CART Report

Compassionate Alternative Response Team Logo

Download a PDF version of the report here.


Table of Contents


Contributing Community Organizations, Service Providers, and City Departments

Executive Summary

Introduction & Context

Street Survey

CART Dispatch Response

CART Street Response



This report is dedicated to everyone who has been killed by police in San Francisco, and to the friends, family, and survivors of police violence who are healing and holding their people with love.

Phillip Randal Dunklin, Charles Blair Hill, Joshua Boling, Bernard Warren, Luis Góngora Pat, Alice Brown, Jessica Williams, Joshua Bowling, Donald Merchant and his pit Pretty Girl, and Jamaica Hampton are the names of a few individuals suffering homelessness who in the last decade have been severely abused, maimed or killed by San Francisco police.

For those who died, may they rest in power, for those who survived, may they rise above the multiple traumas inflicted by police violence and homelessness.

The need for this project arose from these unjust, preventable killings and maimings. This system robbed them of their lives, and robbed their friends and families of their precious loved ones. It robs our entire city of community members who make up the San Francisco we know and love, and that keeps us living and working here despite the high cost of living and rampant inequality.

This project asks what kind of City would be possible if unhoused neighbors were treated as worthy of life and dignity rather than as a nuisance or a threat, and if trauma-informed, unarmed civilians had been called to help rather than control the individuals named above. Many of us who have worked on this effort have personally witnessed and experienced the cruelty of the current system. Whether that be the tears of losing one’s property, the trauma of displacement to nowhere, or the loss of life-saving medications, these practices have led to deaths on the streets from despair, and disconnection from key medical and housing services.

Compassionate Alternative Response Team (CART) imagines that it would be a safer, healthier, and more vital city for the Black and Brown people who live and spend time here, and ultimately for everyone.

Contributing Community Organizations, Service Providers, and City Departments

– Coalition on Homelessness –

– Community Housing Partnership –

– Justice & Honor for Luis Góngora Pat –

– Lawyers’ Committee for Civil Rights of the San Francisco Bay Area (LCCRSF) –

– Code Tenderloin –

– United Council for Human Services –

– Mother Brown’s Dining Room –

– Treatment on Demand Coalition

– Senior and Disability Action –

– San Francisco Outer Circle (SFOC)

– Shelter Grievance Advisory Committee (SGAC) –

– Mission Neighborhood Resource/Health Center

– DOPE Project / National Harm Reduction Coalition –

– San Francisco Suicide Prevention –

– Felton Institute San Francisco –

– Tenderloin Neighborhood Development Corporation (TNDC) –

– San Francisco AIDS Foundation

– Taxpayers for Public Safety –

– Central City SRO Collaborative

– Hospitality House –


– Mental Health Association of San Francisco (MHASF)

– National Alliance on Mental Illness (NAMI-SF)

– Behavioral Health Commission

– Crisis Intervention Team (CIT) Mental Health Working Group–

– Supervisor Matt Haney’s Office –

– San Francisco Public Defender’s Office

– San Francisco Department of Public Health

– San Francisco Police Commission –

CART SF – A Community Plan for San Francisco

was published and released to the public on January 19, 2021

Executive Summary

CART logo: A shopping cart with a heart inside.

CART SF – Executive Summary (Full CART SF Community Plan linked here).

On a typical day, San Francisco Police Department (SFPD) officers respond to 179 homelessness-related incidents, or 1,253 weekly, most often resulting in move-along orders, citations, and destruction of property; systematically limiting homeless people’s access to services, housing, and jobs, while damaging their health, safety, and well-being.

In an attempt to alter the method with which housed residents reported encounters with homelessness the City established a 311 reporting system in 2007. The Healthy Streets Operation Center (HSOC) was established in January of 2018 as a proposed remedy for complaints due to tent encampments. Many members of the public did not want to have their calls diverted away from the police which resulted in a ballooning effect, increasing reports of violations of the Sit/Lie ordinance by 263%. This led to an intensification of resources to increased police presence during Department of Public Works crews coordinated through HSOC. At one point there were upwards of 52 officers working under HSOC. Due to the additional staffing, citations for Sit/Lie ordinance violations increased by 300%, nearly a third of San Francisco jail population identified as homeless, and unsheltered homeless rose by 19%.

Recognizing that additional police resources were not sufficiently exiting individuals out of homeslessness, the San Francisco Police Commission unanimously passed a resolution in January of 2020 calling for a more effective response to homelessness to be developed that would eliminate the use of police officers as a first and primary response. A stakeholder group was established, forming a coalition of those impacted by police violence and community based organizations to work in partnership with representatives from the Mayor’s office, staff from the Board of Supervisors, and City Agencies. Acknowledging the promise of such a program, the Board of Supervisors placed $2 million in reserve during the latest budget process to seed this yet to be determined program.

Despite being sidelined by COVID-19, the process started in July 2020 with over 50 participants working collectively. City departments and officials, community organizations, mental health consumers, people with lived experience with homelessness, service providers, advocates and academics were encouraged to envision a new way of approaching street crisis. From the start, the group was intentional about centering unhoused individuals in the design of the alternative. 95 unhoused neighbors, ages 18 to 67, were surveyed. Their responses were foundational in establishing a new response model that would eliminate police responses which have exacerbated racial disparities and disproportionately left those who are unhoused, disabled, and experiencing poverty feeling as if they are unwanted and disposable.

The new model would be called Compassionate Alternative Response Team (CART). CART reengineers emergency communications, dispatch, and response strategy to address the social and behavioral health needs occurring in public spaces while uplifting the unhoused.

CART is designed with two components:

1) CART Dispatch Response: The first provides a specialized police-alternative dispatch response to calls from and calls for unhoused neighbors in crisis, and the establishment of a new hotline to call CART directly. CART dispatch response recommendations include:

  • Scope of CART Dispatch Response for C-Priority Calls involving unhoused people in the City, on the street or in temporary shelters for the following dispatch radio codes: 800, 801, 910, 915, 916, 917, 919, 920.
  • Establish a new, direct CART hotline
  • Establish a timeline for CART deployment phases with SFDEM
  • Develop CART Dispatch process
  • End the use of SFPD in DPW Encampment responses
  • Eliminate HSOC
  • Repeal Sit/Lie Law
  • Implement CART to SFDEM Dispatch Training and add a CART-specific dispatch code
  • Revise dispatcher questions format to accurately identify CART appropriate scenarios
  • Establish co-trainings between SFDEM, CART, and Street Crisis Response Team (SCRT)
  • Establish an ongoing process of improving 311 and 911 call evaluation by dispatch operators as well as the priority designations given to incidents in which an unhoused person is involved
  • Develop a Collaborating, Learning, and Adapting Working Group (CLA) between CART, SFDEM/DEC and SCRT, and other City departments to review call-taking and dispatch incident data, no less than on a quarterly basis. CLA will monitor and evaluate alternative responses to policing, and determine necessary adaptations
  • Undertake SFDEM data development project to identify how many dispatched police incidents involve unhoused persons There currently is no disaggregated data on homelessness-related incidents for each SFDEM code.
  • Establish open data portal for public (non-confidential) data collected on homelessness related incidents and alternative responses to policing by CART and SCRT

2) CART Street Response: The second component serves as a community-strengthening hub to empower housed neighbors to more “compassionately respond” directly to their unhoused neighbors. CART street response recommendations include:

  • The CART program should consist of two categories of staff and primarily employ individuals with lived experience of poverty and/or homelessness
    • “CART Crisis Response Staff”
    • “CART Community Engagement Staff”
  • CART program’s visual street appearance should be distinguishable and contradistinctive from the appearance of law enforcement
  • CART is funded and staff for 24 hours a day, seven (7) days a week
  • CART team be privately run by a non-governmental organization
  • CART be subject to government and community oversight, with data and ongoing analysis shared in a public venue, to allow for accountability and public feedback

The two-prong scope of the CART program will divert a significant number of homelessness-related calls away from SFPD, while building capacity within San Francisco’s neighborhoods to de-escalate and compassionately resolve homelessness-related conflicts directly between neighbors, thereby reducing the total number of homelessness-related calls made to dispatch in the first place and reducing police interactions with those experiencing homelessness. CART responses focus on the well-being of the unhoused rather than the complaint of the caller, an approach that is foundational to the CAHOOTS model program. This will affirm the civil rights of those experiencing homelessness, as well as educating residents and business owners about life preserving approaches to reduce homelessness.

Measurable Outcomes of the CART Program would include:

  • Reduce police dispatches to homelessness-related quality-of-life complaints
  • Reduce the number of individuals transported to the emergency department for low acuity medical-related issues that could instead be addressed in a pre-hospital care setting.
  • Reduce the number of behavioral health and lower acuity medical calls traditionally responded to by the Police and Fire Departments and improve outcomes for those on the streets.
  • Reduce the number of homelessness-related calls to dispatch, in areas where the CART program’s community-strengthening interventions have occurred.

It is recommended that CART be funded by a diversion of funds from the SFPD budget. This could be achieved by diverting funding currently spent by the SFPD responding to homelessness-related quality-of-life violations (BLA estimates this to be at $18.5 million). The annual budget for the CAHOOTS program is about $2.1 million for responding to 24,000 calls for service. In San Francisco, adjusting for higher salaries, due to higher housing costs, and adjusting for the higher number of calls (65,000 annually) to be responded to, the budget is estimated to be $6,825,000. Implementing CART will yield at least $11 million in savings annually, and interrupt cycles of harm and violence that is all too prevalent when a police response is applied to situations involving an unhoused individual.

CART holds those who are on the margins of our community at the center of proper systems of care that result in dignity — instead of neglect — from institutions, investing in solution providers that live in the community and who see the challenges daily as residents. It is a community-led, government-funded response that is intentionally less violent in approach, focused on building safety for all.

Introduction & Context


The collective inability to deliver long-term solutions to the most vulnerable residents of this city leaves everyone feeling at a loss. Over the years, utilizing police as the first point of contact to address conflicts between housed residents, business owners, and unhoused neighbors have accumulated resentments by endangering lives, exhausting community members, enraging advocacy groups, and frustrating police officers. Despite the San Francisco Police Department’s (SFPD) attempts to re-engineer tactical applications of police interventions in calls for service, including a significant change in the SFPD Use of Force policies in 2016, there has not been a significant decline in the level of homelessness in San Francisco. Newly homeless residents continue to outpace the rate at which current city responses are exiting people from homelessness.

The failure of a police-led response to lead to stable housing has for too long inflicted trauma and pushed individuals deeper into poverty. Recognizing that this over-reliance on police responses to homelessness was not effective, the San Francisco Police Commission unanimously passed a resolution in January of 2020 establishing a stakeholder group, tasking it with the launch of a community-led process to develop an alternative response to the needs of people experiencing homelessness. Community meetings generated a framework for affirmative health-based approaches that we are confident will achieve healing for those traumatized by the revolving door of unnecessary involvement with police, jails, and criminality.

Homelessness must not be treated as a personal failure, and yet relying on a police response signals that these experiences are criminal and adjacent to criminogenic behavior. Though racism, ableism, and disability discrimination do not impact the lives of all who come into contact with police, they do exacerbate the racial disparities that exist among the population of San Francisco residents experiencing homelessness, and lead to an unequal distribution of resources and life chances.

COVID-19 has only further revealed the community’s deep desire for a compassionate and dignified response to struggling neighbors. The pandemic’s challenges will persist long after the majority of San Francisco is vaccinated. Forty-three percent of all U.S. rental households are at risk of eviction. Lethal overdoses due to fentanyl and other street-dealt drugs are three times the rate they occurred during pre-COVID times. Where government supports have fallen short or are nonexistent, community interventions have tried to catch folks. The isolation and lack of resources/services has unleashed immeasurable devastation and the heartbreak of COVID makes it more urgent to have actual pathways to service, a response that can help people get support and care, and out of the self-perpetuating cycle of desperation that often leads to cycles of trauma, crisis, and violence.

Compassionate Alternative Response Team, or CART, will create a system of care where people can ask for help before conflict escalates to violence.

The individual harms experienced by this population form a staggering cumulative disadvantage

The lack of easy access to meaningful recovery pathways is itself an escalation of violence, and mixing that with an unwanted police interaction is often traumatic to a community and potentially lethal.

While death and violence are the most severe negative outcome of a police response to the status of being destitute, there are many other shortfalls to this response. Police respond to approximately over 60,000 calls of these types (see dispatch section, page 27) — whether it be the presence of an encampment or an unhoused person in psychiatric distress. It is not only inhumane to respond to economically displaced individuals with armed officers, it is also ineffective and a poor use of resources. Typically, the outcome is that the person is no longer there, or is moved down the block or across the street. Rarely does a police response lead to ending an episode of homelessness, yet millions of dollars are spent on this same response — millions that could be used to invest in long-term solutions.

In this report, CART lays out a specific vision for a non-police response to homelessness. This alternative to a police response would decrease the workload and hours spent by police responding to homelessness and would necessitate reallocating police funding to resource the alternative.

This coalition comes together to articulate a San Francisco where CART organizes responses to vulnerable, unhoused individuals around care, decency, and dismantling systemic racism codified as barriers to healthy outcomes. The recommendations herein are informed by impacted community members. With the support of Code Tenderloin, Mission Street Neighborhood Services, Senior and Disability Action, Coalition on Homelessness, and community volunteers 95 individuals were surveyed. They were asked about needs which have long been unmet and what they wanted to see for the future. The survey results confirm what advocacy groups have acknowledged for so long — they deeply want structural responses rooted in preserving lives. A CART response will eliminate police responses that disproportionately leave those who are unhoused, disabled, and experiencing poverty feeling as if they are unwanted and disposable.

CART holds those who are on the margins of our community at the center of proper systems of care that result in dignity — instead of neglect — from institutions, investing in solution providers that live in the community and who see the challenges daily as residents. It is a community-led, government-funded response that is intentionally less violent in approach, focused on building safety for all.

None of us can be well unless all of us can be well.

The Shortfalls of San Francisco’s Current Police Response to Homelessness

Policing Perpetuates Homelessness and Harms Physical and Mental Health

In 2015, the Coalition on Homelessness and sociologists from University of California, Berkeley’s Human Rights Center conducted a study on the impact of anti-homeless laws on those experiencing homelessness. The study surveyed 351 unhoused San Franciscans, conducted in-depth interviews with 43, and analyzed a trove of administrative data from City departments. It found that policing nearly always resulted in move-along orders, citations, and destruction of property, which systematically limited homeless people’s access to services, housing, and jobs, while damaging their health, safety, and well-being. This echoes both previous and subsequent findings of studies completed in other U.S. cities and abroad. Since the report’s release, the policing of homelessness has only intensified as the City has more than doubled the number of officers dedicated to policing homelessness.

Below are some key findings of the report:

Experiences of policing due to homelessness was widespread.

  • Move-along orders: 70% of respondents had been forced to move from a public space. 93% of those camping, 88% residing on the streets, 80% residing in vehicles, 72% staying with friends/families, 61% of those in shelter, and 55% of those in a residential hotel had been forced to move from public space.
  • Searches: 56% of respondents reported having been searched while homeless. 21% reported that they had been searched within the month.
  • Property Destruction: 46% of respondents reported having their belongings taken by City officials while homeless, and 38% reported having belongings destroyed by City officials.
  • Citations: 69% of respondents were cited in the past year, and 22% of respondents received more than five citations in the past year. 90% of respondents were unable to pay the fine for their last citation. Unpaid fines create barriers to exiting homelessness, negatively affecting access to jobs, housing, and services.

Police interactions do not result in connection to services.

  • SFPD is far and away the largest displacer — accounting for 84% of displacements, 204 of the 244 most recent displacements reported by respondents.
  • Services or even information on services were rarely offered by SFPD. Only 24 out of the 204 respondents who reported being forced to move were offered services — most often a pamphlet, shelter bed, or sandwich.
  • San Francisco’s shelters consistently have a waiting list of over 1,000 people for a 90-day bed, and each night hundreds of people fall asleep in chairs waiting for a one-night bed. The city has less than 3,000 shelter beds for an unsheltered population of over 9,000 (2019 numbers).

Policing resulted in adverse effects for most of those experiencing homelessness. Most significantly was prolonging homelessness by creating barriers to housing, services, and jobs and worsening physical and mental health.

  • Quality of life-enforcement amplified risks of conflict, sexual assault, and violence by forcing people to move into unfamiliar areas and straining relationships on the street. 30% of participants reported increased insecurity after being forced to move.
  • Move-along orders disrupted sleep, the stability of mentally unwell people, and created challenges for social workers and health care providers to follow-up with their clients.
  • Property confiscation frequently resulted in the loss of critical medications, benefit cards and important paperwork required for services, tools for work, and precious personal mementos
  • .Unpaid citations resulted in bench warrants, increased fines, and spoiled credit ratings creating barriers to accessing housing, services, and work. (*Significant reforms have since been made to reduce these burdens by SF Courts).

Enforcement is Costly and Ineffective

According to San Francisco’s Department of Emergency Management (DEM), SFPD was dispatched to 65,333 homeless related calls in 2019. In other words, SFPD officers responded to 179 homelessness-related incidents on a typical day, or 1,253 each week. Police also responded to thousands of 311 calls as escorts to Department of Public Works (DPW) crews addressing encampments coordinated by the Healthy Street Operations Center (HSOC). Homelessness-related calls do not include all police responses to reported crimes that involve people experiencing homelessness, but rather a limited set of quality-of-life laws directly aimed at homelessness, such as violating the City’s ban on tents, sitting or lying on the sidewalk, illegal lodging, panhandling, blocking a sidewalk, trespassing, etc.

What are the costs and outcomes of this police response to homelessness?

A Budget Legislative Analyst (BLA) report of the City investigated these questions in 2016 and concluded that “current enforcement measures are too expensive” and that the department had “limited results from enforcing quality-of-life laws against the homeless.” The report found that the City incurred approximately $20.6 million in 2015 for sanctioning people experiencing homelessness. SFPD accounted for 90 percent of these costs, or $18.5 million. Although an estimate for more recent years is not available, the associated costs with policing in 2020 are far higher. The primary cost to policing homelessness are officer salaries. At the time of the BLA report, SFPD had anywhere between 19 – 24 officers at any given time responding primarily to homelessness complaints. In 2019, the SFPD had 52 officers working under HSOC responding to homelessness-related 911 calls and escorting DPW crews addressing encampments.The BLA report, which drew from data between 2014 – 2015, found that of the 60,491 homelessness-related incidents, police officers were unable to locate alleged violators in 26.5 percent (15,164) of these cases. At least 8.3 percent (4,711) of cases resulted in a citation directly attributable to homelessness, and only 0.2 percent (125) resulted in arrests. The majority of incidents therefore resulted in a move-along order. A subsequent analysis of citations with more comprehensive court data in 2018 found three times the rate of citation, with 14,888 citations given in 2014. Although the rate of arrest for a homeless related quality-of-life violation is low, according to the city’s 2019 Point-in-Time Count survey, 25% of respondents reported spending at least one night in jail or prison in the past 12 months. Furthermore, in 2019, 472 of approximately 1,500 inmates in San Francisco jails identified as homeless at the time of their arrest on the evening of the point in time count.

One of the main goals of quality-of-life laws is to preserve public spaces in the city. However, there is no evidence that the current police response results in a reduction of visible homelessness. The BLA report noted that the number of homeless individuals increased during the period of their data analysis between 2011 and 2015 according to the City’s biennial point-in-time count. Between 2017 and 2019, years that saw an intensification of resources dedicated to policing homelessness under HSOC, unsheltered homelessness increased by 19% from 4,353 to 5,180.

In the UC Berkeley and Coalition study, survey participants were asked where they relocated following their most recent move-along order from city officials. Only 9% of respondents reported moving indoors. Of these, some reported moving to drop-in centers, but the most common responses were moving to a public library or taking a ride on the bus — indoor spaces that are both public and limited to daytime hours. On the other hand, 91 percent of respondents remained on the streets or in parks, simply moving to a new outdoor location.

The primary reaction following a move-along order was simply to move down the street, around the corner, or to walk around and return after the police had left — a tactic taken by 64 percent following their last displacement.

A Police Response to Homelessness Exacerbates Racial Inequality

The policing of homelessness disproportionately exposes Black people to the criminal justice system. These laws don’t include overtly racist language, so why is anti-homeless policing anti-Black? First off, African Americans experience homelessness far more often than their white counterparts. While only 13 percent of the U.S. population is African American, 41 percent of those counted as homeless are African American. In San Francisco, 37 percent of those surveyed in the city’s 2019 Point-in-Time Count were Black, while Black people represent only 5% of the city’s total population.

Second, not only do those who find themselves homeless face frequent policing, but Black people experiencing homelessness are also more often policed. The Coalition and UC Berkeley study found that unhoused Black San Franciscans experienced citations, searches, arrest, and having their property destroyed more frequently than their unhoused white counterparts. The effects of gentrification and urban revitalization not only exacerbate Black homelessness, but may also intensify the policing of Black bodies. Consider the trend of white callers calling on Black men that have gained prominent media attention, such as the woman who called the police after feeling threatened by a Black bird watcher in New York City’s Central Park, or “BBQ Betty” who called the police on a Black family having a barbecue in an Oakland park. These and other instances highlight how a police response to homelessness contributes to exacerbating the racial inequalities of criminal justice treatment.

Due to the adverse effects of the current policing approach in San Francisco on those experiencing homelessness, its costly and ineffective ability to remove visible homelessness from public space, and its role in perpetuating racial inequality, a new strategy is needed to address street homelessness.

During the pandemic, there has been an inconsistently followed policy of allowing tents to stay if they follow basic guidelines, such as not blocking sidewalks or doorways. This was based on recommendations from the Centers for Disease Control and Prevention (CDC) to allow unhoused people to shelter in place in tents if there are no housing alternatives to minimize spread of the virus and allow for easier contact tracing. However, in some instances, displacements of tents have occurred without adequate housing alternatives, while in other areas of the city, tents have been allowed to stay.

The Invisible. Axel Osterberg

Coalition Process

After the Police Commission passed the resolution calling for the Board of Supervisors to create a stakeholders’ group to develop an alternative, preliminary meetings were pulled together in February 2020 under the leadership of Police Commissioner John Hamasaki to design a collaborative community process. The concept was to have an informal participatory inclusive process that unified community members with key City departments and elected officials. The Departments of Public Health, Emergency Management, and Homelessness and Supportive Housing, as well as the Mayor’s Office, were identified as key departments to be involved. Staff from Supervisor District offices with high numbers of unhoused residents were invited, including those of Supervisors Matt Haney, Dean Preston, Hillary Ronen, and Shamann Walton. In addition, the office of Supervisor Sandra Fewer, gave technical support. The Supervisors secured two million on reserve in the budget process for this program.

Organizations who had a stake in the creation of an alternative to a police department response were invited as well, including Code Tenderloin, United Council of Human Services, Hospitality House, Lawyers’ Committee for Civil Rights, Senior and Disability Action, Glide Foundation, Community Housing Partnership, and many more listed in this report.

The process planning was sidelined for a few months when the pandemic hit, but the formal process started up in July 2020. The Coalition on Homelessness hired Patrick Brown, senior consultant from the Justice Collective, to facilitate the process, and various organizations provided other forms of in-kind support.

Patrick Brown assisted the over 50 participants to collectively form a visioning and goals process, to utilize decision making to create outcomes. City officials, members of poor people’s organizations, mental health consumers, people with lived experience with homelessness, service providers, advocates and academics were encouraged to dream and envision a new way of approaching street crisis. From the start, the group was intentional about centering unhoused individuals in the design of the alternative, seeking their input to form the foundation of the work.

Three subcommittees were formed: research, dispatch and communications, and collectively the committees created a response model, which effectively responds to the needs of unhoused people on the streets, while very deliberately designing a system that no longer relies on unnecessary police responses.

The group decided the new model would be called Compassionate Alternative Response Team (CART).

Street Survey

Survey Design

The survey was designed to elicit a variety of responses regarding what a non-police response should entail. Survey respondents stayed anonymous but were asked to indicate their age, gender, and race they identified with, as well as which city neighborhood the survey was conducted in. The survey included general questions on what a new program should do and what a positive outcome would be, as well as questions in which survey respondents were asked to think of certain situations and whether they should be responded to and, if so, how.

The survey included the following six open-ended questions:

  1. San Francisco is thinking about starting a new program to get police out of dealing with homelessness. What should this program do?
  2. If you are resting outdoors and a business owner calls authorities to complain, how would you want this new program to handle the situation? What should the new program staff do or say to the business owner?
  3. If you are living on the streets and there is a heated argument, would you want this new program to respond and if so, how?
  4. If you or someone you know is having a drug or psychiatric crisis, what would you want this new program to do?
  5. When you have dealt with police, if it wasn’t a good experience, what would you want to have happened instead?
  6. What would you consider a positive outcome from your interaction with this new program?

None of the demographic indicators, nor the qualitative questions, required a response for completion of the survey.


To conduct the survey, unhoused San Franciscans were reached through individual street outreach as well as at drop-in centers of multiple community organizations, such as Mission Neighborhood Resource Center, Glide, Code Tenderloin and Mother Brown’s. A total of 95 surveys were conducted during the months of October, November, and December 2020. Since the survey questions did not require a response, survey participants sometimes declined to answer, resulting in responses not totaling 95 for each question. The survey responses were collected on KoBoToolbox. The responses to each question were analyzed using the method of qualitative coding.


Age range

Graph 1 – Breakdown of respondents by age.
95 people were surveyed ranging in ages from 18 to 67. No minors were surveyed.
Age 15-24: 9 respondents. Age 25-34: 13 respondents. Age 35-44: 19 respondents. Age 45-54: 12 respondents. Age 55-64: 13 respondents. Age 65-70: 5 respondents.

What is your gender?

Graph 2 – Breakdown of respondents by gender
Male – 57.4%. Female – 22.3%. No answer – 19.1%. Trans – 1.1%.

What is your race?

Graph 3 – Breakdown of respondents by race
Black – 34.0%. White – 22.3%. No answer – 18.1%. Latino – 14.9%. Asian – 6.4%. Mixed – 4.3%.

District representation

Graph 4 – Breakdown of respondents by district in which the survey was conducted
Tenderloin – 43.6%. Mission – 28.7%. Bayview – 11.7%. No answer – 8.5%. SOMA – 3.2%. Western Addition – 2.1%. Golden Gate Park – 1.1%. Haight-Ashbury – 1.1%.


Question 1 – San Francisco is thinking about starting a new program to get police out of dealing with homelessness. What should this program do?

95 responses

Overwhelmingly, survey respondents described characteristics of individuals staffing a new response team of an alternative to policing. Most responses included that the staff members should be For Us By Us (FUBU), such that they are from the community or have lived experience of homelessness. Respondents said social workers or people with mental and public health backgrounds should work for this team. The people surveyed shared a desire for the response staff to have a culture of understanding – knowing and empathizing with unhoused residents about the traumas of homelessness. As one participant put it, “This new program should have a culture that has a heart for homeless people, with no bias incentive to be mean.”

Desired characteristics/capabilities of new staff

Graph 5. Question I–Distribution of desired characteristics/capabilities of the new program staff
Lived experience – 9 responses. Culture of understanding – 9 responses. Not police – 9 responses. Harm reduction – 7 responses. Social worker – 4 responses. Network of care / mutual – 4 responses. Autonomy – 3 responses. Nonviolence/compassion – 3 responses. Regular outreach – 3 responses. Problem solving skills – 2 responses. Counselling – 1 response. Interdisciplinary – 1 response.

Additionally, a majority of responses laid out what resources this team should have and deploy to folks on the street. Notably, housing and shelter was mentioned most often. Food, hygiene, medical assistance, harm reduction services, and transportation were popular responses as well. It is clear, no matter what the team looks like, it should be well-equipped with services and resources to meet folks’ immediate needs.

Sample Responses:

“Hire more people who are felons or who have been homeless or can relate to the streets more. Please.”

“The program should be staffed by homeless and formerly homeless people.”

A program that will be hands on to getting to know the community and the people who reside in the community.”

We need them but they need to have understanding and respect.”

This program should be an interdisciplinary team –one like HOT team was supposed to be– consistent of medical, mental health, and harm reduction personnel, as well as generalist advocates. By that I mean someone with a background in social work or social justice, who is passionate about the homeless population and wants to empower them. This new program should respond to people and help them with any needs related to homelessness.”

I think police shouldn’t be dealing with homeless people.”

I would like for the new program to be more compassionate, ask more questions and have more resources like do you need medical attention and do you need Narcan and have some available and paper work for homeless resources.”

If the city can staff the 24 hour pitstops, then they can staff an ambassador program 24 hours a day. When the city does the budget, they pay police officers to deal with homelessness. Police need to deal with violent people and with life and death situations. This new program should have a culture that has a heart for homeless people with no bias incentive to be mean. In the past, police could do anything, be mean to people and take their stuff. People that have a heart for homeless people have no incentive to be mean, and so they should be staffed.”

They should create networks of care for trans and homeless users of drugs so they are properly taken care of.”

Question 2 – If you are resting outdoors and a business owner calls authorities to complain, how would you want this new program to handle the situation? What should the new program staff do or say to the business owner?

95 responses

The situation posed led participants to many varied responses including education, Right to Rest, and displacement of people experiencing homelessness (PEH) from public spaces. A majority of respondents who suggested “education” stated the business person should be educated about homelessness and the dearth of safe and dignified public places unhoused residents are often barred from. Few respondents replied educating the person experiencing homelessness about accessibility and blocking sidewalks.

Education Response

Graph 6. Question II–Distribution of responses suggesting an educational form of response
Educat business owner – 24 responses. Educate both (business owner + person experiencing homelessness) – 4 responses. Educate person experiencing homelessness – 4 responses.

Another commonality was the Right to Rest, or decriminalizing homelessness to allow folks to sit on the sidewalk, sleep on a bench, or lodge in a tent. Twenty percent of the respondents did not think the person experiencing homelessness should move, but explicitly said that their right to rest be respected. On the other hand, there were respondents who said the person experiencing homelessness should move, with some following up to say they should be brought into housing or shelter while others did not mention where the displaced person should then go. Lastly, there were a handful of respondents who did not answer the question or plainly responded, “I don’t know,” which demonstrates the difficulty some had with thinking outside the cultural norm of business owners’ rights superseding the rights of those on the street.

Other response components

Graph 7. Questions II–Distribution of other desired response components
Right to Rest – 19 responses. Culture of Understanding – 7 responses. Information and Referral – 7 responses. Compromise – 6 responses. Stores need to provide more for the community – 6 responses. Know your rights (as a homeless person) – 4 responses. Supplies/resources (e.g. for cleaning) – 3 responses. Laws should be clear/enforced – 3 responses. Mediation – 2 responses.

Sample Responses:

I wonder why homeless people cannot just rest outdoors without causing any trouble to anyone. The new program staff should do nothing if the homeless person did nothing wrong.”

“The program should help the homeless person find a spot to rest, and

teach the business owner compassion for the homeless.”

New program should inform both the business owner and the person of their rights.”

I mean the new program should tell the business owner something like ‘as long

as the homeless person isn’t bothering customers, just let them rest.'”

The staff should tell the homeless person a few suggestions on how to avoid the complaints, like cleaning up the street. The staff should tell the business owner to be patient, and teach them to have empathy with the homeless and explain that there is a reason why they are homeless. The staff should not give instructions, but suggestions –not like the police does it.”

Question 3 – If you are living on the streets and there is a heated argument, would you want this new program to respond and if so, how?

82 responses

Similarly to previous questions, participants expressed a desire for a non-carceral, community-centered response which is multi-tiered. A few respondents did not think that the given situation should be responded to; However, the majority of people said the new program should respond. A number of people answered that the team’s response should depend on the circumstances and the skill level and approach of available staff (i.e. non-punitive, patience and understanding).

Should the new program respond?

Graph 8 – Breakdown of responses to whether the given scenario should be responded to.
Should respond – 67.9%. Should not respond – 15.1%. Not respond, unless… – 9.4%. Situation dependent – 7.5%.

Due to the multitude of potential outcomes that could occur within any given situation, participants suggested that the response team should be a well-equipped, diverse, and interdisciplinary team that would create the potential for more robust responses, as well as de-escalation of minor conflicts. A common theme among the responses was a non-judgemental team with a culture of understanding and active listening. It is also important to note that a majority of those favoring a team response to the above scenario stated that the responders should hear both sides of the conflict, mediate without punishment, and ensure the safety of all people involved. Lastly, a response from the team should always be a trauma informed approach which prioritizes the autonomy and agency of people experiencing homelessness.

Key components of a desired response

Graph 9 – Distribution of key components of a desired response.
Intervene, resolve and diffuse – 10 responses. Assure safety – 9 responses. Mediate, facilitate, and negotiate – 4 responses. De-escalate – 3 responses. Respect – 3 responses. Understanding – 3 responses. Preventative – 1 response.

Sample Responses:

“Diffuse the situation. Prevent violence with a caring attitude. Make sure staff are trained and knowledgeable about violence. Staff should be from/live around the neighborhood so the people will know them. Keep the staff racially balanced and culturally sensitive.”

“The program should offer de-escalation advice.”

Heated arguments and situations like that, really require at least two staff people to separate both persons and get each person’s side of the argument, hopefully this will lead to both parties shaking hands and no one feeling that they are the loser.”

Yes, this program should respond, because being unhoused comes with a

special set of stressors and it’s hard to maintain “homeostasis.” So, yes,

a team responding to mediate would be welcome. It could be important

to check on the safety of other people involved and it could be an

opportunity to do a mental health check-in. This is all assuming that

you could always deny the service/interference of this new program.”

Yes, the program should make sure no one gets hurt.”

People out here are here 24/7, for sure there are arguments and there is

pressure to release. Authorities are no help in those situations. There should

be no response, unless they come with a lot of patience. People help

each other out here, because we know each other. Strangers are of no help.”

Question 4 – If you or someone you know is having a drug or psychiatric crisis, what would you want this new program to do?

90 responses

Participants overwhelmingly responded they would want help in a given crisis from this new program, stressing the importance of resources and services for mental health, substance use, and harm reduction. It is clear this team must link or refer people to actual, on-demand treatment options with capacity to meet them where they are at in their treatment. Similar to the responses to Question III, participants suggested the use of a multi-tiered system capable of responding to a diversity of immediate needs. Many individuals also emphasized the critical need for a non-judgemental, non-punitive response to these crises that prioritizes de-escalation and preventing harm without involving law enforcement.

Response components for the given crisis

Graph 10 – Distribution of desired response components for the given crisis.
Help – 18 responses. Appropriate programs – 13 responses. Medicine/medical – 12 responses. Refer/Link to treatment – 12 responses. Resources and – 9 responses. No jail/arrest/no police – 6 responses. No harm/threat/force – 4 responses. Narcan – 4 responses. Safe space – 2 responses. Non-judgmental – 2 responses.

Sample Responses:

“The program should evaluate the condition and provide services and resources as needed. It’s important to try and get at the cause of the crisis. ‘Small’ things such as just needing someone to talk or listen or being in dire need of a shower could cause someone to act out. The program shouldn’t stigmatize but act with patience and an open-mind.”

“Offer interventions and suggest treatment.”

The program should be able to refer or provide a stable rehab facility, which is currently not available. I’ve seen people who are in the window of wanting rehab, so they request it but are told they have to wait 6 months and then something dramatic happens. It is so hard to sustain that window, so rehab needs to be there right when the person is ready.”

Understand, be curious, give them water, food, medicine and time to settle down.”

Question 5 – When you have dealt with police, if it wasn’t a good experience, what would you want to have happened instead?

82 responses

While this was not specifically asked, in answering this question, 35% of respondents explicitly noted they had a negative experience with the police in the past, many of which described this bad experience in detail, ranging from lack of respect and intimidation to in-depth descriptions of horrifying and traumatic police violence. Furthermore, aggression, stigmatizing behavior, and abuse of power by police were also detailed in participants’ narratives. Some respondents also described specific shortcomings they observed in police officers’ behavior such as lack of respect, lack of communication, and lack of support. Only four of 82 respondents recalled a somewhat positive experience with the police. One respondent wrote: “I’d say I’ve had a good experience. I’ve been stopped though, but not arrested.” A couple of people also responded that they try to avoid interactions with the police whenever they can.

The majority of the respondents (65%), shared their vision of the kind of interaction or alternative response they preferred. Because most of these responses include comparative words such as “more respect” and “less aggression, it is possible to infer these recommendations to be based on firsthand experiences of police interactions with disrespectful and belligerent officers. Essentially, rather than being met with an assumption-based, stigmatizing attitude, respondents expressed a desire for basic respect and humane treatment, better communication, less force and aggression, and increased understanding of the traumas of homelessness.

What people would have wanted instead

Graph 11 – Interaction components that respondents would have wanted instead.
Humane and respectful treatment – 14 responses. Better communication – 10 responses. More understanding – 9 responses. Staff that got sensitivity training – 5 responses. Mediation and oversight – 5 responses. Less force/aggression – 5 responses. Calmness/patience – 4 responses. Alternative resolutions – 4 responses. Thoroughness – 3 responses. Fairness and justice – 2 responses. Direct line of help – 2 responses. Better support for filing complaints – 2 responses.

Sample Responses:

“Police should have been more understanding. Police should have been respectful. They don’t respect us and make our situation worse often times. Some police are okay, but they are the ones who come with patience and understanding.”

“Not be interrupted or disrespected by them; not have my life threatened by them (I’ve been told 3 separate times that if I didn’t respond the way they wanted, they would blow my f-ing brains out).”

Make sure new program doesn’t abuse their power and authority. Have oversight. Never talk to the police. Have police go through sensitivity training to learn how to approach a person.”

During a sweep for example, the police should have shut up and helped us gather our essentials and keep track of those things. Instead they would not stop talking to you and confuse you and not help you collect your thoughts. You’re being intimidated and interrogated. They should be equipped with knowing what a person’s essentials are and walk you through that list

We (me and my boyfriend) luckily don’t have too many interactions with police. There have been, however, incidences of threats of arrest by the police. In those situations police didn’t act appropriately. Instead of being quick to threaten with arrest – because they know you are vulnerable – I would have wanted a response without threats, one with appropriate warning given ahead, proper treatment of my property and with respect to my personhood. A response that mediates and is transparent about the availability of alternatives, options and services”’

“I would have wanted them to be more understanding and supportive. Just don’t judge.”

Question 6 – What would you consider a positive outcome from your interaction with this new program?

93 responses

Many responses to this question seemed to reflect how people are affected by their negative experiences with the police. The sentiment that any new program staffed with social workers –as opposed to police officers– would be better, appeared in multiple responses. Many people responded that a positive outcome would plainly be if an arrest and jail time were averted. A lot of respondents took this question as an opportunity to reiterate how the police mistreated them and how they prefer dealing with an alternative program. Similarly, some people did not necessarily specify what they would consider a positive outcome, but voiced that an alternative to police response itself would be a positive outcome. Respondents who did make specifications named a wide variety of things they consider positive outcomes, such as being linked to resources and services, accessing housing or shelter, receiving job assistance, as well as connecting with mental health and medical assistance, and more generally receiving help with homelessness-related problems and needs.

Positive Outcomes

Graph 12 – Distribution of what respondents would consider positive outcomes.
Resources: housing, benefits, food – 9 responses. Help/assistance with homeless related needs – 8 responses. Getting housed or sheltered – 5 responses. Job assistance – 5 responses. Mental health and medical assistance – 5 responses. Better living situation – 5 responses. Political advocacy and funding – 4 responses. Ending homelessness – 4 responses. No arrests or jail time – 3 responses. Not making peoples’ situation worse – 3 responses. Services (showers, transportation) – 2 responses.

Many people also took this question as an opportunity to describe desired characteristics of homelessness response. Respectful and humane treatment, consideration and understanding were the most prominent elements. Additionally, a number of respondents said it would have to be a non-punitive, non-threatening and community-based response, as well as staff with lived experience.

Characteristics of a Positive Response

Graph 13 – Distribution of identified characteristics of a positive response.
Respectful treatment – 8 responses. Consideration and understanding – 7 responses. Non-punitive and non-threatening response – 5 responses. Care and encouragement – 5 responses. Good communication – 5 responses. Community-based – 3 responses. Community-based staff with lived experience – 2 responses.

Sample Responses:

“Smiles on the faces of my people from dealing with the program.”

“To be treated like a human being.”

Anything that provides some kind of relief to the unhoused, as encounters with the police right now only bring the opposite. Whether it’s services, resources, a chat or a shower. I think, this team should be able to take any person, that wants or needs to, straight to a shower to restore this minimum sense of dignity, that helps tremendously in carrying on with your day.”

More support that is community based!”

A response that doesn’t treat us like their prey would be a positive outcome.”

Involve all people in the community to help prevent homelessness.”

A positive outcome would be if this program would stick around and not lose funding.”



The Dispatch Committee reimagined an emergency communications dispatch strategy that uplifts people impacted by homelessness and addresses the social and behavioral health needs occurring in public spaces. Recommendations were developed through conversation, collaboration, data gathering, and review of emergency communications protocols with input from City officials in charge of these services and other relevant public services.

One objective is to change protocols at the public safety answering points (e.g. public calls to 911 and 311) in order to divert dispatched responses away from law enforcement officers in situations where an alternative compassionate response is better suited for addressing the needs of individuals suffering conditions of homelessness on the street or in shelters. The Committee’s recommendations aim to ensure that call evaluation at the answering point more accurately identifies needs on the ground and dispatches a response from CART in those situations.

The successful implementation of alternative responses to policing requires good faith and commitment from the government of San Francisco, including the Division of Emergency Communications, the Police Department, the Sheriff’s Department, the Fire Department, the Department of Public Works, the Homeless Outreach Team, the Mobile Crisis Team, the Street Crisis Response Team (SCRT) and CART in order to appropriately implement, as well as monitor and evaluate existing and new alternative responses to ensure positive outcomes for the unhoused residents of San Francisco.

Dispatch recommendations were developed in response to three basic questions:

  1. Scope: What dispatch codes should CART respond to? (And, therefore, which codes will CART not respond to?)
  2. Systemic Change: What changes to training and protocols to dispatch communications services are needed to dispatch a CART response?
  3. Outcomes: How should CART be monitored and evaluated and other new alternative responses to ensure positive outcomes for the City’s homeless population?

To facilitate an understanding of the recommendations, the Committee began by providing some basic background on radio dispatch codes and existing data on dispatched police incidents involving unhoused individuals.

Background to 911 radio dispatch codes and current law enforcement responses

The Division of Emergency Communications (DEC) within the Department of Emergency Management (SFDEM) uses an established set of radio dispatch codes to communicate with the Police Department, Sheriff’s Department, Fire Department, and medical responders. There are over 130 radio dispatch codes for law enforcement and a couple hundred more for fire and medical related emergencies. Dispatch codes are used to identify a diversity of incidents. Primarily, this section reviews radio dispatch codes that require a law enforcement response. Dispatch codes are also always issued with an accompanying priority designation. The following are SFDEM descriptions for priority designations for communicating with law enforcement:

  • A-priority is highest priority and demands an immediate citywide response to a high risk situation, such as crimes involving life-threatening situations or serious property crimes. A-priority calls have a target response time of 7 minutes.
  • B-priority requires a district wide response to an incident, for example, with potential for bodily harm or damage to property, but without the immediacy implied by A-priority calls. This may be the case of a crime that has occurred or where the suspect may still be in an area. B-priority calls have a target response time of 20 minutes.
  • C-priority requires a sector response and refers to situations without present or potential danger to life or property. For example, a crime scene has been protected and the suspect is not in the area. C-priority calls have a target response time of 60 minutes, however, response times regularly take longer. (On response times, see also London Breed, “Alternatives to Police for Responding to Non-Violent 911 Calls”, Medium, Sept. 8, 2020.)
  • I-Priority calls are for broadcasting information (no units dispatched). This can be because there is not a location to dispatch but when a broadcast of information is most appropriate, such as a hit-and-run in which the suspect vehicle’s location is unknown but the description is broadcast.

The Fire Department and medical responders use a different prioritization designation. From highest to lowest priority, these are Code 3, Code 2 and Code 1.

According to data from the Department of Emergency Management (SFDEM), in 2019, there were 415,867 dispatched police incidents that were initiated by calls to 911 or 311. Of all dispatched police incidents, 22% were dispatched with A-priority, 36% with B-priority, and 40% with C-priority designations.

The need for urgency or expediency in responding within a specified window of time is determined by the priority code. In developing alternative responses to policing, it is important to keep in mind that alternative responses will need to be able to deploy a response that meets the priority code for an incident.

Homelessness-related incidents and dispatch codes.

The SFDEM defines “homelessness-related incidents” as violations committed by an unhoused person to “quality-of-life” ordinances such as the sit-lie law, illegal lodging, blocking sidewalks, camping, trespassing, panhandling, etc. Homelessness-related incidents are generally speaking dispatched requesting a C-priority response. The law enforcement dispatch codes most often used to describe a homelessness related incident are the following:

  • 601 – trespasser
  • 915 – unhoused person (this code was retired as a primary code in 2018)
  • 917 – suspicious person
  • 919 – sit/lie ordinance violation

A study developed by the SFDEM shows that from 2012-2019 45% of C priority calls were dispatch codes 601, 915, 917 and 919. As explained below, these are the codes most closely related to homelessness.

2012-2019 Homelessness Related C-Priority Calls

(Source: DEM; Dispatch Codes 601, 915 including incoming from 311 App, 917 and 919)

All other C-Priority calls: 2012 – 75,477. 2013 – 73,481. 2014 – 74,115. 2015 – 76,304. 2016 – 103,516. 2017 – 99,730. 2018 – 93,189. 2019 – 92,499.
Manly homelessness related: 2012 – 46,505. 2013 – 53,245. 2014 – 65,590. 2015 – 78,184. 2016 – 89,101. 2017 – 88,762. 2018 – 82,751. 2019 – 65,333.

The SFDEM does not have a breakdown of how many calls per each dispatch code involved an unhoused person, including for the homelessness-related calls. This is because the SFDEM database functionalities cannot automatically generate information about the population characteristics of individuals who are the subjects of police-dispatched incidents. This population characteristics can be found in the message content of a 911 call. However, access to the message content of a call is restricted by law to SFDEM staff and POST-certified personnel with “need to know” authority and requires a process of laboriously redacting restricted information first in order to make this information public. Therefore, the SFDEM would need to carry out a data development project with POST-certified personnel in order to provide homelessness-related data for each code.

Despite the lack of disaggregated data on homelessness-related incidents for each code, the SFDEM provided the following information:

  • Although 601 and 917 codes are not specific to homeless people, these codes often involve unhoused individuals.
    • Based on a recent sampling of the message content of 601 calls in one day, the SFDEM estimates that from one-half to two-thirds of all 601-C priority calls are homelessness-related.
    • Code 917 is commonly used by dispatchers as a catch-all code to describe a variety of potential illegal behavior, from suspecting a person of drug dealing to stealing a bicycle, which is why this code is often attached to describe a variety of homelessness-related conducts observed on the street.
  • As mentioned in the introduction, the gravity of the suspected illegal behavior will also prompt the type of priority response code assigned. For example, with the same level of actionable information, depending on factors such as location (near a school for example), some 917 drug sales might be a B-priority while a 917 drug use might be a C-priority.
  • The vast majority, if not all, 919 calls involve homelessness-related incidents given the subject matter of the Sit/Lie Ordinance.
  • In 2018, the 915 dispatch code designating a homelessness-related incident was retired as a primary code by the SFDEM, when the Healthy Streets Operation Center (HSOC) began receiving 915 service call requests through the 311 call center and 311 App.
Homelessness related C-priority dispatch calls

(2019; Source: DEM; Only dispatch codes 601, 915, including incoming from 311 App, 917, and 919)

41% – Mainly homelessness related. 59% – All other C-priority calls.

We share the SFDEM data with the knowledge that the use of 601, 917 and 919 dispatch codes may often and unjustly criminalize the very housing status of unhoused people surviving egregious human rights violations day to day. It points also to the need to educate the public that a police response to homelessness is neither effective nor humane.

Mental and Behavioral Health Incidents

In addition to the law enforcement dispatch codes mentioned above, the following dispatch codes describe mental and behavioral health incidents that would require a police, fire or medical response. Though not specific to unhoused people, these incidents may also be relevant to people residing on the street or in a shelter:

  • 800 – mentally disturbed person
  • 801 – person attempting suicide
  • 806 – juvenile beyond control
  • 910 – well being check
2019 Behavioral health and well-being dispatch calls

800, 801, 806, 910 dispatch codes

(Source: SFDEM)

Initial Code Assigned by DispatchA-priorityB-priorityC-priorityTotal
800 mentally disturbed person4,60410,41221515,231
801 persona= attempting suicide3,6951339*3,837
806 juvenile beyond control*1431788329
910 well being check9,24817,51422026,982
*Note: CART will not respond to 806. 801 is never a C priority unless it is a report after the fact, or another reason that does not indicate a speedy response.

SFDEM (San Francisco Department of Emergency Management) data for 2019 shows a total of 31,148 dispatched incidents involving codes 800, 801, 806, and 910, across A, B, and C priorities. While the city does not have information about how many of these dispatched incidents were homelessness related, there is information demonstrating that less than one percent (1%) of those calls were C priority, that for the most part are “report” calls not involving an active incident.

Many B-priority calls may also not involve an active incident. However, 910s are almost all dispatched as A or B priorities, meaning there is an expectation of urgency for the person’s health and safety. Any response instead of police would need to have the same or better response time. 915s and 919s, by contrast, are almost always C priorities where there is not a level of urgency that an alternative response to law enforcement would need to meet.

Dispatch or police officers once on site may also designate a code 5150 – mental health detention.

Dispatch calls may also be accompanied by a suffix providing further information about the the type of response required:

  1. CR (a police crisis intervention team dispatched, primarily when incident involves a weapon)
  2. DV (domestic violence/intimate partner incident)
  3. EA (potential elder abuse)
  4. CA (potential child abuse)
  5. 222 (person with a knife)
2019 Behavioral health and well-being dispatch calls

5150 and Suffix codes

(Source: SFDEM)

Initial Code Assigned by DispatchA-priorityB-priorityC-priorityTotal
5150 (or code attached by officers on site)291190148

Street Crisis Response Team (SCRT)

On November 30, 2020, the City of San Francisco rolled out a new Street Crisis Response Team (SCRT). SCRT is a partnership between the Department of Public Health and Fire Department with significant support from the SFDEM and in a collaborative relationship with SFPD. SCRT was pushed for by the Fire Department who notes a large portion of their calls were homelessness-related and wanted additional resources with paramedics. SCRT will offer a non-police response to 800 dispatch calls involving individuals suffering from mental health and substance use issues on City streets. However, while the team has the ability to conduct 5150s, it is not responding to those who are a threat to others, and will not transport those posing physical threat.

Each SCRT team includes a community paramedic, a behavioral health clinician, and a behavioral health peer specialist. The goal of the new program is to provide an appropriate non-law enforcement response to behavioral health emergencies in San Francisco and divert individuals in crisis away from emergency rooms and criminal legal settings into behavioral health treatment. The SCRT aims to provide trauma-informed clinical interventions and care coordination for people who experience behavioral health crises on the streets of San Francisco.

That said, SCRT began as a pilot program with one team in the Tenderloin with plans to expand to at least six teams by the end of March 2021, in high-need and small geographic areas, and will for now only respond to adults in crisis. This model costs about $17 million annually, and the team will generally respond to B and C level 800 calls. In the initial rollout phase of the SCRT program, dispatchers will begin identifying dispatch calls to SCRT with a new medical dispatch code 25A0. Currently, DEM simultaneously creates a police and an SCRT call for what had been 800 B calls. SCRT responds when the call is determined to be in-scope and SCRT is available. The plan is in Phase II in April, police will no longer respond to what had been 800 B calls and they will all be SCRT or EMS responses. Eventually, SCRT will be responding to all of what had been 800 B priority law enforcement and law enforcement will not be sent. Recommendations below consider the deployment of SCRT.

311 calls and the “HSOC Balloon Effect” that Needlessly Dispatches Police Responses to Unhoused People

In January 2018, the Healthy Streets Operations Center (HSOC) initiated operations and the public was oriented to use the 311 non-emergency response number or mobile or web-based app to report homelessness-related calls. 911 dispatch operators were trained to divert any incident reports involving a homeless person that did not require a police, fire or medical response to the 311 non-emergency call center. However, SFPD became the agency lead of HSOC and

policing of homelessness intensified through HSOC. An SFPD Commander directed day-to-day activities as the Operations Commander, over 50 SFPD officers were assigned to the unit, and SFPD escorts to DPW crews dramatically increased.

Calls regarding encampments are re-routed through 311, and sent to HSOC, where HSOC develops a response. As of 2018, when a private citizen calls or otherwise reports an issue related to homelessness to 311, the City’s Central “Customer Service Center,” requests are self-categorized by the caller as either an:

  1. Encampments – Encampment Reports
  2. Encampments – Encampment Items, or
  3. Homeless Concerns – Individual Concerns

The 311 Dispatcher further categorizes the request details as:

  1. Encampment Cleanup
  2. Homeless Other
  3. Wellbeing Check
  4. Cart Pickup, or
  5. Aggressive Behavior

Then, the request is flagged for the DPW Ops Queue, Police – Homeless Concerns Queue, or 311 Supervisor Queue.

Hyper-focused on the removal of large tent encampments, in a response to encampments HSOC may involve the SFPD, Department of Public Works (DPW), Department of Public Health (DPH) and Homeless Outreach Team (HOT Team). More often than not, the response involves SFPD, especially when DPW is sent out to clear an encampment. HSOC responses frequently occur without proper notice to unhoused people, involve illegal confiscation of property, and typically offer no housing to targeted individuals, who are displaced from one block to another. These responses also use public resources that could be better spent on root solutions to homelessness.

In addition, when resources are offered to unhoused individuals through HSOC operations, it results in inequities of resource distribution as it mostly involves a complaint-driven system as opposed to need-based response. For example, the individuals whom HSOC responds to might be those who are located in an upscale or recently gentrified area generating the greatest number of complaints, and therefore, these individuals receive scarce housing resources, while individuals in more impoverished areas continue to suffer quietly with higher levels of acuity. Occasionally, HSOC will concentrate on a geographic area where they will partner with neighborhood agencies, survey the area for where people are, and coordinate with services.

The SFDEM also noted that pushback from members of the public who did not want to have their calls diverted away from the police saw a significant increase in the report of 601 (86% increase), 917 (25% increase) and 919 (263% increase) dispatched incidents in the period from 2016 to 2019. This is often known as the “balloon effect” whereby law enforcement responses are squeezed out to other departments, but for diverse reasons, often related to implementation, remain or return to the scope of activities of the police.

Police responses to homelessness-related incidents needlessly results in the criminalization of homeless individuals, which is contrary to the Supreme Court ruling (Martin v. Boise) that upheld a decision in the 9th U.S. Circuit Court of Appeals stating that homeless persons cannot be punished for sleeping outside on public property in the absence of adequate alternatives. People experiencing unsheltered homelessness — at least according to the 9th Circuit — should sleep without facing criminal punishment for simply trying to survive on the streets. In 2019, a panel of the 9th Circuit held that “as long as there is no option of sleeping indoors, the government cannot criminalize indigent, homeless people for sleeping outdoors, on public property, on the false premise they had a choice in the matter.” The 9th Circuit court’s decision can be read here.

CDC Interim Guidance on Unsheltered Homelessness and Coronavirus Disease

In addition, in 2020 the CDC issued an interim guideline on unsheltered homeless individuals during the COVID-19 pandemic. Considerations for encampments state that if individual housing options are not available, people who are living unsheltered or in encampments should be allowed to remain where they are, since clearing encampments can cause people to disperse throughout the community and break connections with service providers. This increases the potential for infectious disease spread.


Recommendations: CART Scope of Dispatch Responses

Proposed Scope of Dispatch Codes for a CART Response

Only C-priority

Only persons suffering conditions of homelessness

Dispatch CodeDescriptionHomeless SheltersStreet or street facing incident
800 (police / 25A0 (medical)Mentally disturbed person, but only overflow calls from the SCRT team or paramedic first responders; will exceptionally take B-priority calls same as SCRTYESYES
801Person attempting suicideNOYES
910Well-being checkYESYES
919Sit/lie ordinance violationsN/AYES
920Aggressive panhandlingN/AYES
915Homeless encampment, currently routed to 311N/AYES
601Trespassing/ unauthorized person occupying property/ premises of anotherN/AYES
916Suspicious person in carN/AYES
917Suspicious personN/AYES

Considering the frequency of use of specific dispatch codes in homelessness-related calls, and in consideration of CART’s goals, the Committee recommends the initial phase of the program focus on responding only to C-priority calls involving unhoused people in the City, on the street or in temporary shelters for the following dispatch radio codes:

  • 800 – mentally disturbed person, but only for overflow calls from the SCRT team and paramedic first responders; exceptionally will also take B-priority calls same as SCRT.
  • 801 – person attempting suicide, but only when on the street. These calls are rarely C priority as they require a speedy response, however if CART is able to have less than 7 minute response times, this would be appropriate to expand to in future.
  • All 920 – aggressively panhandling

The following two codes tend to cover the presence of homeless people and do not typically contain harmful or troubling behavior or incident beyond the visible effect of government disinvestment from housing for impoverished people and structural inequities. It is key that careful screening of these calls take place by CART staff to ensure they are not the result of class or race bias, but do deserve a team response. These calls should be reviewed to see if they are meaningful and deserve a response. If they are clearly harassment calls, CART will not respond with a team, but work with the caller to educate them on the lack of shelter and housing options available. If there is no response, the police will not be dispatched either.

  • Select 919 – sit/lie ordinance violations, but only when the CART team determines there is a substantive reason to respond, and there is no harassment being conducted.
  • Select 915 – homeless encampment, these are currently routed to 311 (are no longer used by 911 dispatch), and will instead be routed to CART. Once at 311, they currently refer to HSOC. 311 calls of this type would go to CART instead..

In addition, it is recommended that CART be dispatched to additional incidents instead of police involving the following C-priority related to an unhoused people on the street:

  • 601 – unauthorized person occupying property/ premises of another (trespasser)
  • 916/917 – suspicious person in a car/suspicious person
  • All 910 – well being check

If not explicitly included above, then CART will not respond to any other dispatch codes. This working group did evaluate the possibility of responding to other codes, such as 602 (breaking and entry), but concluded that it is very important to avoid CART from becoming an “alternative policing response” rather than an alternative response to policing. This differentiation will be key to the success of the team. Critical to successful outcomes is the development of trust with unhoused community members who at best have experienced a string of broken promises from city outreach teams, and at worst have been demonized, criminalized and shamed for their very destitution. For this same reason, CART would also not want to respond to 917 calls related to drug sales, even in an encampment. CART interventions should remain true to its purpose of addressing the social and behavioral health needs of homeless individuals.

The purpose of CART is to support compassionate alternative responses to policing of unhoused peoples. However the planning body fully supports ongoing efforts for alternative responses to policing related to non homeless specific situations, including but not limited to sex work or prostitution and drug sales.

Establish a direct CART hotline. It is also recommended CART be allowed to take in calls directly from the public through a dedicated hotline number.

Establish a timeline for the CART deployment phases. In coordination with the SFDEM, establish a timeline to deploy a CART response.

Develop CART dispatch process. Similar to SCRT, the working group recommends that it should be CART’s prerogative whether to accept or reject a dispatch call. On receiving a dispatch notice, CART staff will read the text of the call, verify the geographic range and determine if the call is (a) within specified dispatch codes, (b) specified geographic scope, (c) on the street or in a homeless shelter, (d) related to an unhoused or presumably unhoused individual (either adults or youth if in a condition of homelessess), and (e) whether there is a team available to respond. CART’s capacity to respond will be clearly delineated in a Memorandum of Understanding with the City/ agreement/ ordinance allowing for CART.CART’s recommendation is aligned to best practices found in similar programs (e.g. Eugene, Denver, Portland) run by civil society organizations.

Divert police funding to finance a CART response. Considering that CART will take on dispatch response workloads currently assigned to the Police Department, a CART response should be financed with funds diverted from the Police Department. Given the very large number of dispatch calls which are homelessness-related, this would eventually and significantly decrease police work load on C-priority calls, especially as additional A and B-priority calls are reclassified as appropriate for a CART response. Under no circumstance should a CART response be financed by Proposition C funding or other funding currently assigned to finding long-term solutions to homelessness or programming assigned to treating homelessness and unhoused individuals, because taking dollars away from long-term solutions to homelessness is a self-defeating perspective.

Never dispatch SFPD to a DPW encampment response. Where the 311 call center dispatches DPW to an encampment, it is recommended that the Police Department NEVER be dispatched with DPW as a police escort or law enforcement dispatch call. United States Interagency Council on Homelessness has developed federal guidelines which provide best practices to avoid displacement.

In addition, the HOT Team should be trained to understand alternative responses to policing, so that they never call in an armed police response if an alternative compassionate response is best suited to the needs of the individual involved.

Eliminate HSOC. HSOC was ill-conceived from the beginning as a tent and police-centered response to homelessness. Elimination of HSOC will ensure a response to homelessness based on the self-identified needs of unhoused people, and will no longer be necessary with CART responding to these calls.

Repeal Sit/Lie Law. This ordinance targets the most vulnerable residents of the city and calls for a punitive approach to homelessness, which increases police contact with and potential for abuse of unhoused individuals. We ask that the Board of Supervisors and Mayor place a measure on ballot to repeal the Sit/Lie Law and instead focus City efforts on finding long term solutions to homelessness.

Recommendations: Training Changes Needed to Implement a CART Response

Add a dispatch code or suffix that dispatches a CART response. Based on SFDEM data that shows that a significant percentage of C-priority dispatch codes for 601, 917 and 919 are homelessness-related incidents where a compassionate alternative response would be in the best interest of people suffering conditions of homelessness, the recommendation is for SFDEM to develop a non-law enforcement dispatch code that would prompt a CART response.

Update dispatch protocols. All protocols, including training bulletins, operational updates, and line announcements for dispatch codes 601, 800 (overflow), 801, 910, 916, 917, 919, and 920 will need to be revised to allow for CART to be dispatched to respond to homelessness-related C-priority incidents.

  • Change dispatcher question format. Specifically, the general question format followed by dispatchers for police, fire, or medical incidents will require asking information that identifies: (a) a homelessness-related incident or homeless person as the primary subject of a call, (b) geographic scope of CART response, and (c) street or shelter location.
  • Timely updates. Protocols be revised in to reflect any expanded level of CART response as the program moves through deployment stages.

Co-trainings. Co-trainings are recommended between CART and SFDEM/San Francisco Department of Emergency Communications dispatchers and SCRT, as well as other City response services as required.

Improve call evaluation and priority evaluation to enhance scope of work that may merit an alternative response to policing.

CART recommends establishing an ongoing process of improving call evaluation by dispatch operators as well as the priority designations given to incidents in which an unhoused person is involved, in order to improve assessment of whether a compassionate alternative response to policing would provide a more positive outcome for the individual involved.

  • Improve call evaluation of 601, 916, and 917 homelessness-related incidents. CART recommends establishing an on-going process of improving call evaluation by dispatch operators of 601, 916, and 917 calls to identify when an unhoused person is involved, and whether a compassionate alternative response to policing would provide a more positive outcome for the individual involved. The ways operators could identify housing status is by asking the caller if the individual is sleeping, sitting or laying down, and if they have bags or carts with them. In that case, calls should be classified as C-priority. For example, a 601 (trespassing) or 916 (suspicious car) call could be diverted to CART (or another City homelessness service) if the unhoused individual is simply seeking shelter on a front stoop or in a car. Similarly, CART could be dispatched to attend a 917 (suspicious person) call, when the individual involved is not a suspicious person but simply existing while homeless.
  • Improve evaluation of priority response when an unhoused individual is involved. Many dispatch calls that could potentially be an area of focus of a CART response may currently be receiving an A or B priority designation, and, therefore, CART could miss opportunities to work with unhoused individuals who would benefit from an alternative response to policing. To that end, CART recommends that the CLA group (mentioned below) or a subgroup composed of members from the SFDEM and other relevant departments be tasked with improving call evaluation in order to identify situations in which the appropriate priority designation involving an unhoused person could merit an alternative response to policing. Our hope is that all poverty-related calls will eventually receive C-priority designation. For example, if 601 squatting currently receives a B-priority, a revision could reclassify as a C-priority situation, where the situation involves an individual squatting in a vacant lot.

    Similarly, in an incident involving an unhoused individual in a mental health crisis (800 call), who may be threatening to damage or damaging an inanimate object of public or private property (but where no threat of violence to a person is involved) may be considered an A or B-priority situation, but could be handled by CART, if CART is able to respond in a timely manner. The objective of CART is to attend to a person in crisis, and where no threat of violence to a person is involved, a compassionate alternative response may attend to the needs of that unhoused person best.

    Related to the above, incidents involving persons with a knife (suffix code 222) or other non-firearm weapon (e.g. metal bar or other blunt object) may immediately increase the level of priority of a call. However, it is well known that unhoused people carry knives and other tools for utilitarian purposes for living on the street. CART therefore recommends that the working group established with the SFDEM assess situations that currently are given an A- or B-priority designation due to the presence of a knife or potential weapon, but to which a CART response could be deployed if the dispatcher identifies the presence of a nonviolent unhoused individual as the primary subject of the call. For example, assessing if the weapon is being used to threaten an actual individual, if the person is sitting down, and so forth. This would imply changes to protocols and training as well.

    Our recommendations also refer only to dispatched incidents versus incidents which are initiated directly by police officers in the field. However, CART hopes that as the program evolves, police officers in the field will also identify situations for which CART is better suited to respond.
  • Improve call evaluation of 915 and 919 dispatch calls entering through 311 (or 911). We have recommended above the repeal of the Sit/Lie ordinance and elimination of HSOC. However, until such time as this happens, we recommend an ongoing process of improving call evaluation by 311 dispatch operators (and 911, if necessary) to ensure a response from CART or City workers that is in the best interest of unhoused peoples.

The case of Luis Góngora Pat

Luis’s Story

Luis Demetrio Góngora Pat arrived in San Francisco in 2002 from his hometown of Teabo, Yucatán. Like so many thousands of other Mayan Yucatecans in the Bay Area before him, he began working in the restaurant industry, as a dishwasher at Mel’s Diner. He was first and foremost a family man, a primary provider for his wife Carmen, his two sons Luis Jr. and Ángel, and his daughter Rossana, as well as his elderly parents Demetrio and Estela. For a long decade, he had a stable life working, sending remittances back home and living in a rental apartment in the Mission District with his brother José and another Mayan friend. Their friend helped them with the logistics of managing rent until the day he departed back to the Yucatan.

In the fall of 2013, during the peak period of gentrification in the Mission, Luis arrived back home one day to find his belongings in the garbage dump. The circumstances of his abrupt eviction are not clear. His brother José was out of town at the time, but suspects that Luis had language and literacy barriers that complicated his relationship with the landlord. Luis began living on the streets, spiralling into deep disappointment. José also suffered homelessness, but was taken in by his cousin Luis Poot Pat. Back on his feet, José and his cousin made a plan to help Luis, but they would not get that chance.

A 311 call, an SFHOT team, and a failed non-emergency call led to a dispatch code 222 A-priority and an unnecessary, brutal, and fatal encounter with police.

On the morning of April 7, 2016, two SFHOT members responded to a 311 call about a baby crying in an encampment on Shotwell Street near the corner of 19th Street in the Mission District. No baby was present, but the SFHOT members interacted with a pregnant resident of the encampment. As they were leaving, they saw Luis wielding a knife, stabbing at some trash. Luis was not threatening anyone, nor did anyone in the encampment feel threatened. However, the SFHOT members decided to report Luis and the knife to the non-emergency police number. After several unsuccessful calls, they chose to call 911 from their car to report a man “waving a large kitchen knife.” Police were dispatched with a code 222 (man with a knife) and the highest Priority-A response. It was 10:04 a.m. and by 10:08 a.m. Luis would be fatally shot on the ground a few paces away from the safety of his tent and his encampment community, unjustifiably killed by two SFPD police officers.

By all eyewitness accounts, both housed and unhoused people, Luis was simply sitting on the ground with his back against a garage roll-up gate, minding his own business, at the time when the police arrived. Officer Michael Mellone and Sergeant Nate Steger exited their patrol car, and within 28 seconds Mellone yelled commands in English, shot five rifle bean bag rounds into Luis’s back as he crouched away, and Sgt. Steger backed him up by shooting his firearm. Luis was first propelled to his feet, and then immediately fell wounded to the ground. Michael Mellone then took out his firearm, and still holding the bean bag rifle in his other hand, executed Luis with a shot to the head when he lay helpless on the ground. It was the only fatal wound on Luis’s body according to the Medical Examiner’s report.

Luis’s killers were found by the Department of Police Accountability and the SFPD Internal Affairs Department to have violated every SFPD rule in the book on use of force when they breached time and distance policies without justification, and fired two different types of lethal weapons that resulted in Luis’s death. And yet, their only consequence was a suspension. Luis’s family has requested Chesa Boudin, the recently elected District Attorney, to open a criminal case against his killers.

Since the day Luis died, his family of courageous working class Mayan warriors have led and sustained his justice cause, on two sides of the border. They will not rest until Luis’s life is honored and justice is found. In honor of his brother, José specifically requests that funds from SFPD be diverted towards homelessness solutions.

Mural by Marina Perez-Wong & Elaine Chu of Twin Walls Mural Company

Recommendation to the San Francisco Police Commission

Impose severe disciplinary sanctions to officers who breach Use of Force Time and Distance policies. CART recommends that the Police Commission revise the Use of Force Department General Order that mandates the observance and enforcement of time and distance policies to state that breach of this policy will result in immediate termination of an officer given the unnecessary endangerment of an individual, when this policy is breached. The continued breach of time and distance policies by police officers impedes the employment of alternative de-escalation responses, and endangers lives.

Recommendations: Dispatch Monitoring Evaluation to Ensure Positive Outcomes

Develop a Collaborating, Learning and Adapting (CLA) Working Group. CART recommends that a Collaborating, Learning, and Adapting (CLA) working group between CART, SFDEM/ DEC and SCRT, and other City departments be developed to review call-taking and dispatch incident data on a periodic basis, as determined by the parties but not less than on a quarterly basis, in order to monitor and evaluate alternative responses to policing, and determine necessary adaptations and requirements to ensure positive outcomes for unhoused individuals. The objective of the CLA working group will be to prioritize the well-being of homeless persons over their criminalization by improving the functioning of new alternative responses to policing.

To that end, the CLA working group will develop outcome and output indicators to track ongoing deployment of alternative responses to policing. The CLA will also determine the specific data to be collected and shared by parties, including on behalf of SFDEM, the number of homelessness-related calls requiring an alternative response to policing that were left unmet. Data to be tracked should include basic information about the person, critical medical information, follow-up appointments, services provided to them, and what referrals and attempts at linkage were done. This information could also be used by individuals providing continuing care.

In order for the CLA working group to properly function, the SFDEM, SCRT, and other City departments, as appropriate, will share dispatch and response data for 911 and 311 calls related to the universe of calls that CART responds to in a timely manner with CART, including sharing population characteristics where available.

An SFDEM data development project is needed to identify how many dispatched police incidents involve unhoused persons. Given the lack of disaggregated data on homelessness-related incidents for each code, and that the SFDEM can only grant access to the 911 call center data to POST-certified individuals with a “need-to-know” authority, the SFDEM will need to carry out a data development project to identify how many dispatched police incidents involve unhoused persons in order to generate more accurate and useful public information that can be used to expand alternative responses to policing when homeless individuals are involved.

Open data portal. CART also recommends that the public (and non-confidential) data collected on homelessness-related incidents and alternative responses to policing from CART and SCRT be shared through the City’s open data portal with useful information summaries in order to provide the public at large with information about these ongoing efforts in order to promote journalistic and community investigation into these new programs.


Overview – Research Recommendations for a Compassionate Street Response

In order to inform the recommendations of an alternative street response, a literature review was carried out to understand the current landscape of police-alternative programs that would be relevant to San Francisco’s homelessness situation. This literature review included over 30 programs, ranging from U.S.-based programs to international programs, in an attempt to draw inspiration from a wide range of experiences. In order to learn from programs that are contextually similar to San Francisco, the literature review included many programs from the West Coast of the U.S., many specific to California, and a few specific to the Bay Area.

Many similarities existed between the model programs. In effect, all of the programs fell within one of three broad groupings:

  1. Police-alternative dispatch response;
  2. Police-supplemental dispatch response; and
  3. Police-alternative community strengthening response.

Police-alternative dispatch responses involve rerouting 911 and 311 calls that had previously been dispatched to police to teams of other professionals with absolutely no police officer involvement. Police-supplemental dispatch responses involve pairing police officers in interdisciplinary teams with social workers, medics, mental health professionals, and sanitation teams.

San Francisco has already attempted the police-supplemental dispatch to some degree with HSOC that has increased the coordination between SFPD and the HSH’s HOT team. Unfortunately this has resulted in the intensification of policing in terms of property confiscation, increased police resources, and increased distrust among San Francisco’s unhoused not only towards police, but also HOT team workers according to our survey interviews and outreach.

Based on San Francisco’s past experiences of utilizing supplemental police responses, coupled with the city’s commitment to develop a true police alternative to non-violent homeless response, the recommendations herein draw upon the model programs implementing police-alternative programming. Furthermore, the recommendations herein highlight best practices across the police-alternative model programs, tailoring them to San Francisco’s context.

CAHOOTS Model Explained

Crisis Assistance Helping Out On The Streets

CAHOOTS is a community-based mobile crisis response program operating in the City of Eugene, Oregon. CAHOOTS was launched by the non-governmental Health Center White Bird Clinic in 1989. The CAHOOTS program dispatches two-person teams consisting of a medic (a nurse, paramedic, or EMT) and a crisis worker to address mental health crises, as well as other conflicts or crisis situations including situations related to homelessness, substance use, and basic medical needs. The teams rely on trauma-informed de-escalation and harm reduction techniques to provide a non-violent resolution of crisis situations. Incoming calls come through Eugene’s 911 system or the police non-emergency number. At dispatch, staff is trained to recognize non-violent, behavioral-health-related situations and route those calls to CAHOOTS. CAHOOTS is funded through Eugene’s Public Safety Budget and is estimated to save the city of Eugene around $8.5 million in public safety spending annually.

The CART Program

Scope of CART : A Police-Alternative Response & Community-Strengthening Hub

The scope of the CART program includes two components. The first provides a specialized police-alternative response to calls from and calls for unhoused neighbors in crisis. This component draws upon best practices across various CAHOOTS-based model programs (e.g. Eugene, Denver, Dallas, Portland).

The second component serves as a community-strengthening hub to empower housed neighbors to more “compassionately respond” directly to their unhoused neighbors, drawing upon various community-based model programs (e.g. Oakland, New York City, Sacramento, Durham, Montreal, South Africa). The two-prong scope of the CART program will divert a significant number of homelessness-related calls away from SFPD, while building capacity within San Francisco’s neighborhoods to de-escalate and compassionately resolve homelessness-related conflicts directly between neighbors, thereby reducing the total number of homelessness-related calls made to dispatch in the first place.

The CART program’s scope therefore addresses the excessive cost and proven ineffectiveness of policing San Francisco’s homelessness community with its two-prong approach (BLA report). The geographic scope of the CART program is the City and County of San Francisco.

Core Values of the CART Program

The CART program embodies five core values in order to successfully implement its scope.

  • demonstrate compassion toward San Francisco’s unhoused community.
  • provide a direct source of support and help for unhoused community members.
  • build trust and uphold the dignity of all members of the community.
  • provide services to the community that are free of charge, confidential, voluntary, non-threatening, non-punitive, life-affirming, decriminalizing, and participant-centered.
  • increase community resiliency by empowering community members with resources to resolve their conflicts directly, rather than believing that their only choice is to outsource their interpersonal conflicts to the city, i.e. calling dispatch.

These values reflect the responses from the survey among San Franciscans currently experiencing homelessness, as covered in the Street Survey section of this report. These core values are also embedded across various model programs: Eugene, Denver, Dallas, Portland, Montreal, South Africa, Oakland, New York City, Sacramento, Durham.

Services of the CART Program

The CART program will offer a variety of services to fulfill its two-prong scope. These services broadly fall within “Persons-in-crisis” services and “Community-strengthening” services. The “Persons-in-crisis” services will include: first aid and non-emergency medical services, substance/addiction referrals/resources, acute/subacute transportation services (e.g. hospitals, service providers, etc.), de-escalation intervention and interpersonal conflict resolution, street counseling and mental wellness referrals/resources, suicide prevention, housing referrals/resources, as well as referrals/resources for other homelessness-related issues present in the community. The “Community-strengthening” services will include: identifying, networking, and supporting neighborhood-based “compassionate responders” (e.g. non-government organizations, civil society organizations, faith communities, businesses, natural leaders within the community), offer trainings to these “compassionate responders” to educate them on the systemic causes of homelessnes and to equip them with the relevant “Persons-in-crisis” skills and referrals/resources. The CART program will also document and disseminate lessons learned related to San Francisco’s compassionate response to its unhoused community to inspire and catalyze a greater level of compassionate responses across the community.

Skills of the CART Program

The essential skills required to successfully implement the CART program fall into four categories:

  • crisis-response skills
  • community-development skills
  • interpersonal skills; and
  • language skills.

Crisis-response skills are vital to help stabilize distressed persons, in order to provide additional support. These skills include trauma-informed care, de-escalation techniques, suicide prevention, street-level counseling, and first-aid and non-emergency street medicine expertise. These crisis-response skills are present in the various CAHOOTS-based model programs and have proven to be effective in existing San Francisco-based programs (e.g. CARE Ambassadors).

Community-development skills are necessary to increase resiliency through community strengthening. These skills include the ability to educate the public on the systemic roots of chronic homelessness, community engagement, and negotiation skills, as well as a knowledge of relevant referral resources. Interpersonal skills are essential to positive interactions with community members and include: listening, caring, empathy, understanding, patience, and the use of people-first language. These community-development and interpersonal skills are present throughout the CAHOOTS-based and community-strengthening model programs. Additionally, language skills (e.g., English, Spanish, Chinese, etc.) will help overcome potential language barriers in the multilingual setting of San Francisco.

Staff, Recipients, and Partners of the CART program

The core staff of the CART program should be of an implementing non-government organization (NGO). This separation from a government body is in line with the model programs in Eugene (CAHOOTS), Sacramento and Oakland. The CART program should primarily employ people with lived experience (e.g. formerly homeless people, with priority for Black, Latinx, and trans staff). This priority reflects the responses from the CART survey among San Franciscans currently experiencing homelessness, as covered in the Street Survey section of this report. People with lived experience are equipped with a wealth of knowledge and understanding that is indispensable to the work demanded from CART staff and therefore central to the response effectiveness, as confirmed by the model programs in Sacramento and Oakland. Those with lived experience should be paid no less than those with professional training, and both must be paid solid living wages to ensure continuity in the work.

The CART program should consist of two general categories of staff:

  • “CART Response Staff” equipped with crisis-response skills and
  • “CART Community Engagement Staff” who will be tasked to work on community engagement and education, and will be equipped with the necessary community-development skills.

These two categories of staff are in line with the combination of the model programs that inspire the scope of CART: CAHOOTS-based model programs (e.g. Eugene, Denver, Dallas, Portland) and community-based model programs (e.g. Montreal, South Africa, Oakland, New York City, Sacramento).

The primary recipients of the CART program’s services are unhoused residents of San Francisco. Other recipients are also housed residents of San Francisco via education about the context and hardships of homelessness and via training to empower them to compassionately respond to their unhoused neighbors.

Key partners of the CART program should be service providers throughout the city, such as harm reduction centers, SF Suicide Prevention, Mental Health Association, and shelter providers. Other partners include neighborhood associations (e.g. Rad Mission Neighbors, Cole Valley Haight Allies), advocacy and activist groups (e.g. Coalition on Homelessness, Copwatch, Shelter Grievance Advocates), as well as faith communities and local businesses. The CART program should also be in coordination with city agencies (e.g. Board of Supervisors, HSH, DPH, DPW, SFPD, SFFD, HSOC).

Staff Training of the CART Program

Training of CART staff will deliberately be much more intensive than a typical street outreach team training. Key to the success with the CAHOOTS program in Eugene is the 500 hours of field training that occurs for staff, as well as 20 hours of classroom training and regular follow-up training sessions. For CART, this will be even more key, as CAHOOTS relies on professional level staff, often those with EMT certificates and other formal training. For San Francisco, CART is prioritizing hiring those who have lived experience with poverty and homelessness; Black and Indigenous people of color (BIPOC); and transgender individuals to do this work. Part of the training of those who don’t have it already, would be getting EMT certification by working with City College and sending staff through those classes.

The general training for all CART staff would include a deep-rooted vision centering on the difference between policing and supporting, dispatch protocol training, de-escalation training, conflict mediation training, education on a framework on how to approach situations, mindfulness training on power dynamics, mental health education, as well as an education on substance use and harm reduction practices. The training would also include an education on all the services CART provides including the homeless service and behavioral health system in San Francisco and access points, coordinated entry, overdose prevention, first aid and CPR. In regards to community engagement, the curriculum would

include training on community engagement, consensus building, negotiation, and community-based education. Finally, CART staff would be trained in active listening, gender and race, cultural competency and homeless culture competency, including discussions on intersectionality.

The training would also include an education on all the services CART provides including the homeless service and behavioral health system in San Francisco and access points, coordinated entry, overdose prevention, first aid and CPR. In regards to community engagement, the curriculum would include training on community engagement, consensus building, negotiation, and community-based education. Finally, CART staff would be trained in active listening, gender and race, cultural competency and homeless culture competency, including discussions on intersectionality. The training would include group training curriculum as well as specific practice hours in the field where daily debriefs and evaluation would occur with team members to ensure growth. Group training sessions when possible will be open to other street based services such as SCRT members. In numbers, CART staff training would encompass 500 hours of field training as well as at least 40 classroom hours, a requirement that CART based off the CAHOOTS training hours curriculum.

Street appearance of the CART program

Generally, the CART program’s visual street appearance should be distinguishable and contradistinctive from the appearance of law enforcement. This is proven to be a constructive and effective feature of the CAHOOTS model program as it lowers the level of intimidation and threat perceived by recipients of the response service and helps in building trust. CART vehicles should have the CART logo clearly visible and have no sirens. CART staff’s appearance should resemble that of an outreach worker or medical professional with the CART logo visible and their name and area of expertise on their ID.

The CART Program’s Response – 24/7, Citywide, and Community Invitation

The CART program staff will work in pairs, providing a homeless-centered response, focusing on the well-being of the unhoused person rather than the complaint of the caller, an approach that is also endorsed by the CAHOOTS model program. The CART program should respond to all level C-priority homelessness-related calls, as detailed in the previous section on Dispatch. Ideally CART should be able to respond 24 hours a day, seven (7) days a week, as done in the CAHOOTS program in Eugene, Oregon. If hours are limited for any reason, or in a roll-out pilot phase, it is still important that police never respond to any level-C homelessness-related calls during off hours. As Portland, Oregon and other cities are piloting the programs with limited hours, there is concern in these cities that an uneven roll-out may undermine public confidence among both the housed and unhoused communities. Having police available and still responding to homelessness-related calls

during some hours, or in some neighborhoods rather than others could create a number of conflicts. Some callers who desire a harsher police response may wait for particular hours to file their complaints. Some callers who assume that they will be receiving a CART response may lose trust in the program if they call 911 and police respond due to CART’s limited hours or in specific neighborhoods. If CART and the SFPD respond to the same site on the same day or week during different hours, the SFPD could easily undermine the work of CART.

The CART program staff will listen to and encourage communication between conflicting parties, and give referrals to appropriate services and suggestions for actions and resolutions, an approach that the Social Mediators program in Montreal is based on. CART community engagement staff should also respond more broadly through targeted community outreach in situations of neighborhood outrage or frustration. CAHOOTS has stated that it would be beneficial if they had more staff and time to attend community forums, district and neighborhood meetings, merchant association meetings, etc. to educate residents and the business groups about homeless services, their limitations, and their response.

The CART Program’s Relation to Government

Existing police-alternative programs generally operate under two models. One are those privately-run by nonprofits, but publicly funded such as CAHOOTS in Eugene, Oregon. Others are run directly through government agencies, such as the pilots in Portland, Denver, and Dallas. CART recommends to follow CAHOOTS in Eugene in a public-private partnership model for a number of reasons.

First, a core recommendation from our survey respondents was creating an outreach team that included those with lived experiences of homelessness. Due to various restrictions and limitations of government hiring, this may prove a significant barrier, as has been the case in Portland’s pilot program.

Second, there is always the danger that a government agency, given how politicized the issue is, may stray from its original core principles and function when directed by their executive branch bosses. CART has seen this with San Francisco’s HOT team, which has, depending on mayoral control, coordinated more or less closely with the police and distributed resources such as shelter and rehab placements based on political power struggles and business complaints rather than the needs of unhoused individuals.

Third, an independent organization would help overcome the distrust of existing government agencies that was expressed by a number of those unhoused surveyed.

The CART Program’s Relation to other Programs and Agencies

This program would be created within a system that already exists. It is important to understand with the creation of CART, how this would be differentiated from other services, how this fits into the current landscape of street-based services, and ways they could shift. This is not meant to be a comprehensive list of street based services, but captures the main effort:

Program NameTarget PopulationHow response is activated# unhoused people servedStaffServices Provided
STREET MEDICINE TEAMPeople experiencing homelessness with medical, mental health, and substance use needsNo Direct Call Response. Mostly scheduled outreach.1,200 encounters/mo. Annually, 3,000 unduplicated patients with 15,000 encounters.8 FTE, including 0.5 Behavioral Health Clinician and 2 RNsProvides outreach with SFHOT and special outreach on an urgent basis. Outreach teams work in small, neighborhood-based teams.

Provides assessment of need for emergency care; episodic care for acute medical conditions and exacerbations of chronic conditions to the extent that care can be delivered safely in the outreach setting.
EMS6high users of emergency medical services(1) Requested by first responder units
(2) Initiates outreach with individuals who meet their criteria 

6am-2pm daily
6,413 encounters in 2020with 1,681 unique individuals$2.3m for FY20-21.Immediate stabilization of medical emergencies and behavioral health crises, Sobering Center, Shelter, Joe Healy Detox, Psychiatric Services, Substance/Dual Diagnosis Treatment
Consult, Outreach, Case Conference, Housing Assessments
Comprehensive Crisis Services (including Mobile Crisis)individuals experiencing an acute behavioral health crisis and (*not limited to PEH)(1) Hotline
(2) Respond to calls from SFPD to assist CIT officers needed
(3) No Proactive Outreach.

Availability: 24/7
400 adults and 330 children per year33 FTEReferral to Case Management; AOT; 5150 to PES; Emergency crisis assessment/intervention services conducted in the field; Early intervention; Assistance with linkage to outpatient mental health services; 5150 evaluation capacity; Short-term medication services may be available

CCS also includes Child Crisis 5585 response
HSOC – COVID Responseresolving encampments not meeting safe
sleeping guidelines
311 and 911 OnlyHSOC made over 700 placements from June 10 – August 15, 2020Data not available*Note: HSOC role has shifted during COVID. 

Makes offers of sleeping alternatives, including congregate shelters, safe sleep sites, Shelter In Place hotels, etc. When safe sleeping alternatives are not available, HSOC does
outreach, education and cleaning.
CART (Compassionate Alternative Response Team) – PROPOSEDHomeless people on streets and in shelter911, 311 and direct dial

TBD/respond to 65k callsTBDFirst aid and non-emergency medical services, substance/addiction referrals/resources, acute/subacute transportation services (e.g. hospitals, service providers, etc.), de-escalation intervention and interpersonal conflict resolution, street counseling and mental wellness referrals/resources, suicide prevention, housing referrals/resources, as well as referrals/resources for other homelessness-related issues present in the community,  identifying, networking, and supporting neighborhood-based “compassionate responders”
HOT (Homeless Outreach Team)homeless people living on the streets(1) 311
(2) Direct outreach to HSH
(3) Proactive Outreach

No weekend hours

Note: Until 2020, HOT could be contacted directly.
2,054 in Sep 2019

1,450 in Dec 2019
Data not availableShelter Placement, Coordinated Entry Assessment, Problem Solving Conversation, Homeward Bound, Resource Or Drop-in Center, Shelter Services, Navigation Services, Mental Health Services, Benefits Services, Substance Abuse, Medical Services, General Services (Shirts/ Pants, Socks, Shoes, Food/ Snacks, Water, Hygiene Kits, Emergency Blankets, Face Masks
Street Crisis Response TeamAdults in nonviolent crisis on the streets911 Only

Plans for 24/7 coverage in 2021
Estimated ability to respond to 11,000 – 17,000 calls for service per year with six teams

Full rollout by March 31, 2021 will be 6 teams. Each team has 3 staff.
Trauma-informed clinical interventions and care coordination
Felton Institute – Felton Engagement Specialist Teamchronically homeless, substance use disorder, behavioral health issues, & justice impacted311 Only. No Proactive Outreach

FEST responds to dispatches directly from HSOC and 311.
10-15 per day5.0 FTELink and refer clients to external agencies that offer temporary and sustainable housing, and support for substance use, mental health, and physical health needs.
Code Tenderloin – CARE and Calming Corners AmbassadorsHoused and unhoused adults in Tenderloin and parts of SOMAProactive Outreach. Can send team members in response to calls from hotline, but that is not the intention.3,000 /mo2 case managers. Street Teams approx 15
No direct placement authority or connection to City Resources. Connect with services through Case Management, ShelterTech, or live problem-solving. Ambassadors often bring people to Access Points. Provide consults with medical staff.

Homeless Outreach Team (HOT)

The Homeless Outreach Team is operated by non-profit Heluna Health and contracted by HSH. Currently, HOT does a mix of ongoing case management style support for individuals on the streets and responding to complaints. They typically have access to some level of housing options for those on the streets, which varies depending on availability. These include shelter and navigation beds, stabilization rooms, and during COVID, Shelter-In-Place (SIP) hotel rooms. They also are able to do mobile Coordinated Entry assessments and placements, which is the primary avenue for permanent housing in San Francisco. Until late 2019, HOT could be contacted directly by members of the public or people in need of services, but that is no longer an option. In December 2019, HOT reported 1,450 encounters with unhoused people, 84% of whom were not part of an encampment.

Over time, HOT has been used by various political administrations to respond to complaints, assist with sweeps and engage in activities that have diminished trust among the unhoused community. During Mayor Ed Lee’s administration, for example, the workers had lined up individuals for beds in navigation centers based on need, when the Mayor’s office had them switch gears and instead offer those beds to residents residing in the Embarcadero area where the Super Bowl parties were being hosted. Homeless people and their supporters expressed deep frustration with this move and it led to mass protest. With the creation of CART, HOT could focus exclusively on case management-style support, and separate the offering of services from complaints. Our system which currently distributes resources based on complaints leads to great inequities —gentrifying neighbors tend to be the squeaky wheels and get a response from the City, leading to the offering of services to homeless people in those areas over homeless people residing in more impoverished areas such as the Bayview. HOT and CART would be in communication with each other, as CART could handle the front-end work addressing the immediate issues on the streets, while HOT could provide any ongoing support to the individual needed.

Healthy Street Operations Center (HSOC)

HSOC was developed to be a multi-department collaborative to address tent encampments, and more specifically the response to complaints about large encampments and the removal of them. It grew out of the Police Department, which led the unit until recently, when the Department of Emergency Management (DEM) took over. The focus on tents has been criticized, because resources have been prioritized for those “lucky” enough to own a tent over others quietly suffering without. In addition, in many HSOC operations tents have been removed without adequate provision of alternative housing arrangements, and have been accompanied by illegal property confiscation and improper noticing. The same inequities of a complaint-based system take place. Lastly, in a 2019 evaluation of HSOC’s success, management reported that of the HSOC encounters, only 2.4% of police encounters and 17% of HOT Team encounters resulted in service connections. And of the clients placed into HSOC shelter beds, 619 (95.2%) returned to the street, based on HSH data.

In early 2020 before the COVID-19 emergency, HSOC management planned to refresh the HSOC charter and operations plans. Although the HSOC model has shifted to respond to the ongoing health emergency, that “refresh” has not happened formally and the future of HSOC is in limbo. With the creation of CART, HSOC would no longer be necessary. CART would shift from a politically driven response to encampment removal to a humane community-driven approach. CART would work with both housed and unhoused neighbors to mitigate harm, address conflict, and of course work with folks on the streets to connect them with services. In contrast to HSOC, unhoused people would have agency alongside housed neighbors. There would not be a decrease in the number of people moved from the streets into housing — that would occur at the same rate, as that is entirely determined by resources that are available. Those resources, however, would be distributed more equitably based on need.

Street Crisis Response Team (SCRT)

The Street Crisis Response Team as described earlier in this report, is a three-person team made up of a paramedic, peer and clinician, with overall management from the SF Fire Department. This team is expected to respond to 600 calls per year in particular geographic areas of the city, for those in psychiatric distress. CART would respond to overflow 600 calls, along with other homelessness-related calls. The two teams would work in close collaboration, with co-training occurring.


This body was the precursor to SCRT, and is essentially a paramedic team devoted to responding to dispatches regarding high users in the system. CART does not anticipate any changes to the operations of this team with the arrival of CART, however there does need to be tight coordination to ensure the teams are not duplicating each other’s efforts.

Other Street Outreach Teams

There are a variety of street outreach teams, such as Glide’s, Code Tenderloin, Night Ministries, At the Crossroads, Larkin, Homeless Youth Alliance, St. James Infirmary, and many more who provide a variety of services from harm reduction, spiritual support, service offering, and more. These teams would likely not be impacted by CART, but collaboration could occur with individual cases.

Overwhelmingly, the City does not coordinate with these nonprofit-operated street outreach teams, and these teams do not have access to the City’s Homelessness Response System (shelter and housing referrals, the ability to conduct Coordinated Entry primary assessments, or the ability to leverage data systems like RTZ, a comprehensive information system that coordinates, tracks, and manages adult service referral and utilization of City funded services). As a result of this lack of integration, most street outreach teams have to rely on their personal relationships with City staff to gain access to City shelters or housing, or outreach staff have to jump through hoops to connect clients to City services. This is a weakness of the City’s current system of homeless response, as thousands of interactions every month cannot lead directly to connection to City services. Further, various City departments fund these organizations at different levels and with different intended outcomes, making street response even more disjointed.

Street Medicine Team

This is one of the key services that the City and County of San Francisco provides to those residing on the streets, by bringing medical and behavioral health medicine directly to those on the streets. The Street Medicine Team is overseen by the Department of Public Health and its Health Commission. The creation of CART would not impact this work directly, however, it will be key for CART to work closely with Street Medicine as they identify those in medical distress on the streets, especially those with chronic conditions outside of care.

Mobile Crisis Team

Mobile Crisis is a team out of the Department of Public Health made up of a diverse, multidisciplinary staff providing psychiatric crisis intervention services for adults located in the City and County of San Francisco. The team provides crisis assessments and interventions including 5150 evaluation and determination of appropriate level of care. They can also medicate on site. Because SCRT is handling 800 calls, CART suggests that Mobile Crisis respond to 5150 calls and continue responding to addressing trauma in the immediate aftermath of gun violence and similar situations. It would help to limit their scope as currently they do not have near the capacity to respond to the calls they get currently.

San Francisco Police Department (SFPD)

As noted in this report, currently SFPD is the primary response to street homelessness. CART will significantly decrease these police responses, as they will no longer be dispatched to homeless related C priority calls. CART will handle most of these calls, but those that are superfluous will not be responded to at all as determined by the CART team. For example, racist or other hate-based calls, often referred to as “Karen/Karl” calls would not be responded to. If CART does not respond, SFPD will not be brought back in to respond.

Of all the non-SFPD agencies surveyed and researched as part of this report, only one that is supported by the City responds to direct calls from people needing services or the public (as opposed to being routed from 311 or 911), and the target population for that program is people experiencing mental health crises (Department of Public Health’s Mobile Crisis Team).

Housekeys Not Handcuffs protest in front of City Hall.

The CART Program Empowers Neighbors to More Compassionately Respond

In line with the program’s scope, the CART program will serve as a community-strengthening hub to empower housed and unhoused neighbors to more ‘compassionately respond’ directly to their unhoused neighbors. This will be done by identifying and training “Compassionate Response Community-Extension Volunteers” to be a compassionate presence in their neighborhood and to promote the use of the CART program to housed and unhoused neighbors. In many cases, the “Compassionate Response Community-Extension Volunteers’” will consist of community members who are already supporting their unhoused neighbors (e.g., non-government organizations, civil-society organizations, faith communities, businesses, natural leaders within the community), as well as interested community members who are well positioned to compassionately respond to their unhoused neighbors.

The CART program will thereby support these “Compassionate Response Community-Extension Volunteers” in creating or amplifying existing practices that effectively respond to community street conflicts without relying on the police or calling dispatch. These volunteers will use their personal relationships, social networks, and knowledge of their communities to resolve neighborhood conflicts involving unhoused neighbors, thereby providing a neighborhood-based, public health-oriented approach to conflict reduction.

The CART program will work to de-stigmatize a person’s housing status (e.g. homelessness) through the use of social media, relevant materials, and highlighting real voices of persons experiencing homelessness, as opposed to stereotypes (e.g. elevating voices of those who are unhoused in community spaces). The CART program will also facilitate conversations between housed and unhoused neighbors in confrontational (crisis/short-term) and community-strengthening (mid to long-term) situations. In confrontational situations, CART staff will mediate conversations between unhoused neighbors and the business owners or housed residents who called in a complaint to dispatch, drawing upon the example of CAHOOTS in Eugene. In community-strengthening situations, the CART program will build trust with service providers, neighborhood groups, and local neighbor leaders (housed and unhoused) by equipping them with updated information and resources, and encourage them to directly call the CART program, as needed. The CAHOOTS program in Eugene has stated that their program would be strengthened if they had a stronger link to the community more broadly, resulting in community members reaching out to the CAHOOTS program directly, as opposed to calling dispatch.

Oversight and Evaluation of the CART Program

Although it’s recommended that CART be managed by a non-governmental entity, it’s funding through public funds should not put it outside of government and community oversight. Just as San Francisco’s Local Homeless Coordinating Board currently receives quarterly updates on HSOC and regular reports on other aspects of the Homeless Response System, it should also receive regular reports about CART.

As CART is rolled out it will also be important to collect and analyze data on a regular basis. This could be done by the Controller’s Office. There are two examples of reporting by the Controller’s Office that were reviewed for this report: (1) The Controller Office’s independent 6-month and 1-year evaluation of the Navigation Centers, which relied on intensive data collection, various reports, and a partnership with UC Berkeley’s School of Social Work could serve as a model; and (2) Until December of 2019, the Controller’s Office was also responsible for maintaining a public HSOC dashboard that was updated monthly and could serve as an alternate reporting model. Other partnerships with local universities such as the UCSF Benioff Initiative, SF State, or the University of San Francisco may also be considered. Portland, Oregon is currently partnered with an interdisciplinary team of researchers at Portland State University to evaluate its program.

The data and ongoing analysis should be shared in a public venue, like the Local Homeless Coordinating Board, to allow for public feedback. It is recommended that the data also be shared alongside data on the proposed outcomes of the CART Program, as well as data for other City-funded street outreach programs.

Sophia Thibodeaux speaks at the Housekeys Not Handcuffs protest in front of City Hall.

Measurable Outcomes of the CART Program

Beyond meeting the core values and providing the services outlined above, the CART program will be focused on accomplishing the following measurable outcomes:

  1. Reduce the number of police dispatches to homelessness-related quality-of-life complaints where other criminal activity is not present. The Police Commission passed a unanimous resolution that the SFPD should not be the first-responders to homelessness in San Francisco. This past summer’s uprisings against the over-policing of people of color included widespread calls to defund or disaggregate the police, especially in response to mental health crises and homelessness related quality of life complaints where other criminal activity is not present.
  2. Reduce the number of individuals transported to the emergency department for low acuity medical-related issues that could instead be addressed in a pre-hospital care setting.
  3. Reduce the number of behavioral health and lower acuity medical calls traditionally responded to by the Police and Fire Departments.
  4. Reduce the number of homelessness-related calls to dispatch, in areas where the CART program’s community-strengthening interventions have occurred.

CAHOOTS and other stakeholders in pilot programs across the country have stressed the importance in communicating to the public that such police-alternative responses are not designed to solely end homelessness or reduce visible homelessness. Such outcomes require broader system-wide investment in making such housing, shelter, and services available. Nor are such responses aimed at getting people to move or be removed from public spaces. The CART program outcomes are aimed at reducing police interactions with those experiencing homelessness, and in the process deescalate conflicts, provide immediate medical and counseling assistance, connect people to available services, educating those experiencing homelessness about their rights, as well as residents and business owners about homelessness and the city’s homeless response.

Impact if the CART Program is Not Implemented

If the CART program is not implemented, unhoused San Franciscans will further be entrapped into a life of chronic homelessness as increased numbers of interactions with law enforcement are highly correlated to prolonged homelessness. Unhoused, vulnerable San Franciscans will likely not receive the compassionate care they need to survive homelessness, and instead be further criminalized and stigmatized for their housing status. Furthermore, if not implemented, people experiencing homelessness will continue to lack a direct line of contact to call when there is an immediate issue or when they feel unsafe. Finally, many of San Francisco’s neighborhoods will continue to have low levels of community resilience in the face of homelessness, if the CART program is not implemented.

CART Deployment Timeline With Phases

Budget and Finance Committee release $2 million on reserveFebruary 2021
Lay out program budget including program and training costsFebruary 2021
DPH craft and release RFP $2 million based on this proposalStarted in January, RFP ready to be released by time Budget and Finance releases funds in February.
DEM assigned working group to rework Dispatch training and start differentiating housing status and instructions to route dispatches to CART January 2021
DPH awards RFPMarch 2021
Set up working tables with parallel groups such as SCRT, Mobile Crisis, EMS6March 2021 as soon as funding is awarded
Budget Process: move budget dollars from SFPD to CARTIncluded in Mayor’s Budget, June 1 2021
Program awarded funding designs training curriculum with technical assistance from CAHOOTSMarch – April 2021
Program awarded fundings designs job descriptions, and starts recruitment and hiringApril and May 2021
Program is operationalMay 2021

Narrative change

Over time, due to the consistent and vociferous scapegoating of unhoused people, something important has been lost in the collective consciousness of many San Franciscans. Homelessness has systemic causes — racism, disablism, classism, and greed are drivers of the defunding of housing and neglect of the very poorest members of our society. Yet, that narrative is buried deep beneath the easier belief that people are homeless because of their personal faults and shortcomings, and, even more absurd, by personal choice. These beliefs run thick and are difficult to shift. Just as according to a Fox news poll, 36% of Americans believe that Trump had the election stolen from him because some of our leaders and elements of our traditional and social media seeded and reinforced this belief, San Franciscans have had this personal blame belief seeded and reinforced over many years. This belief has created obstacles to opening new housing and shelter programs for unhoused people as neighbors attempt to block developments based on the belief that they will create safety issues or lower their property values. It has also meant that some San Franciscans believe police will be able to address homelessness — by making it uncomfortable for those on the streets, they will leave this lifestyle and choose another. This has led to increases in complaints, and even exaggerating descriptions of behavior to ensure a police response when conversing with 911 operators. Changing this narrative must occur deliberately and thoughtfully.

In the rolling out of CART, a public education campaign should happen simultaneously that explains why a police response is ineffective, inappropriate and inhumane. In addition, the successes of the new team should be amplified. City websites that give conflicting messages, such as the before and after photos of tent encampments on the DPW website should be redesigned to stop communicating that displacing tents is acceptable and instead that relocation to housing is the goal. Policy makers should be mindful of how they talk about homelessness and reflect on whether their messaging is reinforcing negative stereotypes and halting progress on addressing homelessness. Operators at 911 and 311 should be trained on how to talk about these systemic issues and let people know that until housing is available for everyone, we will have impoverished people on the streets. Lastly, these conversations can happen directly with housed neighbors as CART is deployed. CAHOOTS in Oregon has many of these conversations, managing expectations and trying to address the key issues that are presented.

New number to call CART

When CAHOOTS was asked “If you did anything different when you created this program, what would it be?”, they responded that they would have had a separate line from the start. We recommend such a phone number for CART as well. CART would take calls from dispatch and also take direct referrals. This would allow unhoused people and housed people who are uncomfortable calling the city, fearing that it will lead to a police response, to call without risk. Many in San Francisco, for a variety of reasons — fear of City interactions because of immigration status, lack of faith in government entities, distrust of police and more — would prefer to be able to call a program for help directly.

Funding should come from additional cuts to the police budget

After a nationwide uprising began this summer calling to defund the police, Mayor Breed proposed a mere 2.6% cut to the SFPD’s 2020-2021 budget, which Supervisor Hillary Ronen referred to as “a slap in the face.” [8] At an SFPD budget hearing in August, supervisors questioned why SFPD based their anticipated funding needs for SFO International Airport on data from 2018, when the pandemic has seen a drastic reduction in the number of people flying in and out of SFO. Supervisors also questioned the need for mounted officers on horseback, which the department justifies as good for morale and public relations. When most SF government departments are expected to continue functioning with a 10% budget cut in 2020-2021 despite the increased needs resulting from the pandemic, why should the police department continue to monopolize 10% of the city’s budget?[9]

As of now, CAHOOTS in Eugene, Oregon is the only citywide police-alternative in operation. While it is the goal of Portland, Denver, Dallas, and other cities to expand their programs, they have each started or are starting as smaller pilot projects. Although an initial pilot may be useful in scoping and budgeting a citywide program It is our hope that the political and popular momentum of the current moment, the Mayor’s call to end the policing of mental health and homeless emergencies, and San Francisco’s long standing leadership in pioneering progressive social policy would be able to immediately commit to long-term budgeting in rolling out a comprehensive citywide initiative to end the policing of homelessness.

The CART program will need to be funded by a diversion of funds from the SFPD budget in order for it to be a success. This could be achieved by diverting funding currently spent by the SFPD responding to homelessness-related quality-of-life violations as a baseline for CART’s budget. While a 2015 analysis by the Budget and Legislative Analyst’s office showed this to amount to $18.5 million a year, the recent expansion of police resources dedicated to homelessness through HSOC suggest that this amount is now significantly higher. A more up-to-date analysis should be done while the first phase of CART teams is rolled out in a number of neighborhoods. This would provide more accurate estimates on the costs of scaling up.

It is morally imperative that San Francisco reduces reliance on policing to respond to homelessness when it so often leads to physical or psychic injury or death, and when other services are desperately needed. Through the process of researching and writing this report, we’ve considered what would be possible with a response to conflict on the streets or complaints about homeless people if police were not the first responders. In order to make this possible, CART needs to redirect funding from SFPD’s nearly $700 million budget to the CART response. The annual budget for Street Response in Portland Oregon, a city with one-third of the number of unsheltered homeless has secured an annual budget of $4.8 million. The annual budget for the CAHOOTS program is about $2.1 million for responding to 24,000 calls for service. In San Francisco, adjusting for higher salaries, due to higher housing costs, and adjusting for the higher number of calls (65,000 annually) to be responded to, the budget is estimated to be $6,825,000. This is already partially funded with $2 million on reserve. If this funding is released this fiscal year, to be used in full for the last two months, CART would need to annualize an additional $4.825 million to fund this program for a total annual budget of $6.825 million (assuming a 20% higher per call cost then CAHOOTS). In addition to being morally just, the CART response will redirect calls from the police department, leading to less work for officers to do. In other city departments, reduction in work leads to reduction in funding, and it should be no different for the police department.

Mural that reads “Today was a good day!” Photo by Heidi Alletzhauer.


How to share and use this report

We consider this report to not only be an educational resource for those interested in alternatives to police response to homelessness, but also an advocacy tool. We believe the more people who become familiar with our recommendations, the more we raise public awareness of the need to restructure the city’s policing system, the homelessness response policies, and the concrete ways in which this can be achieved. As members of the public and concerned San Francisco residents learn from the CART report, they will become empowered to raise their concerns and advocate for changes from the City’s leadership. Anyone can use this resource to have conversations about reducing police involvement in homelessness issues and how we can respond in a more compassionate, less dangerous, and more effective way to calls and complaints about people who are without homes.

We hope that with this tool, citizens will be more equipped to make the case to their neighbors, colleagues, family, etc. that as a City we must and we can conduct a wholesale redesign of our response to concerns about and needs of the houseless in our community. Thus we encourage readers to share this far and wide —perhaps incorporate it into teaching curriculums or bring it to your congregations and social action groups. If you are part of a community-based organization, you can link to the report on your website, post it on social media, write about it in your newsletters, and distribute it to your member lists. It is easily available to share digitally at

There is also an opportunity with this report to advocate directly with elected officials who have the power to implement the CART program model. We highly encourage people to use information or excerpts from the report when engaging with those decision makers. Set up meetings with Board of Supervisors members in your district, for instance, to review the report and explain why you think CART should become a reality in San Francisco. Send emails to the Mayor or the Police Commission endorsing the CART recommendations. Provide public comment at pertinent city hearings (for instance, budget items) armed with the justifications and recommendations from the CART report.

CART will more likely be adopted by the City if there is a greater groundswell crying for it. Communications about the report will help greatly in creating that groundswell. Please feel free to share it on social media, send letters to the editor about it, talk to reporters and news outlets, and otherwise push it out in any ways you see fit.

Additional resources

As a companion to this report, CART has created a website with additional resources and information about the proposed program. Please visit the website to learn about CART and alternatives to police research and programs in general.


We would like to give our greatest appreciation to all of the many community-based organizations, service providers and City Departments who gave many hours of staff time to help create this report and to the individuals who volunteered their time to research, plan, coordinate and write the recommendations.

Special thanks to the following city officials and department staff who provided technical assistance and resources to the CART working group:

Police Commissioner John Hamasaki

Robin Candler, SFDPH

Robert Smuts, DEM

Courtney McDonald, Supervisor Matt Haney’s office

Amy Sawyer, Office of the Mayor

Carolyn Goossen, Office of the Public Defender

Special thanks to Coalition on Homelessness and Community Housing Partnership who provided funding.

Special thanks to the many community members and organizations that provided assistance designing, proctoring and analyzing survey data including Stella Kunkat, Code Tenderloin, Mission Neighborhood Resource Center, Glide, Mother Brown’s Dining Room, Flo Kelley, Olivia Glowacki, Ruth Grace Wong, and Fellows with San Francisco District Supervisor Matt Haney.

Special thanks to Christopher Herring, PhD (Harvard University) and Professor Dilara Yarbrough, PhD (San Francisco State University) for academic support.

Special thanks to John Stiefel, Terezie Bohrer, Peter Murphy, Sakina Bambot and David Elliot Lewis for research assistance.Special thanks to TJ Johnston who provided editing services.

Special thanks to Vinny Eng, who provided design services.

Facilitator and Process Design – Patrick Brown, The Justice Collective

Administrative Support – Stella Kunkat

Committee Chairs

Research – Legislative Aide Courtney McDonald and Commissioner John Hamasaki

Dispatch – Adriana Camarena and Jennifer Friedenbach

Communications – Sara Shortt and Samantha Lew

Appendix A – SF Police Commission Resolution For Effective Response to Homelessness and Complaints Regarding Presence of Homeless People


WHEREAS, the roots of our local homelessness crisis can be traced, in part, to federal divestment from the funding of affordable housing, and local municipalities have been left to manage the problem without the tax base the federal government enjoys,

WHEREAS, local police departments have been placed by default at the front lines of the homelessness response. As homelessness continues to grow across the country, communities are increasingly turning to police to address the issue. Instead of helping people escape life on the streets, this creates a costly revolving door that circulates individuals experiencing homelessness from corner to corner,

WHEREAS, the United States Interagency Council on Homelessness (USIAC) issued a report in August of 2015 outlining best practices for addressing the presence of encampments entitled “Ending Homelessness for People Living in Encampments: Advancing the Dialogue”. This report includes guidelines on how to address homelessness and recommends that the linking of homeless people with an appropriate level of housing is the only lasting solution,

WHEREAS, the City of County of San Francisco received last year more than $44 million in federal McKinney-Vento funding for projects addressing homelessness, and the Department of Housing and Urban Development (HUD) has created funding incentives for communities to ensure that persons experiencing homelessness are not deemed criminals because of their use of public space for survival,

WHEREAS, the Ninth Circuit Court of Appeals rejected and en banc petition by the City of Boise in Martin v. Boise (formerly Bell v. Boise), upholding its September 2018 ruling that homeless persons may not be punished for sleeping outside on public property in the absence of adequate alternatives,

WHEREAS, the 2019 San Francisco Homeless Count and Survey found 5,180 unsheltered individuals,

WHEREAS, number of individuals experiencing homelessness in San Francisco far exceeds the number of available shelter beds,

WHEREAS, the number of police officers devoted to responding to homelessness has increased from 24 in 2017 to over 80 in 2019,

WHEREAS, when calls are made to the city expressing and concern for unhoused people, SFPD remains the primary agency dispatched in response,

WHEREAS, a report issued by the Budget and Legislative Analyst’s Office in May, 2016 found that police officers dispatched incidents related to quality of life laws produced limited results given the increase in homelessness on the streets and recommended that the Board of Supervisors should consider implementing a new strategy to address these issues that shifts responsibility to quality of life incidents from the police department to other City agencies, including the Department of Homelessness and Supportive Housing,

WHEREAS alternative models exist such as Crisis Assistance Helping Out on the Streets or “CAHOOTS,” a mobile crisis intervention response and support team located in Eugene, Oregon, in which health workers, instead of police, act as first responders,

RESOLVED, that the San Francisco Police Commission calls on the Board of Supervisors and the Mayor’s Office to convene a current or new stakeholder group that should include the departments of Homelessness and Supportive Housing, Public Health, the Police department, other related departments, community organizations and those personally impacted by homelessness tasked with developing alternatives to a police response to homelessness.

RESOLVED, that the stakeholders group explore alternatives to a police response that exist in the United States and other Countries and weigh the feasibility of implementation of such programs locally.

RESOLVED, that this stakeholders’ groups should identify funding sources, appropriate dispatch protocol, necessary system changes and appropriate service model to move from a police response to a more effective health and human services response to homelessness.

RESOLVED, that the stakeholders group should meet for a time limited period, and make recommendations to the Board of Supervisors, Mayor’s Office, and appropriate commissions on how to transform our response to homelessness from one led by law enforcement into one led by trained health and human services workers.

A person at a protest with a red bicycle and a sign that reads “Housing is a right not a commodity”.
Compassionate Alternative Response Team logo: a shopping cart with a heart inside.