An alternative first response program that will launch in Berkeley this month claims to be on the verge of making history: according to Samantha Russell, the Director of Crisis Services at the non-profit that will run the unit, “Berkeley will once again be making history in the way that they are running things.” The program, called the Specialized Care Unit, will, like other alternative first responders that are being piloted across the country, respond to drug and mental health-related 911 calls. However, unlike the pilot of San Francisco’s non-police first responders, the Street Crisis Response Team, Berkeley’s Specialized Care Unit will be able to enter private residences. And, unlike the pilot of Oakland’s alternative responder program, the Mobile Assistance Community Responders of Oakland, Berkeley’s unit will immediately begin operating 24 hours a day, 7 days a week. Russell hopes that the design of the Specialized Care Unit will provide more access to civilian first responders than any other program in the Bay Area.

This may sound ambitious for a program that has yet to be launched. Even though at first glance Berkeley appears to be moving more slowly than its neighbors in Oakland and San Francisco in developing a non-police responder, it has been researching and planning this program for the past three years. In 2020, in response to the nationwide protests against police brutality, the Berkeley City Council gave the green light to developing an alternative first responder system as a way to limit the Berkeley Police Department’s scope of work. The city contracted a consulting firm to write a report on the pre-existing crisis response models that were in operation at that point and release a set of detailed recommendations about the design of this program. The pilot was supposed to start last September, but policymakers pushed the date further back to June 2023.

The differences that have been pointed out between the programs in Berkeley, San Francisco, and Oakland illustrate the wide range of choices that policymakers must pick and choose from when launching alternative first response programs. Alternative first response aims to become a fourth branch of the emergency response system, complementing the existing police, fire, and emergency medicine departments. However, unlike those departments that have well-established foundations, alternative first response operates more like a start-up: the programs are often housed under another department and have a significantly smaller staff and budget, even while they attempt to make a data-driven argument to various stakeholders and the public that they are a worthwhile investment. As a result, program directors have commented that it can feel like there is “pressure built into” the early months of these programs to prove effectiveness and expand. While Berkeley’s plans may seem bold, they are in line with the kind of innovation and success that policymakers and the public expect from alternative first-response.

Complicating matters, as discussed in our previous article, is the fact that there is no default model that policymakers can rely on for building and scaling alternative first-response programs. On a recent panel on Scaling Alternative Response Programs co-hosted by the Policing Project at NYU Law and the Harvard Kennedy School’s Government Performance Lab, Andrew Dameron, the Director of Emergency Communications for the Support Team Assistance Response program in Denver, drew attention to the fact that “needs are different everywhere across the country.” Dameron advised that policymakers should “pick and choose the pieces [of different programs] that are going to work for [each] community.” This approach to alternative response suggests that understanding various design choices that have already been made can be informative, leading to a toolkit of effective alternate response design choices that can be customized for cities. 

In this article, as Berkeley prepares to launch a new alternative first response program, we will ask: what lessons have been learned so far from the implementation and scaling of the alternative first response teams in San Francisco, Oakland, and Santa Clara County? We will consider the insights that we gleaned from talking to the team behind San Francisco’s relatively mature two-year-old Street Crisis Response Team program, as well as the evolution of the Mobile Assistance Community Responders of Oakland, and the state of the Santa Clara County crisis teams that cover San Jose. Perhaps Berkeley can pick and choose from the valuable lessons that have recently been learned in the communities surrounding it in the Bay. We will also consider the common hurdles that recur in the rollout of alternative first responses, including limited funding, and hiring issues. In conjunction with these questions, this article will discuss how communities can manage the public pressure to succeed, and how they can build a clear, accurate, and data-driven case to stakeholders that demonstrates how promising these projects can be for innovating community safety.

Making a Case for Alternative Response: Statistics and Stories

One strategy to showcase effectiveness, as implemented in San Francisco and Oakland, is to publish monthly reports that are easily accessible to the public, containing data that measures the success of the program. The San Francisco Crisis Response Team currently posts monthly updates with operations updates and key performance indicators. The program in Oakland publishes monthly updates with performance indicators, as well as success stories about people to whom the program responded. Collecting and telling anecdotal stories may be another powerful way to qualitatively illustrate the work that alternative first response performs. Oakland also hosts an interactive dashboard that visualizes the incidents its teams responded to on a map, as well as the number of blankets and water bottles that were given out. These varied methods of reporting suggest that alternative first response programs may benefit from finding creative, accurate ways to effectively convey to policymakers, stakeholders, and the public the kind of outcomes that they can achieve.

Regarding which quantitative metrics to record, first response program directors from across the U.S. recommend consulting with experts at the onset of a pilot program to figure out which data will be useful to collect. Directors of the San Francisco Crisis Response Team advised first figuring out which metrics may be useful to measure and then working backward from there, reverse-engineering the design of the program to optimize for its effectiveness. Deciding on data collection protocols later on in the lifecycle of the program may result in missed opportunities, which happened with the pilot of Denver’s Support Team Assistance Response program. Initially, the program only collected data during the limited hours and in the constrained geographic region of the pilot of the alternative first responder. As such, 911 operators only flagged eligible calls when and where the program was in operation. Later, the program realized that having 24/7 city-wide information would have been helpful because it would justify the growth of the program beyond the initial limited pilot phase.

Programs vary in how much they publicly report on metrics that are not directly related to the program. San Francisco’s program includes the percentage of eligible calls that received an alternative first response in its monthly reports. This data point demonstrates growth in how, in April 2021, the Street Crisis Response Team responded to 24% of eligible calls, whereas in February 2023, after two years of rapid expansion, it responded to 72% of eligible calls (criteria for eligibility did change during this time period, affecting the direct comparability of these data, but the data does show general growth in the program). Based on public reports that we reviewed, Oakland does not publicly report incidents that were eligible but not responded to. In contrast, the crisis teams in Santa Clara County do not seem to have publicly available monthly reports about the status of their teams at all. If Berkeley’s program adheres to the recommendations that the city released, it looks like the program plans to publish a variety of metrics, including the percentage of eligible calls responded to, on Berkeley’s Open Data Portal. The variation in data-sharing practices between the Bay Area programs underscores the need for researchers and policymakers to actively work towards establishing a standardized set of best data-collection practices. 

No Impact? Troubleshooting Effectiveness 

If a pilot program does not cause the impact that one would expect, being able to take on an experimental methodology and troubleshoot the pilot is important. The Berkeley recommendations emphasize the need for a “pilot approach” in which the program would “test, modify, and expand specific aspects of each crisis response model based on program successes, challenges, and consumer feedback.” If there are no early results, experts suggest that the program should reconsider the choices that went into designing the pilot and try altering them. For example, a program may reconsider the constraints that were put on calls in order to be considered pilot eligible. Some jurisdictions will initially limit the pilot based on geography, such as the Oakland program, which initially only served East and West Oakland, and the San Francisco program which initially focused on the Tenderloin neighborhood.  Often, programs limit their pilots to operating only during the conventional working day, as it is more difficult to recruit staff for nighttime incidents. Other programs may limit the scope of their operations based on client characteristics, such as only serving those who are 18+. If a pilot does not have a perceptible impact, it may be worth widening these constraints and asking questions about how the program can expand and serve more people. This may involve providing additional training sessions for responders, forming new relationships with non-profit partners, or adding new roles, such as peer specialists, to the team.

At the same time, stakeholders should keep in mind that pilots are often initially small and thus may have a modest, gradual impact. Gabriela Solis, the Project Leader at the Harvard Government Performance Lab observed that small impacts run the risk of generating rhetoric about ineffectiveness. The question of how to manage political and public expectations of the program while experimentally building an effective model is relevant to many issues in alternative response. Even though pilot programs are foundationally dependent on showing results, the experimental and operational reality of learning how to run and operate a pilot requires patience and time. 

Recurring Obstacles to Alternative First Response: Funding and Staffing Issues

Common obstacles that occur in alternative response include having a limited budget and staffing troubles. The Berkeley program that is about to launch is working with a significantly lower budget than researchers recommended: it is currently operating on $5.5 million, which is much less than the $10 million budget that experts suggested. Funding alternative first response is a topic that has not been satisfactorily answered. Currently, local taxes are the primary funding source for most programs. The federal government has encouraged using Medicaid as a way to support mobile crisis intervention; however, Medicaid funding has not been consistent for program purposes. The Policing Project recommends considering a dedicated tax for the program and being efficient in deciding who is staffed on the program. In San Francisco, the designers behind the Street Crisis Response Team agreed that staffing a focused team is expensive, and does not save money in terms of budgeting.

Additionally, hiring issues plague many alternative first-response pilots. In our interview with the San Francisco Crisis Response Team, they revealed that agencies throughout the city are facing challenges in hiring for every behavioral health job opening. This problem is exacerbated for programs like the Street Crisis Response Team which is operating 24/7, as nights and weekends are not desirable, especially for highly trained workers. Compounding this issue is the fact that in San Francisco these roles are filled by those who are contracted through a private employer with a lower wage scale. The paramedic role, in contrast, does not have staffing issues, perhaps because it is employed directly by the Fire Department which has a higher wage scale.

An Experimental Field Rife with Research Questions

This article has provided a peek into the lessons learned so far about alternative first response, and the challenges that remain for the models that are in operation. Many questions remain for researchers and policymakers to answer about the success of these programs: What, if any, standardized practices can be created to guide collecting data, experimenting with pilots, and deciding the scope of pilot programs? How can policymakers manage the high expectations that accompany political and public scrutiny while building a brand-new program from scratch? What does a consistent, sustainable funding structure for alternative response look like? And what methods of data sharing, story-telling, and monthly reporting will help build an honest and accurate picture of the work that alternative first response entails?

In the next essay, we will take a step back from the specifics of these programs and consider how alternative first response fits into the larger systems and processes of the 911 department, including rethinking and retraining a critical role: the 911 dispatcher.