs to site visits to halfway houses and psychiatric wards. But perhaps the most important thing officers learn in such courses is attitude adjustment.

On the first day of class, Simonds struck a typical police officer pose: feet planted firmly, hand on gun.

“Doesn’t exactly say, I’m here to help you, does it?” Simonds asks. “More like, I’m here for target practice.” Everybody laughs and Simonds explains that when a crisis intervention officer encounters someone in psychiatric crisis, he should approach with caution and patient talk: “How are you doing today, sir? Looks like maybe you’re having a bad day. Are you taking any medication?”

Time and distance are the allies of a crisis intervention team, and for police officers used to demanding and receiving immediate answers, such an approach can seem a revolutionary concept. But the advice isn’t exactly new: Florida police psychologist Dr. Laurence Miller has been writing columns for PoliceOne.com for years about an officer’s most important weapons: his brain and mouth. “Ninety percent of potentially lethal situations in law enforcement can be talked down through negotiation,” Miller says.

Joe Hight tells me he believes that if the Oklahoma City police had had crisis intervention training in 2000, his brother would be alive. Capt. Nash, who runs Oklahoma City’s crisis intervention program, agrees, saying that it helps officers sharpen communication skills and has led to a drop in citizen deaths and injuries at the hands of police. “Now that families trust us,” he says, “we’re getting more and more early calls when a loved one begins to slip into a crisis, instead of waiting to call when things have deteriorated so much that communication is difficult.” Also, he says, the concepts taught in the program — patience, empathy and persistence — have trickled down to other officers.

“A couple months ago, we had an incident where an officer was called to a house where shots had been fired,” Nash recalls. The officer hadn’t taken crisis intervention training, but on previous visits to the house — the home of a man who suffered from extreme paranoia and often called police when he was battling delusions — the officer had taken the time to sit down and listen. That experience paid off. Instead of barging in with his gun drawn and, perhaps, provoking a shooting — a shooting that might well later be classified a suicide by cop — he was able to defuse the situation.

“He made contact with the man, reminded him who he was and was able to talk him into coming outside without the shotgun,” Nash says. “The man hadn’t hurt anyone, so he didn’t need to go to jail. The officer took him to the crisis center, and he was able to get the help he needed.”

This kind of background information might have made a difference the night Lt. Poisson shot Seth. “If I had thought he might be trying to use me to commit suicide,” Poisson says, “I would have pulled up much farther away from him, so I could keep my distance and not have to make the decision to shoot.”

Seth is now in his 30s, and when he looks back on that night, he isn’t sure Poisson could have done much differently. A Taser might have been better than a bullet, but in 1997, that controversial, “less lethal” option wasn’t available.

“Maybe if he had more knowledge about people with mental disease and troubles, he wouldn’t have fired as soon,” Seth says. “But in the end, I was so hopped up and crazed I don’t know if he could have talked me down. I was committed.”

This article was reported in collaboration with The Investigative Fund at The Nation Institute, now known as Type Investigations; the John Jay/H.F. Guggeneheim Fellowship provided additional support.