Shifting Police Functions

May 2021

Police officers in contemporary American society are on the frontlines of all of society’s ills. They are trained to keep the peace and enforce the law, but in reality, they are required to be street-level problem solvers, responding to people in all manner of trouble or crisis. Police are asked to be skilled in de-escalating family conflicts, hold the expertise of mental health clinicians, navigate situations involving drug use and overdoses, and know the range of community treatment and housing services that exist to serve people in need of support. In recent months, calls to offload police functions – and to defund the police accordingly – have increased, especially within communities of color. People in such communities say that they are over-policed for minor infractions or crisis situations that too often result in a use of force or an arrest. These perceptions are reflected in the data: while arrest rates for disorder offenses have plummeted in recent years, along with disparities in such rates, Black arrest rates for disorder offenses remain substantially higher than White arrest rates (see Policing by the Numbers). These interactions not only lead to Black people being arrested disproportionately, but can also prompt the use of unnecessary force. Specialized training and co-responder models aim to help officers serve in their varied roles or to team them with trained professionals to do so. Efforts to offload some share of police responsibilities are designed to allocate these roles to people who are better situated to handle them. This shift reduces the amount of contact officers have with members of the public in situations that carry a higher potential for disproportionate arrests and unnecessary uses of force and increases the amount of time officers have for the most serious incidents. Revisions and alternatives to the 911 system have been proposed to facilitate both co-responding and the offloading of police functions.

Summary Assessment

  • Co-responder models hold promise but are under-researched. Key challenges include ensuring that co-responders are available around the clock to respond to calls in partnership with police and clarifying the role non-police responders should play in situations that involve weapons.

  • Mental health-related calls represent a small share of both calls to police (between 1% and 4%) and officer time spent on calls (2%). This suggests that co-responder models may be feasible for many agencies, but also means that offloading all such calls is unlikely to substantially reduce the footprint of police. The primary benefit of these alternative responses is improving the alignment of services to those in need.

  • Non-police response to mental health crisis calls also may reduce fatal police shootings, given that roughly one in four people killed by police has a mental health condition.

  • The evidence on Crisis Intervention Team (CIT) programs is thin, in part because these programs vary widely, with some representing basic officer awareness training and others composed of full-fledged and well-funded co-responder programs. However, the evidence on the impact of de-escalation training, which includes instructing police in how to identify and respond to people in crisis, is relatively strong.

  • Traffic stops and enforcement consume approximately 18% of officers’ time, and some of these activities could potentially be handled by unarmed officials. But American models of alternate traffic enforcement are almost non-existent, and thus there is no research about their potential impacts on traffic safety, responder safety, or criminal activity.
  • Making improvements to the call triage functions of 911 systems and implementing differential response mechanisms are likely to be more effective than relying on alternative systems to divert calls to non-police responders.

  • Cutting police budgets without analyzing the share of calls and encounters associated with offloaded duties may produce unintended consequences, such as reduced funding for needed training and enhanced accountability mechanisms.

Police are asked to do a lot – arguably far too much. Some share of people they encounter are involved in repeated domestic altercations, are wanted on outstanding criminal warrants, have mental health challenges, suffer from serious substance use disorders, or are chronically homeless. Unfortunately, police recruit training is not sufficient to prepare officers adequately to take on this complex role, and some argue that it is unreasonable to expect them to do so.

The following are three strategies to address this monumental challenge:

  1. Alter the way police officers are trained to increase their preparedness for these difficult and deep-seated societal challenges.
  2. Team officers with professionals who possess the knowledge, skills, and experience to address them.
  3. Offload some share of police officer responsibilities to people who are better trained to deal with them and have chosen a professional path to do so.

A fourth, longer-term strategy can only be noted briefly here. Local governments could do much more to alleviate the longstanding social and economic deprivations in communities designated as “high crime” areas. Local elected officials could address the stressors of poverty and lack of opportunity that contribute to the desperation behind much street crime with targeted investments in affordable housing, eviction prevention, homeless services, mental health services, living wage employment, environmental protection, youth employment and recreation, public health, arts programs, and other initiatives that support the safety, security, and well-being of residents. State and federal resources could augment tight local budgets; such investments are long overdue.

Training Police to Interact with People in Crisis

While examples exist of training that prepares officers to deal with children at domestic violence or crime scenes and use a trauma-informed approach to crime victims, documentation of such programs is scant and evaluation research is virtually nonexistent. The vast majority of police training programs specific to community members in need pertains to those with mental health issues and, to a lesser extent, people with intellectual or developmental disabilities or who are living with autism or other spectrum disorders. Police officers routinely encounter people with mental health conditions in the course of their work, and they are often the first to be dispatched to 911 calls involving people experiencing mental health crises. Given that roughly one in four people killed by police has a mental health condition, programs that train officers on how to respond to people in crisis hold the potential to save lives. According to a 2016 survey, 41 of the 42 responding states reported having training standards for police for interacting with people who have mental health issues. Most states focus on academy training and, to a lesser degree, in-service training, but few states reported having specialized training standards specific to people with mental health conditions. Many of these trainings are characterized as Crisis Intervention Team programs, but such programs are loosely defined. As such, one jurisdiction’s model might be limited to officer awareness training, while another’s program might involve a robust collaboration and co-responding model with police officers alongside mental health professionals. The One Mind Campaign seeks to ensure successful interactions between law enforcement and people in crisis. Agencies that join the campaign commit to partner with community health organizations, develop crisis intervention policies, and train all officers in mental health awareness. Since its launch in 2016, over 550 agencies have taken the pledge.

Given that roughly one in four people killed by police has a mental health condition, programs that train officers on how to respond to people in crisis hold the potential to save lives.

Only one study employing a randomized controlled trial design has been conducted on the impact of training police to recognize and interact with people with mental health challenges. Officers in England were randomized to receive either routine training or specialized one-day mental health awareness training. Researchers found that the mental health training improved officer identification of people with mental health conditions, in that the officers were more likely to accurately record encounters involving people with mental illness. Several pre-post studies have examined the impact of the officer training component of Crisis Intervention Team programs on officer attitudes and behaviors, but these suffer from the inability to control for competing factors. A key methodological challenge in evaluating such training is the inability to disentangle the impact of officer training from other aspects of these programs, such as teaming officers with co-responders and offering robust alternatives to jail detention (see Crisis Intervention Team training). A randomized controlled trial study of the impact of training officers on awareness of people with autism spectrum disorders found statistically significant differences between treatment and control groups in knowledge of autism and ways to identify and respond to it, but no indication that participants had fully mastered the training material. The study was not designed to evaluate the impact of the training on officer behaviors in the field.

Mental Health First Aid for Public Safety (MHFA-PS) is an adaptation of Mental Health First Aid (MHFA), a widely disseminated program to raise awareness on the part of the public of the signs and symptoms of mental health disorders and better equip them to aid people in need. MHFA-PS is a relatively new training specific to public safety personnel, but its content is very similar to the original program and is the curriculum recommended by the One Mind Campaign. While no evaluative research on MHFA-PS has been published to date, a systematic review of 18 experimental trials of the MHFA curriculum on adult (non-law enforcement) participants found small to moderate improvements in their ability to recognize mental health disorders and their confidence in helping someone with a mental health problem, with more modest reductions in stigma. These impacts, however, were primarily within the first six months, and waned over time. By contrast, de-escalation training has a more substantial and persuasive evidence base. One randomized controlled trial found reductions in use of force, public complaints, and officer injury, while another study using a strong quasi-experimental design observed meaningful reductions in use of force. Both studies evaluated the Police Executive Research Forums' (PERF) Integrating Communications, Assessment, and Tactics (ICAT) training program, which includes a module that trains officers in how to identify people experiencing mental health crises, substance use disorders, or intellectual disabilities.

Co-Responder/Partnership Models

Multiple models exist of police teaming with social workers, psychologists, and mental health clinicians to respond to certain types of calls for service. Examples include police partnerships with mental health professionals for team interventions with children who experience, witness, or perpetrate violence; police responding alongside child services caseworkers to substance use overdose calls for which children are present; and officers partnering with mental health clinicians for calls involving people who are perceived as a threat to their own safety or that of others. This latter model, which sometimes appears under the loosely defined “Crisis Intervention Team” heading, is by far the most common and may include programs with case managers embedded in police agencies to identify high utilizers of police and emergency services. One randomized controlled trial study of a co-responder model in Roanoke (VA) found very few differences between treatment and control groups on outcomes such as the number of mental health-related calls for service or the length of time co-responder teams spent on each call. However, the study design was compromised by the fact that only 35.5% of people assigned to the treatment group consented to receive treatment. Both extensive literature reviews and systematic reviews highlight the dearth of experimental research on the impact of co-responder models on key outcomes and the lack of evaluation research altogether on co-responder models serving people with intellectual or developmental disabilities. However, considerable research has been conducted on Crisis Intervention Teams, many of which include co-responders.

Crisis Intervention Teams

Crisis Intervention Team (CIT) programs are designed to help officers identify when people are in crisis, employ tactics that defuse volatile situations, and secure mental health services and supports for those in need. This multi-pronged strategy can include dispatch training, officer training, dedicated specialized units, and partnerships with first responder agencies, mental health service providers, and advocates. The goals of CIT are numerous. They include reducing the frequency and severity of use of force, minimizing injuries to community members and officers, lowering rates of arrest, reducing jail stays, redirecting people to the healthcare system, and improving individuals' mental health outcomes. The CIT model was established by the Memphis Police Department in the late 1980s. Since that time, CIT programs have proliferated across the country. According to the University of Memphis CIT Center, there were 2,700 CIT programs in 2019, accounting for between 15% and 17% of all police departments. Because CIT training is more common in larger, urban departments than in small, rural ones, the share of people engaging with CIT-trained officers is likely higher than 17%.

A study comparing Chicago police officers in four districts, some of whom were trained in CIT, found that training increased the number of mental health referrals only in districts where there was greater availability of mental health services. Referrals also increased in districts that had a substantial share of officers who were trained in CIT.

A meta-analysis of CIT training evaluations found no impact on arrests of individuals with mental illness or police use of force. The evidence is insufficient to determine whether CIT training is effective enough to merit widespread adoption. Moreover, the significant differences across agencies in CIT program content and implementation modality makes comparing research findings across different programs ill-advised. Descriptive literature suggests that CIT programs could improve officers’ awareness of mental health issues and their knowledge of and preference for de-escalation strategies when encountering someone in crisis. These changes may help to improve linkages to mental health services and reduce emergency room visits and repeated mental-health-related calls for service. However, a study comparing Chicago police officers in four districts, some of whom were trained in CIT, found that training increased the number of mental health referrals only in districts where there was greater availability of mental health services. Referrals also increased in districts that had a substantial share of officers who were trained in CIT. Given the lack of rigorous evaluation research on CIT programs, studies on the degree to which they are cost beneficial should be considered with a high degree of circumspection. Nonetheless, a cost-effectiveness study of the Memphis, TN, CIT program concluded that the program was associated with increased costs owing to diversion to hospital, which is costlier than jail. The cost of a hospital stay might not be the appropriate metric, however, if hospitalization ultimately delivers quality medical services that are not available in correctional settings. In contrast, a more recent study of Louisville, KY, CIT found that the program yielded net savings. Given the considerable variation in scope and cost of CIT programs, it is difficult to draw conclusions regarding their overall cost-effectiveness.

Law Enforcement Assisted Diversion

Evaluations of other co-responder models include the prominent Law Enforcement Assisted Diversion (LEAD) program, implemented in Seattle and replicated in agencies throughout the country. LEAD is primarily a diversion program intended to prevent people who unlawfully use or sell controlled substances or engage in prostitution from undergoing repeated arrests and jail stays, directing them instead to case management and supportive services. Police officers, as first responders, continue to make arrests but instead of being booked into jail, candidates for the program are first screened by an officer at the precinct level for eligibility in the program, in coordination with a dedicated case manager. In addition to diverting the arrestee from prosecution and jail, the officer and case worker collaborate to ensure that the person’s housing, healthcare, substance use treatment, and employment needs are met by referring them to resources in the community. If participants do not report to intake, repeatedly fail to comply with terms of the program, or commit new crimes the prosecutor may proceed with filing charges.

A quasi-experimental evaluation of the impact of Seattle’s LEAD program found that participants were 60% less likely to be arrested after six months, 58% less likely to be arrested for a felony, and 39% less likely to be charged with a felony than those in a matched comparison group. A subsequent study estimated that these impacts were associated with 1.4 fewer average yearly jail bookings, 41 fewer days in jail per year, and an 88% reduced likelihood of going to prison compared with control group members, along with substantial reductions in legal costs.

Participant Outcomes: Seattle LEAD program


Less likely to be arrested after six months


Fewer days spent in jail per year


Less likely to go to prison

Sources: "Seattle's Law Enforcement Assisted Diversion (LEAD): Program effects on recidivism outcomes" by Susan E Collins, Heather S Lonczak, and Seema L Clifasefi; "Seattle’s law enforcement assisted diversion (LEAD): program effects on criminal justice and legal system utilization and costs" by Susan E Collins, Heather S Lonczak, and Seema L Clifasefi.

Offloading Police Functions

Unlike the co-responder model, some programs completely offload police response to cases involving people experiencing mental health, intellectual, developmental, or substance use disorders. This approach is designed to divert certain cases to experts who are better trained to respond to them, shift or save resources, and reduce unnecessary interactions between the police and community members, thereby reducing the odds of use of force and police violence. The three types of offloading programs that have the largest bodies of research are mobile crisis teams, which respond to calls for service in lieu of the police; community-responder programs, which include responses by nonprofits and health and human services departments instead of police; and neighborhood safety offices, which are staffed by civilians to address all manner of public safety and social welfare issues with non-police responses. In some jurisdictions, efforts to offload traffic enforcement functions are also under consideration as a means of reducing the police footprint.

Mobile Crisis Teams

Mobile Crisis Teams (MCT) have been in use for several decades, the best-known example of which is Crisis Assistance Helping Out On The Streets (CAHOOTS), established in Eugene, OR, in 1989 as a community policing initiative. CAHOOTS involves a two-person crisis team, combining a nurse, paramedic, or EMT with a trained mental health professional. The unit is dispatched to scenes involving a variety of mental health-related issues, including people in crisis, domestic altercations, substance use, and suicide threats. The team’s approach is grounded in harm reduction rather than punitive measures, focusing on non-violent resolution. The impact of CAHOOTS is an open question. Some documentation suggests that it handles 17% of the police department’s call volume and only requires police backup 1% of the time, saving $8.5 million annually (according to the White Bird Clinic, which operates CAHOOTS). But an internal evaluation by the Eugene Police Department Crime Analysis Unit estimated that the diversion rate ranges between 5% and 8% of calls for police service, and that emergency police backup (involving lights and siren) was requested by the CAHOOTS mobile unit approximately 8% of the time.

The CAHOOTS model is spreading rapidly. Olympia, WA, started a similar program in 2019 and Denver instituted one a year later. One study notes that pilots have been launched in Austin; San Francisco; Portland; and Rochester, NY. Similar MCT models, while not direct adaptations of CAHOOTs, exist in Washington, DC, Baltimore, MD, Salt Lake County, UT, and throughout the state of Connecticut through its Department of Children and Families. Further adoption is poised to grow exponentially, given that the CAHOOTS model is a component of the American Rescue Plan Act of 2021, signed into law by President Biden on March 11, 2021. The law provides federal medical assistance reimbursement of 85% for three years to states to provide MCT services for people experiencing a mental health or substance use crisis, along with $15 million in planning grants to establish or expand upon such units.

Despite this investment, the evaluation literature is sparse and quite dated. Perhaps the most rigorous of MCT studies dates back to 2002, when researchers examined people receiving MCT services in Cuyahoga County, OH, who were matched to those who did not receive MCT services on several key variables to create a comparison group. Researchers found treatment group members were 17% more likely to access mental health services and a related study found that comparison group members were 50% more likely to be hospitalized than those receiving mobile crisis team services. A comprehensive literature review of MCTs concluded that while they have been successful in linking clients to treatment and reducing costly hospital stays, more robust and more recent studies are needed to explore these and other outcomes. An examination of the impact of offloading mental health-related calls to mobile teams on the volume of repeated calls for service, arrests, non-dedicated officer time, and use of force would be particularly helpful to jurisdictions assessing offloading options. While MCTs are the most common form of offloading of policing functions, other models exist that are tailored to specific issues and subpopulations. The Family Crisis Intervention Team in Rochester, NY, is housed in the city’s Department of Recreation and Youth Services and responds to calls about domestic issues, including family disputes, youth behavioral challenges, and child abuse. In Brooklyn, NY, the nonprofit Man Up! Inc. resolves minor community issues such as shoplifting by youth without police involvement through the use of “credible messengers.” These models and others like them are consistent with the community-led safety movement – a priority for the Biden-Harris administration, which has called for a $5-billion investment in community-based violence

Neighborhood Safety Offices

A relatively new development in the move to offload police functions is the establishment of offices of neighborhood safety. Funded by the municipality, these entities are staffed by non-law enforcement personnel and operate outside of the traditional criminal justice system. These offices develop solutions to a wide array of community challenges that are related to public safety but extend beyond it to address community needs such as job readiness and mentoring programs. They often hire community members who have experienced the criminal justice system and typically work in partnership with residents to define and prioritize community problems. The City of Richmond has been operating an Office of Neighborhood Safety since 2007 and some have credited this model with reductions in violent crime, although no research has identified a causal relationship between the two. A related model is Neighborhood Stat, a New York City initiative that engages residents of public housing in developing a public safety approach that is consistent with the needs and priorities of the community. Similarly, in 2017, Washington, DC, established an Office of Neighborhood Safety and Engagement, offering an array of community services, including those targeting young adults at risk of being involved with or impacted by violence. These models have not been rigorously evaluated; indeed, doing so is methodologically challenging given the many concurrent public safety initiatives that are typically underway.

Civilian Traffic Enforcement

Traffic enforcement and responding to traffic collisions have historically been key components of a police officer’s job. A Bureau of Justice Statistics analysis of the 2018 Police-Public Contact Survey documented 24.4 million contacts of drivers or passengers that were initiated by law enforcement and 8.9 million contacts between police and residents aged 16 and over owing to traffic collisions. Taken together, these traffic-related contacts represented 13.7% of all police-resident contacts in 2018. The most recent comprehensive report on traffic enforcement is a decade old, relying on data from 2011. This national survey of residents aged 16 and older regarding their contacts with the police found that among those involved in a traffic stop, 46.5% of stops were for speeding violations, 7% for an illegal turn or lane change, 6.7% for a stoplight or stop sign violation, and 1.3% for sobriety checks (61.5% total). Stops for vehicle defects (14.1%), records checks (9.7%), and seatbelt or cell phone violations (6.6%) represented another 30.4% of stops. Among all drivers who experience traffic stops, about 1% reported that the officer used physical force against them; of those, 55% believed the police behaved properly during the stop.

Racial Disparities in Traffic Enforcement

Source: "A large-scale analysis of racial disparities in police stops across the United States" by Emma Pierson, Camelia Simoiu, Jan Overgoor, Sam Corbett-Davies, Daniel Jenson, Amy Shoemaker, Vignesh Ramachandran, Phoebe Barghouty, Cheryl Phillips, Ravi Shroff & Sharad Goel

The racial disparities associated with police-initiated traffic stops are well documented. A study of 35 municipal police agencies found that the annual rate at which Black drivers are stopped by municipal police officers is substantially higher than for White drivers, at 0.20 compared to 0.14 per capita, respectively. The researchers also explored traffic stops by race and time of day, confirming that disparities were less pronounced after sunset, when it is more difficult for an officer to discern the race of the driver. Search rates are also higher for Black drivers despite two credible studies finding that “hit rates” – the share of searches that turn up contraband – are the same for Black and White drivers and a third analysis that found that hit rates for Black drivers were slightly lower. Roughly 10% of the 994 police fatal shootings in 2015 occurred at traffic stops, and one in three victims was Black.

The racial disparities associated with police-initiated traffic stops are well documented. A study of 35 municipal police agencies found that the annual rate at which Black drivers are stopped by municipal police officers is substantially higher than for White drivers, at 0.20 compared to 0.14 per capita, respectively.

Given the relationship between traffic stops and racially disparate policing and use of force, scholars and public officials alike are questioning whether police should retain the role of traffic enforcement or offload it to civilian employees in an agency outside of the police department. In July 2020, the Berkeley, CA, City Council voted to create a new city agency for that purpose, although the details have yet to be finalized. Similarly, the Lansing, MI, City Council instructed officers to stop pulling over motorists for “secondary traffic violations” such as cracked windshields and tail lights, items dangling from the rearview mirror, or loud exhaust. No other documented examples of offloading traffic enforcement in the U.S. exist, nor has research explored the impact of issuing tickets for broken tail lights and expired tags by mail rather than engaging in traffic stops. Experiences with traffic control in the U.K. and Europe, where traffic police are unarmed, may hold insights for U.S. police departments. The use of technology holds promise in efforts to offload enforcement of speeding and red-light violations. Research finds that red-light cameras issue tickets in accordance with the demographics of the population in the immediate vicinity of the camera. In addition, license plate recognition software can detect expired plates and outstanding warrants in a manner that is more race neutral than relying on the discretion of a patrol officer. However, such devices can perpetuate disparities if they are disproportionately sited in communities of color and can also contribute to the accrual of fees and fines among low-income drivers with limited ability to pay them, fueling jail incarceration rates.

Critics of offloading traffic enforcement cite the dangers of traffic stops as an argument that officers are better equipped to handle them. The FBI reported that six of the 48 officers (12.5%) feloniously killed in 2019 died while conducting traffic stops. A detailed analysis of traffic stop data from more than 200 Florida police departments over a 10-year period found the risk of officer injury associated with traffic stops is 1 in 361,111 stops, while the risk of felonious killing is 1 in every 6.5 million stops.

Call Taking and Dispatch

Before adopting any type of police co-responding or offloading approach, jurisdictions must fully understand public demand for public safety services and the best means of identifying those services that are identified for differential treatment. At the planning stages of changes to police or alternative service responses, examining incoming 911 calls alongside call taking, triaging, and dispatching processes can help assess the nature of requests for services by the type and severity of need. Analyses of 911 computer-aided dispatch data can provide insights about law enforcement responses to community member calls for service and quantify what share of police time is spent responding to calls, and at what cost. Since the types of services needed varies greatly between jurisdictions, information on 911 call systems and alternative non-emergency call-taking systems can highlight the pros and cons of triaging via existing 911 systems versus using existing or creating new non-emergency or non-law enforcement call systems.

Enhancements and Alternatives to 911

The most rigorous studies pertaining to 911 systems are somewhat dated and focused primarily on improving call taking and dispatching functions. These include several field tests conducted in the 1980s that used experimental designs to randomize various changes to the 911 processes, such as asking callers to report by phone rather than receive an in-person response by dispatched officers; delaying officer dispatch by 30 or 60 minutes, or scheduling an appointment for a later time based on the urgency of the request for service; or referring callers to another agency. These studies found that such alternative responses increased the share of calls for service that were handled with non-patrol responses and freed up patrol units to spend time on other police activities. These benefits had little bearing on community member satisfaction with the response they received; that level of satisfaction was uniformly high, though marginally higher among those who received a traditional in-person patrol response. Other studies pertain to the use of alternatives to 911 systems, such as developing and guiding community members in the use of 211 systems (which typically direct callers toward community resources) and 311 systems (designed to field non-criminal grievances and complaints). The most rigorous evaluation examined the impact of 311 systems and the degree to which they successfully diverted non-emergency calls from the police to other city services. Researchers found that many calls for potential police response came into the 311 system, while the 911 system continued to receive a high volume of non-emergency calls. Similarly, a more recent study of a large call center found high volumes of non-emergency calls coming into the 911 system and that four in ten calls to non-emergency lines needed to be re-routed to dispatch for police response based on the nature of the call or severity of the incident that generated it. The research on enhancements and alternatives to 911 systems suggests that making improvements to the triage functions and differential response mechanisms to existing 911 systems is more effective than relying on alternative systems to divert calls to non-police responders. However, this could change over time with support from a strong public awareness communications campaign.

Researchers found that many calls for potential police response came into the 311 system, while the 911 system continued to receive a high volume of non-emergency calls.

Dispatching Processes

The 911 system is the gateway to public requests for service and the allocation of police and EMS staff in response to those requests. Since the establishment of the 911 system in 1968, the public has become increasingly reliant on police for a wide range of problems, a trend that has expanded the police footprint considerably. Call takers are the true first responders for the system, answering calls, collecting information about each request for service, categorizing the nature of the request, and passing along some share of calls to dispatchers. Dispatchers then assign the appropriate emergency responders to the call and send them to the location of incident. Call takers have an important role in screening out calls that do not merit police response. A systematic social observation study of call takers in one of the largest public communications centers in the country found that they screen out about half of all calls. These calls are handled directly by the call takers and range from requests for information or assistance to non-emergency traffic collisions and false alarms. In fact, a third of all calls related to non-vice disorders and property calls were resolved by call takers, as were 22% of traffic-related calls. As such, call takers could have a more meaningful role in offloading calls that would have gone to the police or to other responders. However, the authority of call takers to offload calls is constrained by organizational rules and policies, and once calls are transferred to dispatch, they almost always generate a police response. Moreover, subjective biases may come into play as call takers develop narratives on the reason for the call, which can influence how dispatchers and officers respond to certain incidents. Call takers may also have a role in promoting or preventing the escalation of an event through their responses to the caller and the types of information they pass along to dispatch, suggesting that de-escalation and procedural justice training for call takers could be useful. Efforts to offload calls at the call-taker level include those employed by the Tucson Police Department, which has diverted calls for a wide array of low-level property, quality-of-life, and nuisance issues to non-police responders. The Camden County (NJ) Police Department has changed protocols to reduce the burden on call takers, dispatchers, and first responders; among other things, the agency verifies burglar alarm calls and requires complainants to go to the police department to fill out a form for non-injury automobile collisions, low-level thefts, and lost or missing property.

A third of all calls related to non-vice disorders and property calls were resolved by call takers, as were 22% of traffic-related calls. As such, call takers could have a more meaningful role in offloading calls that would have gone to the police or to other responders.

Assessing Calls for Police Service

Quantifying the nature and volume of calls for police service, as well as officer time spent on those calls, can help agencies anticipate the number of mobile crisis units that might be needed at various times of day and days of the week, or the potential impacts of offloading traffic enforcement on police time and resources. Such quantification can also provide insights about the impact of calls to defund the police, monetizing the labor costs of current police practice in responding to calls designated for offloading. Analyses of computer-aided dispatch data across an array of jurisdictions have identified that the majority of calls that ultimately receive a police response are not particularly serious. One five-site study found that the most common types of calls that generated a police response were non-emergency, predominantly complaint-based in nature, or a request for an officer to check on the welfare of an individual. A study of nine jurisdictions found that disorder-related events were the most common reason for police dispatch across all agencies, with traffic-related calls also featured prominently. The study also found that mental health-related issues made up a very small share of all calls, representing no more than 4% of total calls and averaging about 1.3% of calls across all study sites. In assessing how much police service could be offloaded, jurisdictions should examine not just the volume of calls by call type but also the time spent per call. For example, the nine-site study found that average officer time spent per call response (as measured from arrival to reported completion) ranged from 20 to 60 minutes. However, the time spent on mental health-related calls, which represented a very small share of dispatches, was at the high end of that spectrum, averaging just under an hour, or 2.2% of total officer time. Importantly, when examining the volume of calls by the time spent on each type of call, researchers estimated that traffic calls take up the greatest share of officer time, at 18%. Other calls that monopolize a high share of officer time include property crimes, vandalism, and disorder incidents.

A recent analysis of the Baltimore Police Department by the University of Chicago Crime Lab examined officer time spent on calls in order to estimate the capacity that would be needed to offload some share of those calls to other responders. The Lab found that mental health calls represented 2.1% of call volume and occupied 2.7% of officer time, while auto collisions made up 10.1% of call volume and took up 11.3% of officer time. The auto-related calls consumed 26,748 hours annually, or the equivalent of 13 full-time officers.

Critical Policy Elements

  • Training officers in mental health awareness without ensuring sufficient mental health resources exist in the community may not be effective in diverting people with mental health issues from arrest and incarceration.

  • Individuals with intellectual or developmental disabilities can exhibit behaviors similar to those displayed by people with mental health problems. Additional research is needed on whether training on police responses to those two groups should be similar or different.
  • Regardless of whether an agency has a co-responding unit, some share of officers will likely encounter people in crisis in the course of their routine patrol responsibilities. Given that, foundational training on how to identify and respond to people in crisis is warranted for all officers.

  • Agencies that plan to use existing 911 call systems to divert mental health and non-emergency calls away from law enforcement response should revise their training of call takers and ensure that sufficient non-police resources are available to handle such requests 24/7.

  • A thorough review of public demand for police services and police response to that demand requires analyzing the type and nature of incoming calls and calls that are dispatched, along with the duration of time each dispatched unit spends on calls by call type. Solely analyzing incoming calls will inflate the potential share of calls that could be handled administratively because those calls never reach the dispatch stage.

  • Offloading police functions assumes that the entities that assume those responsibilities are sufficiently resourced to serve effectively and fairly. Effective offloading requires sufficient supply to meet demands around the clock. There is no evidence that these actors behave in a more or less biased manner in their interactions with community members. For example, biases in the provision of healthcare services are well established.

  • Offloading some share of police functions to other actors could free up officer time for more essential functions. For example, officers could spend more time partnering with community members on projects that solve persistent crime and disorder problems, a strategy that research finds can be highly successful for a wide array of public safety issues.

  • Public safety call communications centers should partner with researchers to test new guidelines that afford call takers more discretion in offloading even more calls while ensuring those with public safety implications are dispatched to police.

  • Jurisdictions should develop detailed guidelines to inform dispatch decisions. Without objective criteria for assessing the risk associated with a call for service, subjective decisions about response types could result in disparate impacts (e.g., alternative responders are dispatched more frequently to predominantly White communities, while police responses are dispatched more frequently to communities of color).

Expected Impacts

Preventing Misuse of Force

Co-responders and professionals to whom police functions are offloaded may be more skilled at de-escalating situations involving people in crisis, reducing excessive force. This hypothesis has not been confirmed through rigorous research, though if alternate responders are not armed, encounters that result in fatalities or serious injuries to members of the public would be reduced. Injuries to responders, however, could increase.

Enhancing Transparency and Accountability

Changes to call taking processes may improve responses to public requests for assistance, enhancing accountability for timely and effective service, but research is needed to examine this potential relationship.

Strengthening Community Trust

It is unclear whether efforts to reduce the police footprint would strengthen or erode public trust in the police. Communities have differing experiences with police interactions, so reducing interactions will affect communities differently. Much depends on the nature of the remaining roles police have and how they approach them.

Reducing Racial Disparities

Given biases throughout American society, it is unclear whether co-responding alongside police or offloading police functions would enhance or reduce racial disparities in activities such as traffic enforcement and responses to people in crisis. This question demands further research.

Ensuring Officer Safety

Shrinking the police footprint would reduce police contact with members of the public, but the degree to which it enhances officer safety depends on the types of calls police respond to and what types of proactive policing they conduct. Similarly, more research is needed about the safety of non-police responders to calls that traditionally have been handled by law enforcement.

Promoting Public Safety

Diverting people from the criminal justice system to health and social services that address the underlying causes of criminal behavior may promote public safety. This would occur if such diversion prevents recidivism and frees up criminal justice system actors to focus on serious offenses.