Search About Newsletters Donate
Support independent, nonprofit journalism.

Become a member of The Marshall Project during our year-end member drive. Our journalism has tremendous power to drive change, but we can’t do it without your support.

Analysis

Sending Unarmed Responders Instead of Police: What We’ve Learned

There are more than 100 response teams nationwide, but experts say more research on their impact is needed.

A crisis response team of three people, wearing cargo pants, turquoise shirts and fanny packs, talk to a medium-dark-skinned man sitting on a concrete wall. The man, wearing gray sweatpants, holds a water bottle in one hand.
Members of the Holistic Empathetic Assistance Response Team (HEART), an alternative to police in Durham, N.C., talk with a resident, in June 2024.

In the four years since George Floyd’s murder, many sweeping attempts to reform policing have faltered. But one proposal that has taken hold across the country, and continues to spread, is launching alternative first response units that send unarmed civilians, instead of armed officers, to some emergencies.

In Dayton, Ohio, trained mediators are dispatched to neighbor disputes and trespassing calls. In Los Angeles, outreach workers who have lived through homelessness, incarceration or addiction respond to 911 calls concerning people living on the street. In Anchorage, Alaska, trained clinicians and paramedics are showing up to mental health crises.

This article was published in partnership with Tradeoffs.

As many programs transition out of the pilot phase, they face new challenges: How do they scale up and become big enough to meaningfully reduce the presence of police? How do cities sustainably fund these new agencies and find the right people to staff them? And are these teams on track to become the sea change in public safety that was promised?

“The Fifth Branch,” a podcast series from The Marshall Project and Tradeoffs, examines the results in Durham, North Carolina. (Listen to the podcast.) We also talked to experts across the country — some who are running these programs, some who are researching them, and some who are supporting them. Here’s what they said:

These programs are spreading.

For decades, Eugene, Oregon, was the rare city that sent unarmed crisis workers and EMTs to 911 calls. Now, researchers have tracked over 100 alternative crisis response units operating across the U.S. Over half of the country’s largest cities have created such teams.

How many there are depends on how you define them. Some distinguish between mobile crisis teams, which exclusively send clinicians to mental health emergencies, and community responder programs, which send civilians to a wider range of calls. The key tenets are that they can be the first response to an emergency situation and that they arrive without armed officers. (Many cities also send clinicians alongside police in what’s known as a co-response model.) But the kinds of specialists who are being sent out, and the kinds of calls they respond to, vary significantly.

Not sending one of these teams could be a civil rights violation.

In Washington, D.C., social service nonprofit Bread for the City has sued the city, claiming that sending police to mental health emergencies discriminates against people with mental health disabilities. The city started a program in 2021 to send mental health providers to those kinds of calls. But according to the lawsuit, just 327 calls were referred in the 2022 fiscal year, less than 1% of eligible calls. A similar lawsuit is ongoing in Washington County, Oregon, which includes part of Portland.

“The reason why the police response is so harmful may not be because there’s an excessive force incident — it’s because you’re not getting effective treatment,” said Michael Perloff, interim legal director for the ACLU of D.C. and one of the lawyers on the case. “If you called the EMTs for your broken leg and they sent someone who didn’t know how to set a broken bone, that’s denying you effective care. People with mental health crises, that’s their experience with emergency response services.”

The D.C. program may have been limited by overly strict criteria. At one point, calls would not be transferred to the team if they involved someone under 18, or someone who had ingested alcohol or drugs, the lawsuit states. Even if a call was deemed eligible, the team was frequently unavailable, and police were sent instead.

In an email, a spokesperson for the D.C. Department of Behavioral Health, which runs the city’s Community Response Team, noted that the team had recently increased its number of eligible call types and decreased the number of exclusionary criteria. In a motion to dismiss the case, lawyers for the city emphasized that they are working to improve mental-health emergency services but that "the creation of new services and standards of care presents a policy argument, not a legal claim.”

The U.S. Justice Department has weighed in. In February, department officials submitted a filing in the case, arguing that the Americans with Disabilities Act does apply to emergency response systems.

That filing echoed similar findings from the Civil Rights Division’s investigations into policing in Minneapolis, Louisville, Kentucky and most recently, Phoenix, where police and the city were found to be discriminating against people with behavioral health disabilities. In one instance, a 911 operator in Phoenix failed to refer a call from a mother worried about her 15-year-old daughter in distress to a mobile crisis team. When police showed up, they tackled and handcuffed the daughter and booked her in juvenile detention.

Getting 911 dispatchers on board is essential.

One challenge to scaling up these programs is dispatchers, who are often trained to err on the side of sending police. Cities have designated specific call categories to be sent to community response teams, and often use a “decision tree” of questions to determine whether a situation is safe for alternate response. But those questions take time, and 911 call centers are often overworked and under-resourced. And determining what is “dangerous” is highly subjective.

“Now there are way more layers of decision-making being added on to figure out, is this the type of call that gets police at all?” said Jessica Gillooly, a professor of sociology and criminal justice at Suffolk University in Boston, where she studies 911 dispatch. Gillooly worked as a call taker for two years and saw firsthand the “when in doubt, send [cops] out” mentality.

“In general it’s not a fear about an entire category of calls, it’s more fear based on one incident,” said Daut’e Martin with Law Enforcement Action Partnership, an advocacy organization of police, prosecutors and others working on criminal justice reform. “If you’re a call taker who’s worked for 20 years, you know about that one time an officer showed up to the noise complaint and someone started shooting.”

Cities are looking for ways to get dispatchers on board. In Chicago, the city’s Crisis Assistance Response and Engagement (known locally as CARE) team sends a quarterly update to dispatchers with data on diverted calls and examples of people connected to services. In Durham and other cities, mental health clinicians are embedded within the 911 call center to help determine the proper response.

So far, these programs rarely need to call for police backup.

There have been no known major injuries of any community responder on the job so far, according to experts. And data suggests unarmed responders rarely need to call in police. In Eugene, Oregon, which has operated the Crisis Assistance Helping Out On The Streets (known locally as CAHOOTS) response team since 1989, roughly 1% of their calls end up requiring police backup, according to the organization. Albuquerque responders have asked for police in 1% of calls, as of January. In Denver, the Support Team Assisted Response (STAR) had never called for police backup due to a safety issue as of July 2022, the most recent data available. In Durham, members of the Holistic Empathetic Assistance Response Team (HEART) reported feeling safe on 99% of calls.

Many communities are still sending alternative responders to a narrow subset of calls, and debating whether it’s safe to expand their scope. For example, many cities will only send community responders to situations that are outdoors or in public spaces. Programs are also divided on whether disputes between neighbors or within families are a proper place for crisis responders, or calls involving suicidal threats.

Some programs have avoided using 911.

Many people remain leery of dialing 911 in a crisis, especially when there’s no guarantee that someone will get an alternative responder instead of police. In a survey of hundreds of residents in Portland, Oregon — where the Street Response team has been operating since 2021 — nearly half said they did not feel safe calling 911. That rate was even higher among Black and Latino respondents.

Instead of being dispatched through 911, programs in cities like Atlanta have opted to use a non-emergency line. That way, if a call coming through 311 is deemed better suited for police, the caller can make the decision whether to transfer to 911 or hang up.

“People may have hesitancy about calling 911 because it might result in a police dispatch that they don’t want,” said Moki Macias, executive director of Atlanta’s Policing Alternatives & Diversion Initiative (PAD). Macias notes there’s a tradeoff in this approach. Fewer people know to call 311, meaning the team may miss a number of potentially eligible calls.

But the separation from the police was important, she said. “Community members were asking for an actual alternative,” Macias said. “It has been really important to us to have that clear definition of our team as entirely consent-based. That also means PAD responders cannot hospitalize someone against their will.

Some mobile crisis teams that exclusively send trained specialists to mental health crises are being dispatched through 988, the national mental health hotline launched in 2022. Often, they require the caller to be the person who needs help or a family member, rather than a bystander. Most community responder programs, which work on a wider range of calls, are not connected to the national hotline.

Reaching their full potential will require more money and more staff.

Studies suggest that eventually, a large portion of current police work could be handed off to alternative responders. A 2020 review of 911 calls in eight major cities estimates that up to 68% of calls “could be handled without sending an armed officer,” according to a report by the Center for American Progress and the Law Enforcement Action Partnership.

Getting there will take buy-in from multiple government branches, a sizable and steady budget, and time. In Albuquerque — where the city created an entirely new Community Safety Department in 2021 — roughly 5% of police calls were being diverted as of September 2023. In Durham, roughly 1% of calls have been diverted from police to the city’s two-year-old crisis team so far, staffed by mental health clinicians, peer support specialists and EMTs. In Eugene, the CAHOOTS team answered around 17% of the city’s police department calls as of 2019, the most recent data available.

These are among the most established programs in the country. Most units are operating on a significantly smaller scale. Few programs operate 24/7, and many don’t yet cover their entire city.

Many programs have struggled to find sustainable funding. Some used federal money from the pandemic-related American Rescue Plan — money that is running out. Others have tried to tap into Medicaid funding, but that carries restrictions. Cities that have set aside money in their general funds have to find a way to grow these budgets in the face of deficits. New York City, for example, has halted plans to expand its mobile crisis team after significant budget cuts. The team currently operates in less than half of the city’s police precincts.

Experts also say these jobs need to be a viable career path, and that cities need to ensure there is a pipeline of people to staff the new agencies. The CAHOOTS team in Eugene has faced a staffing crisis for at least the last three years, said Mike Yoshioka, director of programs for the White Bird Clinic, which operates CAHOOTS. The positions require specialized skills, were previously paid much less than other first responders, and can easily lead to burnout as they face increasing demand. Recently, the organization increased pay for responders, which Yoshioka hopes will help with hiring.

We don’t yet know the full impact of this type of response.

There is still significant research to be done on the impact of these alternative response teams. A 2022 study on Denver’s STAR team found a 34% drop in low-level crime in neighborhoods where the team was operating, compared to neighborhoods where it had not yet been rolled out. The impact of these programs on other metrics, such as involuntary commitments, arrests and police use of force, is yet to be seen. Studies on these are ongoing in Durham, Madison, Wisconsin and other cities.

There’s limited data on how communities feel about these new programs. A survey of over 600 Durham residents found that 57% were more likely to call 911 because of the city’s HEART program. A larger survey of Durham residents’ views on HEART is currently underway.

Macias from Atlanta noted that without increasing other social services, like housing and mental health care, the teams risk being merely a Band-Aid for people in crisis. “There’s a lot of pressure put on these response agencies without the local government building up the destination and infrastructure that people need,” she said. “Sure, expand these response teams 24/7, but make sure that at 11 p.m. on a Sunday there’s actually a place to bring somebody to get their needs met. Otherwise you’re just making yourself feel better. But the problem’s not being solved.”

Christie Thompson Twitter Email is a staff writer reporting on mental health, solitary confinement, and prison conditions. Her investigative series with NPR examining violence in double-celled “solitary confinement” won a George Polk Award for Justice Reporting and was a finalist for an IRE Award and the John Bartlow Martin Award.