There is outrage in Philadelphia’s streets after the killing of Walter Wallace Jr., who was experiencing a mental health crisis. Much energy will go into debating the technical legality of the shooting. Many will see another Black man killed by police in comparison without incidents where white people wielding weapons are taken into custody without being killed. Many will defend the police because Wallace had a knife in his hand, claiming the officers had no choice. Others will describe it as “lawful but awful” — a situation in which the use of force by police is disturbing to see but legally justified. While these are important issues, their resolution is unlikely to prevent more deaths.

Right now, we should explore how to get police out of the mental health business as much as possible.

Over the last 50 years we have witnessed the systematic dismantling of mental health services in the United States. Much of this involved the defunding of state mental hospitals that were rife with abuse. The savings achieved from these closings were supposed to go into community-based services that would allow people to live independently with dignity and security. But those community investments never materialized and millions suffer with no or inadequate support, leaving them vulnerable to homelessness, self-medication with illegal drugs and alcohol, and periodic bouts of profound crisis that can lead to incarceration, forced hospitalization, suicide, or death at the hands of police.

Between one-quarter and one-half of all people killed by police are experiencing a mental health crisis. Many cities have responded by giving police more training or developing co-response models in which police work with mental health workers. But the research on these approaches is not encouraging. A 2014 meta-analysis found no improvement in safety for officers or the public from the use of crisis intervention teams. A 2019 study showed that evidence to support the training was contradictory and, in most cases, deploying the teams didn’t affect arrest and use-of-force rates. The only time we see improvements in the outcomes of these calls is when such teams are combined with increases in community-based mental health services.

Even when police responses don’t result in death or serious injury, police have few good options in these encounters. They can arrest someone, force them to go to a hospital, or attempt to resolve the situation informally. None of this leads to sustained high-quality mental health care. The coercive nature of police involvement instead directly contributes to some of the violent outcomes, since people in crisis often fear or dislike police because of past bad encounters. In addition, many violent outcomes are the result of the tragic dynamic of “suicide by cop,” when someone in crisis summons police hoping that if they act in a hreatening manner, police will kill them. In the United Kingdom, where police response to mental health crisis calls has been rare until recently, this phenomenon doesn’t exist.

There is an alternative. In parts of Oregon, a local clinic has developed a non-police outreach capacity called CAHOOTS, integrated into the 911 system. When calls come in about crises related to the overlapping problems of mental health, homelessness, or harmful substance use, they respond with mental health and social workers. Out of 24,000 calls last year, police backup was requested in only 150 of the incidents — less than 1%. CAHOOTS teams now handle 20% of all 911 calls, saving an estimated $8.5 million in law enforcement expenses and additional dollars in diverted emergency-room visits. Similar teams are now operating in Denver, Portland, and San Francisco, and more are planned for Austin, Albuquerque, Minneapolis, and Los Angeles.

Philly We Rise and other groups have called for reducing police spending to free up dollars that could go toward non-police responses. Unfortunately, city leaders have not developed similar plans and continue to rely on officer training programs and some co-response. Commissioner Danielle Outlaw has noted that this approach is not working as planned and in a recent event at the University of Texas Law School told me that she would ideally like to see non-police response capacity increased, though she remains concerned about potentially violent encounters and, not surprisingly, opposes any reduction in police spending.

Clinical response teams have been shown to work and save cities money. But we must also look to build high-quality community-based mental health systems that help people avoid being in crisis. A recent report from the New York City Public Advocate lays out a program that starts with low-intensity, low-cost drop-in centers that will reduce the need for police response, emergency-room visits, and forced hospitalizations. Fixing our mental health system is central to any effort to reduce the burden of police violence.

Alex S. Vitale is professor of sociology and coordinator of the Policing and Social Justice Project at Brooklyn College, and author of The End of Policing.