The Community Responder Model

How Cities Can Send the Right Responder to Every 911 Call

Establishing a new branch of civilian first responders can help cities reduce overreliance on the police.

In this article
An employee answers calls at the Essex County Regional Communications Center in Middleton, Massachusetts, on June 22, 2017. (Getty/The Boston Globe/Nicholas Pfosi)
An employee answers calls at the Essex County Regional Communications Center in Middleton, Massachusetts, on June 22, 2017. (Getty/The Boston Globe/Nicholas Pfosi)

Executive summary

Today, a significant portion of 911 calls are related to quality-of-life and other low-priority incidents that may require a time-sensitive response but are better suited to civilian responders, rather than armed police officers. Some 911 calls may not require a time-sensitive response at all. Recent original analysis conducted by the Center for American Progress (CAP) and the Law Enforcement Action Partnership (LEAP) examined 911 police calls for service from eight cities and found that 23 to 39 percent of calls were low priority or nonurgent, while only 18 to 34 percent of calls were life-threatening emergencies. While many 911 calls do merit an emergency police response, unnecessarily dispatching armed officers to calls where their presence is unnecessary is more than just an ineffective use of safety resources; it can also create substantially adverse outcomes for communities of color, individuals with behavioral health disorders and disabilities, and other groups who have been disproportionately affected by the American criminal justice system.

This site is protected by reCAPTCHA and the Google Privacy Policy and Terms of Service apply.

To improve outcomes for the community and reduce the need for police response, LEAP and CAP propose that cities establish a new branch of civilian first responders, known as “Community Responders.” As envisioned, Community Responders would be dispatched in response to two specific categories of calls for service that do not require police response. First, they could be dispatched to lower-risk 911 calls related to mental health, addiction, and homelessness. This report details several existing programs that send nonpolice responders to handle such issues, including the Crisis Assistance Helping Out On The Streets (CAHOOTS) program in Eugene, Oregon. Second, Community Responders could handle calls unrelated to behavioral health needs, which might be classified as disturbances, suspicious persons, trespassing incidents, noise complaints, other quality-of-life concerns, and lower-risk neighborhood conflicts. These sorts of situations would benefit from the mediation skills and neighborhood experience that credible messengers—a type of outreach worker with a personal history of justice system involvement—already employ in violence prevention initiatives across the country.

Using 911 data from eight cities, this report estimates that between 33 and 68 percent of police calls for service could be handled without sending an armed officer to the scene; between 21 and 38 percent could be addressed by Community Responders; and an additional 13 to 33 percent could be dealt with administratively without sending an armed officer to the scene.

Leading departments are already demonstrating that police are not needed to respond to calls—including those for auto accidents without injury as well as minor larceny, theft, and burglary cases—where the primary purpose is to take reports for insurance companies. Law enforcement agencies can also filter out more false alarms and mistaken 911 calls to avoid dispatching officers unnecessarily. This report discusses the challenges and opportunities of a Community Responder approach and concludes with recommendations on how to successfully implement Community Responder programs in cities across the country—including the need to gather community input. Community Responders are not a silver bullet, but particularly among communities where relationships with police have been damaged by generations of disproportionate enforcement, the model could play an important role in increasing safety, well-being, and trust.


In today’s society, the police are responsible for much more than enforcing the law. Officers are increasingly expected to solve every problem that crops up in the community, from resolving noise complaints and reversing overdoses to disciplining school children and de-escalating behavioral health crises. Yet because police often lack the tools and training to address such concerns, these calls for service can result in unnecessary uses of force, justice system involvement, and other adverse outcomes for civilians, as well as put a strain on public safety resources. Moreover, the harmful effects from these interactions have not been felt equally by all Americans: Communities of color have disproportionately experienced heavy police presence, high rates of arrest, and unduly harsh enforcement tactics. The growth of policing has also negatively affected people with behavioral health disorders and disabilities, whose medical conditions are too often treated like a crime.

Likewise, the ever-expanding role of the police has had a negative impact on officers themselves, many of whom have attested to having too much on their plate. “Every time 911 receives a call, it’s currently the job of police to respond,” explains retired Maj. Neill Franklin, who served as head of training for the Baltimore Police Department. “But many calls don’t involve a crime. And when they do, many of those crimes are minor and related to quality-of-life issues such as homelessness, mental health disorders, or substance misuse. We need to stop expecting police to do social work and start sending the right trained professionals to address low-level crimes and noncriminal calls for service.”1

Today, officers spend a significant portion of their time responding to low-priority 911 calls for service. The Vera Institute of Justice conducted an in-depth analysis of 911 data from five localities—Tucson, Detroit, Seattle, New Orleans, and Camden—and found that nonemergency incidents were the most frequent type of call for service in each jurisdiction.2 In 2017, nearly 25 percent of calls for service in the city of Camden, New Jersey, were related to environmental conditions or nuisances, such as open fire hydrants, disturbances of the peace, and animal complaints. Another 7 percent of calls were related to medical or behavioral health needs. In total, less than 30 percent of calls for service in 2017 were related to any type of crime, with only 1.4 percent concerning violent crime.3

CAP and LEAP found a similar pattern in this report’s analysis of 911 data from eight cities—Detroit, Michigan; Hartford, Connecticut; Minneapolis, Minnesota; New Orleans, Louisiana; Portland, Oregon; Richmond, California; Seattle, Washington; and Tucson, Arizona. The share of low-priority, nonurgent calls ranged from 23 percent in Hartford to 45 percent in Seattle.4 Meanwhile, medium-priority, non-life-threatening incidents comprised a large portion of 911 calls, from 34 percent in Detroit to 42 percent in Hartford. Top-priority, life-threatening emergencies made up the smallest portion of 911 calls, from 18 percent in Seattle to 34 percent in Hartford. (see Table 1)

Sending police to low-priority calls carries significant societal costs. First, stretched thin by low-acuity 911 calls, police officers have less time to devote to serious crimes. And when police spend their days racing between 911 calls, officers have few opportunities to proactively build relationships with community members. Crucially, sending law enforcement is not the best response to resolve certain types of calls for service. The police, for instance, are regularly expected to respond to people with mental health or substance use disorders. Yet officers are not hired for their skill in managing complex behavioral health needs, nor do they have the necessary training or resources to do so effectively or safely. Because the police are not set up to provide the necessary quality of service, police response can create negative outcomes for people with disabilities and those with chronic or acute behavioral health conditions. Often, these individuals are arrested and booked into jail, which can exacerbate their medical needs.

Moreover, police officers may unintentionally escalate a situation, simply by showing up on the scene. In particular, police presence can be triggering for people with behavioral health disorders, as these individuals are more likely to have experienced negative or traumatic contacts with the justice system. In fact, according to the International Association of Chiefs of Police, “the mere presence of a law enforcement vehicle, an officer in uniform, and/or a weapon … has the potential to escalate a situation” when a person is in crisis.5 Police may not understand how a particular disability manifests and may assume that an individual’s reaction—or lack of reaction—to law enforcement is a show of defiance. All too often, these calls for service result in unnecessary fatalities: People with serious mental health disorders are 16 times more likely than the general public to be killed during a police encounter.6

Dispatching the police to intervene in situations where their presence is not necessary is particularly damaging for communities of color. Calls for service related to minor incidents are more likely to result in justice system involvement for Black people, who are disproportionately arrested for low-level and noncriminal infractions such as loitering, having an open container, or disorderly conduct.7 The consequences of these enforcement actions can last a lifetime for individuals, their families, and their communities, since even a minor arrest record can create permanent barriers to housing, education, and employment. Researchers have also documented substantial racial disparities in the use of excessive force: Police use of force is among the leading causes of death for Black men and boys, who are 2 1/2 times more likely than their white peers to be killed by law enforcement.8 Excessive use of force and its disparate impact on the Black community is an issue that every jurisdiction must urgently confront through law, policy, and culture change.

In addition to reforming police culture and policy, cities can help prevent unjust killings by avoiding unnecessary armed response to 911 calls. Many of the tragic killings of Black Americans began with a 911 call that did not require an armed first responder: George Floyd was killed by police officers responding to a call about a counterfeit $20 bill;9 Atatiana Jefferson was killed while playing video games at home after a neighbor called the police to report that her door was ajar;10 Rayshard Brooks was killed after police received a call that Brooks was asleep in his parked car;11 Michael Brown was killed during a police response to shoplifting allegations;12 and Elijah McClain was killed when officers responded to a call that McClain “looked sketchy” while walking down the street.13

Relying on police to handle low-level calls for service has other long-term consequences. Aggressive enforcement of low-level offenses and unnecessary negative police interactions not only affect residents’ health and well-being but also erode public trust in police.14 Loss of trust can have serious ramifications for public safety, including a significant reduction in the likelihood that residents will report crime to law enforcement.15 In 2005, for example, a Black man named Frank Jude was violently assaulted by law enforcement while attending a party with officers in Milwaukee. When the public learned about it, 911 calls for police service dropped precipitously. In Milwaukee’s communities of color, call volumes remained low for more than a year.16

In effect, residents are unlikely to report crime or provide valuable information to a police force perceived as untrustworthy, making it harder for officers to prevent and solve serious crimes, including homicide. Over the past decade, less than half of all murders of Black Americans resulted in an arrest, let alone a conviction.17 Because these offenses often go unaddressed, residents may feel that law enforcement will not protect them. Instead, they may rely on other means to resolve issues—including interpersonal violence.18

The Community Responder model

To strengthen outcomes for the community and law enforcement alike, LEAP and CAP propose the establishment of a new branch of civilian first responders, known as “Community Responders.” Community Responders (CRs) could be dispatched for two broad categories of calls for service that do not always require police presence: First, CRs could respond to calls related to homelessness, behavioral health crises, and substance use—calls that might currently be classified by local dispatchers as “wellness checks,” “disturbances,” “intoxicated persons,” or “mental crises.” Second, CRs could respond to calls related to quality-of-life concerns and low-level community conflicts that do not require a behavioral health intervention, including many calls currently classified as “suspicious persons,” “disorderly conduct,” “noise complaints,” “juvenile disturbances,” or “trespassing.” These two groups of low-acuity calls for service would be more effectively addressed by Community Responders, given their training as well as their personal and professional backgrounds. With a focus on long-term solutions, CRs can help address today’s low-level concerns before they become tomorrow’s emergencies.

LEAP and CAP’s analysis of 911 calls for service in eight cities across the United States suggests that a large percentage of calls could be handled by Community Responders. Estimates for the share of calls that could be handled by CRs range from a low of 21 percent of calls in Detroit to a high of 38 percent in Seattle and Portland. Most of the calls classified as appropriate for CRs were assigned low priority by dispatchers—from 36 percent in Tucson to 85 percent in Detroit. Medium-priority calls comprised anywhere from 14 percent of CR-appropriate calls in Detroit to 58 percent of such calls in Tucson.19 The most common 911 call types that would be appropriate for CRs include disturbance, suspicious person or vehicle, trespassing, intoxicated person, nuisance, noise complaint, adolescent disturbance, wellness check, and mental health crisis.

Behavioral health and social service calls

Community Responders could offer a more appropriate response to the many calls for service that are related to medical and social service concerns, including behavioral health needs, substance use disorders, and homelessness. For these calls, an effective CR would be an individual with a background as paramedic, a clinician, or a crisis intervention specialist. CRs should also include peer navigators, who have personally experienced the challenges facing people involved in these 911 calls and can provide valuable perspectives to outreach teams. In the context of homelessness, for example, research has demonstrated that peer outreach workers are uniquely positioned to connect with and support unhoused individuals, based on their shared experiences with housing insecurity.20

A number of jurisdictions have already adopted nonpolice response programs that incorporate elements of the proposed Community Responder model. Among the most prominent is the CAHOOTS program based in Eugene, Oregon. Since 1989, CAHOOTS has sent teams of medical and crisis intervention professionals in response to behavioral health- and other service-related 911 calls.21 CAHOOTS specialists are trained to provide a wide range of services within the community, including wellness checks, behavioral health crisis interventions, substance use-related de-escalation, family conflict mediation, and basic medical treatment.22 And importantly, CAHOOTS offers connections and transportation to social services to help support clients’ long-term needs. Though the CAHOOTS team is dispatched through the same 911 system as the Eugene Police Department, their goal is to create an alternative to police response for people experiencing social service and behavioral health needs whenever possible.23 The team responded to more than 24,000 calls in 2019—roughly one-fifth of total 911 dispatches—and only called for police backup in 1 percent of those calls.24 The model has also proven cost-effective. According to estimates from the Eugene Police Department, it costs roughly $800 per call to dispatch the police.25 CAHOOTS has substantially cut back on the need for police response, saving an estimated $8.5 million in taxpayer dollars every year.26

In recent years, other cities have taken note of Eugene’s success with civilian first responders. CAHOOTS served as a model for the Support Team Assisted Response (STAR) program in Denver and the Crisis Response Unit (CRU) in Olympia, Washington.27 Like CAHOOTS, STAR and CRU dispatch civilian crisis responders for calls related to behavioral health crises, substance use disorders, and other social service needs. Both programs include peer responders who have lived experience with the concerns they are responding to, making them uniquely credible and effective in engaging difficult-to-reach individuals.28 According to peer navigators, they can truly “walk in the shoes” of people whose actions might cause other responders to give up.29

In addition to addressing behavioral health concerns, existing programs show that CRs could address an even broader range of low-risk calls for service. In Rochester, New York, the Family Crisis Intervention Team (FACIT)—a team of trained social service professionals and mediators—responds to calls related to domestic disputes, youth behavioral challenges, landlord-tenant problems, death notifications, and child abuse.30 FACIT was originally housed within the Rochester Police Department; however, following the killing of Daniel Prude—a 41-year-old Black man who died of asphyxiation after police restrained him during a mental health wellness check—the city announced plans to transfer FACIT out of the police department. FACIT will now be housed within the Rochester Department of Recreation and Youth Services’ newly created Crisis Intervention Services Unit, a civilian office established to strengthen and coordinate non-law enforcement responses to public safety needs.31 The Crisis Intervention Services Unit will be available 24/7 to respond to calls for service through the 911 dispatch system.

Other jurisdictions around the country have implemented programs similar to civilian first responders. (see Appendix) Some programs send responders only through a nonemergency phone number, rather than dispatching calls through 911. Other jurisdictions divert calls to a virtual clinician instead of an in-person responder or provide long-term follow-up to high utilizers of emergency services but do not respond to calls for service.

Quality-of-life and conflict calls

Community Responders could also provide an effective response to 911 calls related to nuisance complaints and nonviolent conflicts. For these calls, however, a Community Responder’s qualifications would differ from those necessary to address behavioral health and social service needs. Rather than clinicians and crisis interventionists, these CRs should be “credible messengers”—individuals with strong ties to the community, oftentimes with a personal history of overcoming violence or justice system involvement, who are able to connect to community residents based on their shared background and experiences. Currently, credible messengers play an integral role in many successful violence interventions. Oftentimes, they serve as outreach workers, known as “violence interrupters,” trained to mediate conflicts before they turn violent. Interrupters spend much of their time building relationships with the small group of individuals at highest risk of violence and staying abreast of the dynamics within the neighborhood. When tensions arise, interrupters have the credibility and the skill set to diffuse the situation and encourage healthier responses to conflict.

Relying on their training and connections within the neighborhood, credible messengers have proven successful at engaging residents at highest risk of violence and defusing potentially dangerous situations. In Sacramento, California, for example, researchers found that credible messengers intervened in nearly 70 instances of imminent gun violence and mediated 220 conflicts over a two-year period from July 2018 to April 2020, contributing to a 14.3 percent decline in firearm violence in the target communities.32 Importantly, the Community Responder program would supplement—not supplant—existing violence interventions that rely on credible messengers. CRs would respond only to low-level conflicts and nuisance complaints, while violence interrupters would retain their focus on engaging the individuals at the highest risk of violence.

Credible messengers are uniquely qualified to respond to certain types of low-level community conflicts that do not require an armed officer or a behavioral health or social service professional. Their portfolio could include some of the leading types of calls for service, including disturbances, noise complaints, suspicious persons, disputes between neighbors, youth behavioral issues, conflicts in schools, trespassing, and even some simple assaults that do not involve weapons. Like many police officers, credible messengers would be trained to de-escalate conflict and administer basic medical care. However, their authority would be rooted in their connection to the community and their credibility with their neighbors, rather than a badge and a firearm.

As credible messengers, CRs will likely prove more effective than the police at addressing conflicts and engaging residents in solutions, particularly those residents who distrust law enforcement. As low-income communities and communities of color continue to experience disproportionately high rates of arrest, incarceration, and aggressive enforcement tactics, many residents are disinclined to engage with officers and, in some cases, to even to call 911 in the first place. In a survey of low-income, predominantly Black neighborhoods across the country, more than 60 percent of residents indicated that they do not feel safe around the police and 30 percent said that they were unlikely to call the police to report suspicious activity in their neighborhood.33 If provided with an emergency response option that does not involve the police, residents may feel safer reporting issues as they arise. And as trusted community members with extensive training, CRs can offer an effective and appropriate resolution to concerns that do not require police response.

In many communities, credible messengers are already doing this work on the ground. One example comes from Brooklyn’s East New York neighborhood, a predominantly Black community whose residents have been affected by aggressive policing tactics and “stop-and-frisk” encounters that unjustly target communities of color.34 East New York is home to the nonprofit organization Man Up! Inc., whose credible messengers have been able to resolve minor community issues without unnecessarily involving the police.35 When a group of young people were caught on camera stealing from a local business, that business called Man Up! Inc. rather than the police. Shneaqua “Coco” Purvis, a site manager with Man Up! Inc., tracked the young people down and told their parents, and they returned the items to the store without police involvement. “Instead of … putting them in jail or putting these kids through the system, police should just give us a call and we’ll handle it,” Purvis said in an interview with a local Brooklyn newspaper.36

Other examples have emerged during the COVID-19 pandemic. In communities nationwide, credible messengers have played a valuable role in combating misinformation about the virus, urging residents to adhere to requirements to wear a mask and practice physical distancing. Not only do these efforts protect residents’ health, they also hold the potential to reduce unnecessary police response and enforcement actions for public health violations, which have been disproportionately levied against communities of color. Early reports revealed that more than 80 percent of the people ticketed by the New York Police Department for violating physical distancing mandates were Black and Latinx,37 despite national-level data demonstrating that Black and Latinx Americans are more likely than white Americans to observe public health safety measures.38 To limit unnecessary enforcement actions, the city of Newark, New Jersey, has dispatched credible messengers to respond to reports of young people gathering in violation of public health mandates.39 With their ties and standing in the community, credible messengers are able to use incidents where health measures are violated as a “teachable moment,” educating young people on the importance of taking public health precautions.

However, even where credible messengers are already working on the ground, local governments are not systematically dispatching them to low-level calls for service. The vast majority of calls go to 911, and there is no pipeline to direct appropriate calls to credible messengers. With a CR model, however, crisis communications centers could directly refer calls to CR teams that would include credible messengers. And by connecting 911 to these teams, local leadership would send a clear message to residents and government agencies alike: Credible messengers are a valuable part of the city’s approach to public safety.

Establishing a CR model can institutionalize and expand efforts to deploy credible messengers to address nonemergency incidents, reducing unnecessary arrests for minor infractions that do not require police response. As discussed above, 911 calls in communities of color are disproportionately likely to lead to aggressive searches, citations, arrests, and officer-involved shootings. CRs alone cannot solve disparities in policing, but by dispatching CRs in lieu of officers for low-acuity calls for service, cities can reduce police contacts that might result in arrests, uses of force, and other negative interactions in communities of color.

Administrative opportunities to reduce police response to 911 calls

In addition to implementing a Community Responder model, 911 data reveal other opportunities to substantially reduce the volume of calls for service to which police must respond. In many jurisdictions, police spend an inordinate amount of time responding to callers who immediately hang up or say nothing on the line, as well as silent alarms from home and business security systems, which are nearly always false. Police must also respond to calls made primarily for insurance purposes—accidents without injuries; minor theft from purses, cars, and homes; and destruction of property—as well as 911 calls that come in for other city services such as animal control.

Police response to 911 hang-up and no-voice calls could be greatly reduced through telephone screening. In cities such as Baltimore, these calls are directed to a Telephone Reporting Unit that redials the callers to determine the nature of the situation.40 In most instances, Telephone Reporting Units confirm that these calls are false alarms that do not require a police response. To cut down on silent alarm responses, police departments can require third-party confirmation of the call. For almost 10 years, the city of Camden has responded only to silent alarm calls that are confirmed by security companies.41

Agencies use a variety of methods to avoid patrol officer response to insurance-driven calls. The city of Baltimore relies on online forms and the Telephone Reporting Unit to take reports for no-injury auto accidents, hit-and-runs, destruction of property, theft, and burglary. Likewise, Camden requires residents to report similar types of issues online or by coming into the station to speak with an investigator. Tucson also encourages online and in-station reporting of these issues, offering appointments for convenience.42 When situations require an official presence on scene, Tucson sends nonsworn “community service officers,” who can also handle runaway young people, blocked driveways, and road hazards.43

This report’s analysis of 911 calls in eight cities suggests that a significant portion of police calls for service could be handled by administrative alternatives without sending an armed officer to the scene. LEAP and CAP estimate this share of calls to range from 13 percent in Detroit, which already uses a telephone reporting unit, to 33 percent in New Orleans. Most of these calls are low-priority, from 54 percent in Hartford to 93 percent in Detroit.44 The majority of them are insurance-driven—for example, car accidents or larceny—and a significant portion are parking-related. Hartford also showed a significant share of calls for service related to alarms and 911 hang-ups. All cities could reduce the need for armed police response through proven administrative alternatives.

How the Community Responder model works

Few people know what happens when they call 911. When a caller dials 911, the phone company determines their location based on the address associated with the landline. If the caller is using a cell phone, their location is determined by the nearest cell phone tower.45 The call is then transferred to the appropriate local 911 center, known as a “public safety answering point,” where it comes up on a call-taker’s screen. In most large cities, the call-taker collects key information from the caller and discerns and enters into the system the call’s priority level, call type or event codes, and other notes. The call file is then sent electronically to a police or fire dispatcher, who typically sits a few feet away from the call-taker. Call files wait in the dispatch queue based on priority level, during which time the call-taker can continue to ask the caller questions and amend the call notes.46

Under the Community Responder model, to divert 911 calls to CRs, jurisdictions could embed a CR dispatcher alongside the 911 call-takers and police and fire dispatchers. Houston, for example, uses this setup for the Crisis Call Diversion (CCD) program, which relies on clinicians to resolve nonemergency mental health calls for service over the phone.47 As calls come into 911, call-takers would flag situations that might be appropriate for CR response. At the same time, the CR dispatcher would continuously review the calls in the police and fire dispatch queues—in particular, the calls flagged by the call-takers—and reach out to certain callers when more information is needed. If a CR dispatcher determines that a call is appropriate for CR response, they would radio the CRs to confirm the appropriate CR team to dispatch to the scene, before removing the call from the police dispatch queue. En route to the scene, the CR team could dial the caller to gather additional details prior to arrival—a practice that patrol officers themselves should adopt.

This system would allow dispatchers to ensure an appropriate response to each call. Importantly, call-takers would determine whether each call involves a medical emergency or a weapon, since CRs should respond only to non-life-threatening calls. The CR dispatcher would screen for potential mental health factors and other root causes underlying the situation. At any point, if the CR dispatcher or CR team decides that the call needs a police response, the CR dispatcher could update the call file and return it to the police dispatch queue.

Community members could also reach the CR dispatcher through a direct number to increase reporting. Many people wait to call 911 until there is a true emergency or a crime has been committed, or they do not call at all, because they do not want to involve the police. Having a direct number for the CR dispatcher would encourage community members to call earlier while situations are easier to de-escalate and before crimes are committed. In Olympia, most calls to the Community Responder Unit (CRU) program come directly from service providers and other groups that do not want to involve the police, rather than through 911. To maximize reporting, it is important to inform the community at large, as well as service providers, businesses, and city agencies such as transit and libraries, about how to reach CR responders.

Community Responders would approach each incident with an eye toward long-term solutions, rather than finding a quick fix and moving onto the next call. A key piece of this work would involve actively facilitating connections to key supportive resources, which might include domestic violence prevention services, crisis stabilization centers, and sites for independent living. The CRs would have a vehicle available to transport individuals to service providers, which has been a key asset for the CRU program in Olympia. For example, if a person in crisis is at risk in their current environment, CRs could provide transportation to 24-hour crisis centers, which serve as a drop-off point for a warm handoff to service providers.

Other community institutions, such as overdose prevention sites, are crucial to avoiding crises in the first place. Sometimes referred to as safe injection facilities, these sites provide a safe and hygienic space for individuals to bring drugs that they are intending to use. Sites are staffed by specialists who prevent overdose death and disease transmission while building supportive relationships with clients that can help stabilize their lives.48 Establishing these facilities would greatly reduce police and ambulance calls for overdoses and public drug use.

While waiting for calls to come in, CRs could do proactive engagement. The CRU team in Olympia spends most of its time on self-initiated preventive efforts rather than responding to calls. For example, CRU checks in with frequent clients up to three times a week, ensuring that they are able to refill prescriptions and keep up with doctors’ appointments and providing other proactive support to prevent crises before they occur.

Olympia takes an even more preventive approach by partnering CRU with Familiar Faces, a long-term follow-up program. When CRU has several interactions with the same individual, it can recommend their enrollment in the Familiar Faces program. Familiar Faces builds long-term relationships with clients, helping them access housing, health care, and other supportive services and following up to maintain progress. Peer support is the backbone of the Familiar Faces program: Clients are identified, engaged, and supported by peer navigators. They relate to clients without judgment and with a personal understanding of their experience. “I’ve actually not only been where they’re at, but I’ve dug even deeper,” said Keith Whiteman, a Familiar Faces peer navigator.49

Tailoring the approach

Each city will need to tailor the CR model to fit its local context and to most effectively meet the needs of its residents. Thus, when designing a CR model, government officials should meaningfully involve the community and engage residents in every step of the process. In particular, local leaders must engage residents who are most likely to come into contact with a CR, whose input can help refine the CR model to most effectively address community needs. As part of this effort, cities should ensure that engagement strategies are accessible and inclusive of diverse communities and audiences. Instead of conducting large public forums at city hall, for example, cities might host a series of smaller meetings at community-based locations, where residents may feel more at ease participating in dialogue. The city might also recruit a team of resident leaders from across the city to jointly develop and refine the CR program. Resident leaders will be better positioned to solicit feedback from their neighbors, which can provide valuable insights for shaping the model’s development. And once the program reaches the implementation stage, resident leaders can help educate their communities on the purpose and role of CRs.

Perhaps the most significant decision in customizing the Community Responder program for a particular community is shaping the relationship between CRs and law enforcement. The Community Responder model could be rendered less effective if residents perceive CRs as being too closely affiliated with the police. Importantly, CRs with close ties to law enforcement may face additional barriers to connecting with clients. In Olympia, for instance, the CRU program is housed with the Olympia Police Department. Because of their affiliation with the police, CRU had to work to gain the community’s trust, explained Anne Larsen, who oversees CRU: “One of the biggest things we had to overcome is the idea that we would be snitches,” she explained in an interview with The Marshall Project.50

A close connection to the police could undermine one of the chief benefits of the CR model: its ability to attract direct calls from people who otherwise would not have called 911. In Olympia, most of the external calls for CRU responders come not through 911 but to the direct CRU phone number from service providers and community members who do not want to call 911. For example, many outreach programs focused on individuals with substance use disorders keep their distance from police in order to maintain the trust of their clients, who may avoid law enforcement out of fear of arrest or harassment. Most violence prevention programs deliberately limit relationships between credible messengers and police to protect their credibility with clients, who are at highest risk of engaging in violence and often distrust law enforcement. CRs will be less effective with residents who distrust law enforcement if they are perceived as working hand in glove with the police—in which case these residents may avoid calling CRs altogether.

However, Community Responders will likely need some relationship with police. At a minimum, CRs would want to summon police quickly if a situation escalated or proved more dangerous than it seemed on the surface. But this does not occur often: The STAR program in Denver has not had to call police for backup once, and the CAHOOTS team needed police backup on less than 1 percent of calls in 2019.51

At the same time, CRs can also benefit from their relationship with the police. In Olympia, a positive working relationship with the police has helped expand CRU’s reach: When law enforcement arrives on a scene that would be more appropriate for CRU, officers call in the CRU team to handle the call. In fact, CRU receives more calls from officers on scene than from 911. In Denver, 40 percent of STAR calls come from police on scene rather than from 911 directly. Because officers are more likely to refer calls to CRs if they have established trust in the program, it can be valuable for CRs to build and maintain a trusting relationship with police. Moreover, if CRs had the option of police providing backup or arriving first and checking a scene for weapons, they could safely handle a broader range of calls.

CRU is also able to access police records, providing insight into clients’ history of law enforcement contacts. However, it is vital that information sharing is never used for enforcement purposes—and that clients trust that CRs will maintain that distinction. “It’s about reassuring folks that we don’t run [their names] for warrants or anything like that,” emphasized Larsen.52 No matter their relationship with police, CRs should strive to reduce justice system contact.

Communities can design their CR program’s relationship to police based on local circumstances. They can tailor-fit a model that balances the benefits of reputational separation from police with the need for adequate police support.

Funding and implementation

The CR model offers a cost-effective alternative to police response to certain calls for service. With the introduction of the CR program, cities can expect to see a substantial reduction in the need for police response—and, with it, substantial cost savings. Eugene’s CAHOOTS program, for example, saves an estimated $8.5 million per year in public safety costs by reducing the need for police response.53 As savings accrue, cities should reinvest funds back into the CR program to expand its reach and promote sustainability.

The CR program should have a permanent place within a government agency to help ensure that it has access to dedicated city funding streams. CRs should be full-time government employees, on par with other branches of first responders. By offering CRs full benefits and fair compensation, local government agencies can help to secure the program’s impact by reducing staff turnover.

Local leaders should consider housing CRs within an Office of Neighborhood Safety—a government agency dedicated specifically to civilian-led approaches for strengthening public safety and well-being. These agencies can provide the necessary infrastructure to support and sustain nonpunitive public safety strategies, including civilian first responders, violence interruption, mentoring, reentry services, and more. Offices of Neighborhood Safety are now popping up in cities across the country as government leaders increasingly recognize the value of community-driven approaches to public safety.54

Like other first responders, CRs should receive extensive training in the skills required for their job. While the specific training needs will vary based on different calls for service, trainings might cover conflict mediation, de-escalation, harm reduction, restorative justice, cognitive behavioral therapy, motivational interviewing, and basic medical care. CRs must also receive training and ongoing resources to support their own emotional and physical well-being, with a focus on helping CRs cope with vicarious trauma and stress associated with their day-to-day responsibilities.

In Eugene, CAHOOTS responders complete 40 hours of classroom training and more than 500 hours of field training, guided by an experienced mentor.55 Trainings focus on de-escalation and crisis intervention while emphasizing the safety of first responders. CAHOOTS provides “extensive training in scene awareness and compassionate communication and verbal de-escalation to prevent things from becoming physical,” according to Tim Black, CAHOOTS’ program manager.56

While jurisdictions should strive to provide a dedicated funding stream for CRs, local leaders may need to explore private funding sources to support the early implementation of the program, especially as governments are facing pandemic-related budget deficits. Localities can expect to save money in the long term, but private funding could help cover startup costs and support initial operations while savings begin to accrue.

Local leaders may also consider piloting the CR program in a specific neighborhood or police district before rolling out the model citywide. A pilot program would help drum up local support for the model and enable cities to compare outcomes in similar areas with and without CRs. Promising evaluations can demonstrate that CRs are a valuable investment for the city, helping local leaders to make a compelling case for expanding the CR model into other areas. As part of the program’s growth, cities will need to employ enough CRs to keep response times short. And importantly, as the model expands into different communities, the city must recruit CRs with specialized neighborhood experience, background, and training to address different types of calls.

Next steps communities can take to move forward on Community Responders

In communities across the country, support for civilian first responders is on the rise.

Groups interested in Community Responder programs should consider the following steps as they move forward:

  • Do your homework: Collect and analyze data that can help demonstrate the need for CRs, including data on 911 calls for service. There is no standardized procedure across jurisdictions for collecting 911 call data, nor is there a centralized database for tracking call data nationwide. Nonetheless, most jurisdictions collect and analyze 911 call data, and many jurisdictions now make these datasets available online.57 This report includes analysis for most cities that make 911 calls-for-service data publicly available in a usable format. Other jurisdictions publish data that mix 911 calls for service with officer-initiated activities such as traffic stops and business checks. Groups may want to submit Freedom of Information Act (FOIA) requests to the local police department to obtain 911 calls-for-service data, including call type and priority level. Collecting data from residents is also vital: Community surveys and interviews can help pinpoint the types of calls best suited for CRs and demonstrate public demand for the CR model.
  • Identify and gain support of local stakeholders: Determine which groups can be brought on board to support Community Responders. In addition to city leadership and law enforcement, important stakeholders include behavioral health agencies and major service providers, city institutions such as transit and libraries, police reform and racial justice advocacy organizations, harm reduction nonprofits, and community associations. It is crucial to mobilize support from individuals and groups in neighborhoods with the highest 911 call volumes—in particular, credible messenger groups already developing skills in mentoring and de-escalation. The CR plan should be framed differently for each stakeholder so that they understand how it will serve their own interests.
  • Learn from experience: Engage residents in shaping the CR model. Residents are experts in their own neighborhood and can help refine the CR model to maximize effectiveness. Groups interested in the Community Responder model can also gain valuable insight from the experiences of peer jurisdictions who are undertaking related work. As mentioned in this report, a number of jurisdictions are implementing civilian first responder programs that share elements with the Community Responder model. Finally, the Law Enforcement Action Partnership and Center for American Progress offer our own local connections to law enforcement, city officials, and advocacy groups in order to help bring Community Responders to neighborhoods across the country.


Today, local elected officials are increasingly facing calls to reduce reliance on police to respond to calls for service.58 Indeed, data clearly demonstrate that law enforcement is often dispatched to address 911 calls that could be more appropriately handled by civilians—a situation that is detrimental to community members and officers alike. By adopting a Community Responder model, cities have an opportunity to reduce the volume of calls to which police must respond, mitigating the negative impacts that can result from an overreliance on policing. Community Responders, with their intensive training and lived experiences with the issues facing communities, are uniquely positioned to provide an effective solution to answering calls for service that do not require a law enforcement response.

Ultimately, a Community Responder program operates with a vision of helping communities resolve problems internally without the justice system. Public safety is about more than just policing; when residents experience crime or disorder, they should have tools beyond the police at their disposal. Americans understand this: A recent survey found that nearly 8 in 10 voters support diverting mental health- and substance use-related 911 calls to trained, nonpolice responders.59 The goal of the CR model is to ensure that every neighborhood has the resources to settle more disturbances and disputes and find long-term solutions outside of the justice system to recurring public safety issues.

As with any new initiative, the Community Responder model may face initial skepticism. Yet the best evidence for the model is to see it in action. Six months after the CRU model was implemented in Olympia, a survey of law enforcement showed that many veteran officers had already gone from skeptics to champions of the model. One officer explained that he had been called to respond to the same individuals over and over for 20 years, and he assumed he would be doing it until he retired. Suddenly, the CRU responders not only took those calls off of his plate but also helped find long-term solutions that worked for the people involved, so much so that the community no longer called about them.60

Promisingly, city leaders are recognizing the imperative shift away from the status quo. The police are “handling these calls the best they can,” said San Francisco Fire Department Capt. Simon Pang, who is leading the city’s effort to develop a team of civilian first responders. But “law enforcement officers [are] responding to nonviolent, noncriminal calls for service for people whose needs are largely social, behavioral or mental. And that’s just not right … It’s glaringly obvious we need to change the model.”61

As more and more Americans recognize the need for change, it is our hope that jurisdictions nationwide will join the movement to bring the Community Responder model home.

About the authors

Amos Irwin is the program director for the Law Enforcement Action Partnership.

Betsy Pearl is an associate director for Criminal Justice Reform at the Center for American Progress.


Appendix: Existing civilian responder models

A number of jurisdictions have adopted civilian first responder programs that include some elements of the Community Responder model. Several of these programs were discussed in the report—including CAHOOTS in Eugene, STAR in Denver, and CRU in Olympia. Additional examples are discussed below. This list is not meant to be exhaustive but instead to provide a sample of the landscape of existing civilian responder programs and illustrate their differences from the Community Responder model.

The following programs include a civilian response to behavioral health and substance use needs. Unlike CRs, these programs are not connected to the 911 call system and focus only on calls related to underlying behavioral health needs.

  • The Community Response Team (Washington, D.C.): The Community Response Team (CRT) is an initiative housed within the city’s Department of Behavioral Health that dispatches social workers, clinicians, and peer outreach workers in response to behavioral health and substance misuse issues. CRT is available 24/7 via a nonemergency phone line to provide services such as overdose prevention and reversal, on-the-spot behavioral health assessments, referrals to long-term supportive resources, and ongoing engagement with individuals with unmet health needs.62
  • Baltimore Crisis Response Inc. (Baltimore): Baltimore Crisis Response Inc. (BCRI) operates the Here2Help Hotline, staffed by trained counselors 24 hours a day.63 If counselors determine that emergency assistance is needed, they can dispatch the Mobile Crisis Team, comprised of mental health specialists, social workers, and nurses.64
  • Mobile Crisis Outreach Team (Salt Lake City): The Mobile Crisis Outreach Team (MCOT) dispatches social workers and peer support specialists to respond to mental health calls that come in through a 24/7 hotline, rather than directly from 911.65 MCOT is a free service operated by the University Neuropsychiatric Institute and offers follow-up supports, including referrals to other health care providers and other social services.66
  • Mobile Crisis Intervention Services (State of Connecticut): Connecticut’s Mobile Crisis Intervention Services is available via 211—the state’s nonemergency alternative to 911—and sends clinicians to support young people experiencing behavioral health needs or crises.67 The state’s Department of Mental Health and Addiction Services operates a similar service for adults, which is reachable via a crisis phone line.68 The crisis service lines can dispatch mobile crisis teams when necessary and offer referrals for continuous treatment and behavioral health care.

The following programs divert 911 calls for service related to behavioral health or social service needs to trained civilians. Unlike CRs, these programs primarily address issues over the phone without an in-person response and focus only on calls related to underlying behavioral health needs.

  • Crisis Call Diversion program (Houston): Houston’s Crisis Call Diversion program—a nonpolice diversion program for 911 calls—redirects calls related to mental health and social service needs to trained counselors who conduct immediate over-the-phone mental health assessments and make connections to appropriate social services and community-based resources. The CCD is a collaboration between the Houston Police Department, the Houston Fire Department, the Houston Emergency Center, and the Harris Center for Mental Health and IDD.69
  • Crisis Call Transfer (Pima County, Arizona): Pima County Sheriff’s Department implemented a new program in partnership with Arizona Complete Health that will triage mental health-related 911 calls. If the 911 call-taker categorizes the call as a mental or behavioral health crisis, they will transfer the caller to the Arizona Crisis Line, which will connect the caller with a trained crisis specialist. If the specialist is unable to de-escalate over the phone, they can dispatch the Crisis Mobile Team to assist the caller.70
  • Behavioral Health Call Diversion (Abilene, Texas): When 911 dispatchers in Abilene receive a call related to mental health needs, they are now able to push a button to start a three-way call with a clinician, who triages the callers’ needs and determines whether the issue can be addressed over the phone or requires an in-person response.71 The program is a partnership between the Abilene Police and Fire departments and the Taylor County Sheriff’s Office, as well as the Betty Hardwick Center, a local mental health service provider, and Avail Solutions, an organization that staffs crisis hotlines.



LEAP and CAP began by identifying jurisdictions with open datasets for 911 calls for service. Jurisdictions were prioritized for analysis if their data included the following variables: description of call type, call priority level, and whether the call was initiated by civilians or by law enforcement. The authors limited the data analysis to calls for service in 2019. The sole exception is the city of Minneapolis, which covers calls received from July 1, 2017, to June 30, 2019. Minneapolis’ data were available only within a data visualization created by the city that analyzed call volumes, broken down by priority level and call type. They were not available for extraction or for refinement by date.72 As a result, the authors manually transcribed all data for Minneapolis.

Data were sorted by call type and priority level to identify percentages of low-priority calls and common low-priority calls. (see Tables 1 and 2) The authors then reviewed the call types with law enforcement professionals from LEAP to estimate what percentage of call types at a particular priority level could be covered by Community Responders or handled administratively. Since the call type and priority level provide only the most basic sketch of the calls, the authors acknowledge that there is great uncertainty in the percentages of calls that could be successfully diverted. These estimates are based on the assumption that a CR dispatcher would be on the crisis communications center floor, calling back those who dialed 911 to gather additional details. Estimates also assume that there are enough CR teams in operation to handle the diverted calls, which would mean that a CR team would be able to respond to calls relatively quickly when necessary, though not as fast as police patrol. These estimates of responder types were combined with the corresponding frequency of each call type and priority level to estimate the overall percentage of 911 police calls for service that could be covered by CRs and handled administratively. (see Table 3)

In the spirit of transparency and in order to continue to improve estimates as new information becomes available, LEAP and CAP are publishing a spreadsheet at containing the full list of calls for each city and all variables and formulas used for this analysis.


The authors benefited from the guidance and feedback of too many individuals to list here. Special thanks to Eddie Bocanegra, DeVone Boggan, Clint Lacey, Chief Ronnie Roberts (Ret.), Anne Larsen, Keith Whiteman, Melissa McKee, Deputy Chief Al Handy, Cmdr. Bryan Bennett, Lt. Patrick Plourde, and LEAP speakers Maj. Neill Franklin (Ret.), Sgt. Carl Tennenbaum (Ret.), Maj. Mike Hilliard (Ret.), Deputy Chief Wayne Harris (Ret.), David Muhammad, Chief Paul Pazen, and Lt. Diane Goldstein (Ret.). The authors would also like to thank Sarah Figgatt, a research assistant for Criminal Justice Reform at the Center for American Progress, for her contributions to this report.


  1. Retired Maj. Neill Franklin, executive director, Law Enforcement Action Partnership, personal communication with author via email, October 8, 2020, on file with author.
  2. S. Rebecca Neusteter and others, “Understanding Police Enforcement: A Multicity 911 Analysis” (New York: Vera Institute of Justice, 2020), available at
  3. Ibid.
  4. Note: Minneapolis, Richmond, and New Orleans were excluded here because they do not group calls into low-, medium-, and high-priority categories.
  5. IACP Law Enforcement Policy Center, “Responding to Persons Experiencing a Mental Health Crisis” (Alexandria, VA: 2018), available at
  6. Doris A. Fuller and others, “Overlooked in the Undercounted: The Role of Mental Illness in Fatal Law Enforcement Encounters” (Arlington, VA: Treatment Advocacy Center, 2015), available at
  7. American Civil Liberties Union of the District of Columbia, “Racial Disparities in D.C. Policing: Descriptive Evidence From 2013–2017” (Washington: 2019), available at; American Civil Liberties Union of New Jersey, “Selective Policing: Racially Disparate Enforcement of Low-Level Offenses in New Jersey” (Newark, New Jersey: 2015), available at; American Civil Liberties Union, “Picking Up the Pieces: A Minneapolis Case Study” (New York: 2015), available at
  8. Frank Edwards, Hedwig Lee, and Michael Esposito, “Risk of being killed by police use of force in the United States by age, race–ethnicity, and sex,” Proceedings of the National Academy of Sciences of the United States of America 116 (34) (2019): 16793–16798, available at
  9. Matt Furber, Audra D. S. Burch and Frances Robles, “What Happened in the Chaotic Moments Before George Floyd Died,” The New York Times, June 10, 2020, available at
  10. Marina Trahan Martinez, Nicholas Bogel-Burroughs, and Dave Montgomery, “Woman Was Playing Video Game With Her Nephew When Shot by Fort Worth Police,” The New York Times, October 13, 2019, available at
  11. Minvonne Brooks, “911 call before Rayshard Brooks’ fatal shooting by police is released,” NBC News, June 16, 2020, available at
  12. Rachel Clarke and Christopher Lett, “What happened when Michael Brown met Officer Darren Wilson,” CNN, November 11, 2014, available at
  13. Lucy Tompkins, “Here’s What You Need to Know About Elijah McClain’s Death,” The New York Times, August 16, 2020, available at
  14. American Public Health Association, “Addressing Law Enforcement Violence as a Public Health Issue” (Washington: 2018), available at
  15. Giffords Law Center to Prevent Gun Violence, “In Pursuit of Peace: Building Police-Community Trust to Break the Cycle of Violence” (San Francisco: 2020), available at

  16. Matthew Desmond, Andrew V. Papachristos, and David S. Kirk, “Police Violence and Citizen Crime Reporting in the Black Community,” American Sociological Review (2016): 1–20, available at
  17. Giffords Law Center to Prevent Gun Violence, “In Pursuit of Peace.”
  18. Ibid.; Stephen Lurie, “There’s No Such Thing as a Dangerous Neighborhood,” Bloomberg CityLab, February 25, 2019, available at
  19. Minneapolis, Richmond, and New Orleans were excluded here because they do not group calls into low-, medium-, and high-priority categories.
  20. Calgary Homeless Foundation, “Dimensions of Promising Practice for Street Outreach Supports in Ending Homelessness” (Calgary, Canada: 2013), available at
  21. White Bird Clinic, “Crisis Assistance Helping Out On The Streets” (Eugene, OR: 2018), available at
  22. Ibid.
  23. Ibid.
  24. Ibid.
  25. Ibid.
  26. Ibid.
  27. Jessica Porter, “New program diverts some 911 calls from police to a mental health team,” Denver 7, July 2, 2020, available at; Angela Ufheil, “The STAR Van Offers an Alternative to Police,” 5280, June 17, 2020, available at; Urban Strategies Council, “Report on Feasibility and Implementation of a Pilot of Mobile Assistance Community Responders of Oakland (MACRO)” (Oakland, CA: 2020), available at; Christian Hill, “Local crisis unit in cahoots with more police agencies,” The Register-Guard, December 14, 2018, available at
  28. Maya Chiodo, “Caring for Denver Issues First Grants Worth About $2 Million,” 5280, February 4, 2020, available at; Denver Police Dept., @DenverPolice, June 11, 2020, 9:05 p.m. ET, Twitter, available at; RI International, “Community Support Teams,” available at (last accessed October 2020); City of Olympia, “Crisis Response & Peer Navigators,” available at (last accessed October 2020).
  29. Keith Whiteman, peer navigator, Familiar Faces, personal communication with author via phone, September 3, 2020, on file with author.
  30. 211 Lifeline, “FACIT Program,” available at (last accessed October 2020).
  31. City of Rochester, New York, “City Announces Creation of New Crisis Intervention Services Unit To Better Serve the Needs of the Community,” Press release, September 16, 2020, available at; Aaron Ross Coleman, “The killing of Daniel Prude, explained,” Vox, September 10, 2020, available at
  32. Jason Corburn and Amanda Fukutome-Lopez, “City of Sacramento/Advance Peace Sacramento Youth Peacemaker Fellowship Program: CalVIP, BSCC Final Local Evaluation Report” (Berkeley, CA: University of California, Berkeley, 2020), available at
  33. Nancy La Vigne, Jocelyn Fontaine, and Anamika Dwivedi, “How Do People in High-Crime, Low-Income Communities View the Police?” (Washington: Urban Institute, 2017), available at
  34. Andrew Gelman, Jeffrey Fagan, and Alex Kiss, “An Analysis of the New York City Police Department’s ‘Stop-and-Frisk’ Policy in the Context of Claims of Racial Bias,” Journal of the American Statistical Association 102 (479) (2007): 813–823, available at; New York City Bar Association, “Report on the NYPD’s Stop-and-Frisk Policy” (New York: 2013), available at
  35. Andre Mitchell, “Man Up! Inc. pushes to rid New York City of crime and violence for good,” New York Daily News, January 3, 2018, available at; Jessy Edwards,” An Alternative to Cops With Guns in NYPD ‘Occupied’ East New York,” BK Reader, July 20, 2020, available at
  36. Edwards, “An Alternative to Cops With Guns.”
  37. Mark Morales and Elizabeth Joseph, “Blacks and Latinos are overwhelmingly ticketed by NYPD for social distancing violations,” CNN, May 9, 2020, available at
  38. Stella M. Rouse, “New data finds Black and Hispanic Americans more likely to take precautions against coronavirus,” The Washington Post, August 10, 2020, available at
  39. David Muhammad and DeVone Boggan, “The Very Essential Work of Street-Level Violence Prevention,” The Trace, April 14, 2020, available at
  40. Baltimore Police Department, “Policy 506: Telephone Reporting Unit (TRU),” February 12, 2020, available at
  41. Deputy Chief Albert Handy, Camden County Police Department, personal communication with author via phone, August 27, 2020, on file with author.
  42. S. Rebecca Neusteter and others, “The 911 Call Processing System: A Review of the Literature as it

    Relates to Policing” (New York: Vera Institute of Justice, 2019), available at

  43. Ibid.
  44. Minneapolis, Richmond, and New Orleans were excluded here because they do not group calls into low-, medium-, and high-priority categories.
  45. Neusteter and others, “The 911 Call Processing System.”
  46. Retired Maj. Mike Hilliard, Baltimore Police Department, personal communication with author via phone, August 10, 2020, on file with author.
  47. Houston Police Department Mental Health Division, “Crisis Call Diversion Program (CCD),” available at (last accessed October 2020).
  48. Betsy Pearl and Maritza Perez, “Ending the War on Drugs” (Washington: Center for American Progress, 2018), available at
  49. Drew Mikkelsen, “Olympia hires former inmates for homeless outreach,” KING 5 News, May 28, 2019, available at
  50. Christie Thompson, “This City Stopped Sending Police to Every 911 Call,” The Marshall Project and The Daily Beast, July 24, 2020, available at
  51. Elise Schmelzer, “Call police for a woman who is changing clothes in an alley? A new program in Denver sends mental health professionals instead.”, The Denver Post, September 7, 2020, available at; Anna V. Smith, “There’s Already an Alternative to Calling the Police,” Mother Jones, June 13, 2020, available at
  52. Thompson, “This City Stopped Sending Police to Every 911 Call.”
  53. White Bird Clinic, “Crisis Assistance Helping Out On The Streets.”
  54. Betsy Pearl, “Beyond Policing: Investing in Offices of Neighborhood Safety” (Washington: Center for American Progress, 2020), available at
  55. Urban Strategies Council, “Report on Feasibility and Implementation of a Pilot of

    Mobile Assistance Community Responders of Oakland (MACRO).” .

  56. Emily Green, “The need for a better first response on Portland streets,” Street Roots, March 15, 2019, available at
  57. Neusteter and others, “Understanding Police Enforcement.”
  58. Ed Chung and Betsy Pearl, “How To Reinvest in Communities When Reducing the Scope of Policing” (Washington: Center for American Progress, 2020), available at
  59. Alliance for Safety and Justice, “Toward Safety: The First Ever National Survey of America’s Safety Gaps” (Oakland, CA: 2020), available at
  60. Anne Larsen, outreach services coordinator, Olympia Police Department, personal communication with author via phone, August 26, 2020, on file with author.
  61. Eric Westervelt, “Removing Cops From Behavioral Crisis Calls: ‘We Need To Change The Model’,” NPR, October 19, 2020, available at
  62. D.C. Department of Behavioral Health, “Community Response Team,” available at (last accessed October 2020).
  63. Baltimore Crisis Response Inc. “Here2Help Hotline,” available at (last accessed October 2020).
  64. Baltimore Crisis Response Inc. “Mobile Crisis Team,” available at (last accessed October 2020).
  65. University of Utah Health, “Community Crisis Intervention & Support Services,” available at (last accessed October 2020).
  66. Optum, “Mental Health Services Hits the Road with New Mobile Crisis Unit,” March 12, 2012, available at
  67. 2-1-1 of Connecticut, “Mental Health Crisis Intervention Services/Mobile Crisis Intervention Services – Connecticut,” available at (last accessed October 2020).
  68. Connecticut State Department of Mental Health and Addiction Services, “Crisis Services,” available at (last accessed October 2020).
  69. Houston Police Department, “Crisis Call Diversion Program (CCD),” available at (last accessed October 2020).
  70. Joshua Morales, “PCSD Communications Center implements crisis call transfer,” KOLD News 13, September 17, 2020, available at
  71. Laura Gutshke, “New Abilene 911 program helps callers in mental health crisis,” Abilene Reporter-News, February 1, 2019, available at
  72. City of Minneapolis, “Public 911 Call Problem Types,” available at, (last accessed September 2020).

The positions of American Progress, and our policy experts, are independent, and the findings and conclusions presented are those of American Progress alone. A full list of supporters is available here. American Progress would like to acknowledge the many generous supporters who make our work possible.


Amos Irwin

Program Director

Law Enforcement Action Partnership

Betsy Pearl

Former Associate Director

Center For American Progress

This site is protected by reCAPTCHA and the Google Privacy Policy and Terms of Service apply.