“It’s something we’re connected to”: Acceptability and adoption of overdose detection technologies implemented in San Francisco permanent supportive housing

Abstract

Introduction:

The transition from homelessness to permanent supportive housing (PSH) is associated with solitary drug use, which heightens the risk for fatal overdose. Overdose detection technologies show promise in reducing fatal overdose in supportive housing environments. We conducted a longitudinal, mixed-methods study in a 50-unit single-room occupancy PSH building in San Francisco, California—tracing the implementation of wall-mounted, push-activated technologies (Brave Buttons) that alert designated responders to potential onsite emergencies, including overdose.

Methods:

Between May 2021 and February 2022, we conducted 35 days of naturalistic observation as well as serial, semi-structured interviews with 8 building staff (e.g., housing services coordinators, front desk clerks, custodians) and 5 tenant specialists, who promoted and sensitized residents to Brave Buttons prior to their installation. We administered surveys to all residents, assessing acceptability and uses of Brave Buttons. We calculated descriptive statistics regarding Brave Button installations and activations using administrative device data. Through inductive, iterative thematic analysis, we synthesized participant narratives to identify early expectations, implementation determinants, and adaptive strategies to bolster acceptability and adoption of Brave Buttons.

Results:

Despite high demand for overdose prevention solutions, staff and tenant specialists initially expressed doubts towards the adoption potential of Brave Buttons in the PSH building, citing resident mistrust of novel technologies with surveillance properties, liability related to (failed) overdose responses, and building staff capacity and willingness to respond to Button activations. Responding to these anticipated implementation constraints, a group of staff and tenant specialists hosted “engagement sessions” to build resident affinity/trust with Brave Buttons and leveraged tenant specialists’ rapport with residents for technology sensitization. By December 2021, Buttons were voluntarily installed in 72 % of resident rooms and were used for multiple purposes, including overdoses and safety/wellness checks. The Buttons also helped formalize informal systems of care among residents and strengthened connections between residents and building staff.

Conclusion:

Adaptive strategic planning, responsive to emerging sources of implementation opposition, was critical to addressing perceived barriers to Brave Button acceptability and adoption in the PSH building.

1. Introduction

Drug overdose remains a leading cause of death in the United States, with over 110,000 overdose deaths recorded in 2023—representing a nearly 30 % increase from 2020 (Ahmad et al., 2024). Most fatal overdoses in the United States are opioid-involved, attributed in recent years to the introduction of fentanyl and other potent synthetic opioids into illicit drug markets (Ciccarone, 2019). An increasing number of overdose deaths, however, occur in the context of polysubstance use, specifically opioid co-use with stimulants like methamphetamine and cocaine (Gladden et al., 2019; Mattson et al., 2021), and there are growing concerns about sedative adulterants (i.e., xylazine) in the illicit opioid supply (Alexander et al., 2022; Gupta et al., 2023). Fatal drug overdoses constitute a public health emergency requiring innovative solutions, including novel approaches to rapidly identify and responding to overdoses.

The overdose crisis overlaps and intersects with a burgeoning housing crisis throughout U.S. metropolitan areas. The proportion of U. S. adults experiencing chronic homelessness has increased sharply in recent years (by 16 % since 2020) (de Sousa et al., 2022)—coinciding with dramatic increases in housing costs in densely populated urban areas (Desmond, 2016; Whittle et al., 2015). Individuals experiencing housing insecurity shoulder disproportionate burdens of overdose deaths, with several studies reporting excess overdose-related mortality in unhoused persons relative to the general U.S. population (Baggett et al., 2013; Fine et al., 2022; Larney et al., 2020).

The transition from homelessness to permanent supportive housing (PSH) can be marked by heightened vulnerabilities to fatal overdose, as the social support PWUD might receive in congregate living contexts like tent encampments or shelters may not accompany them into supportive housing (Dougherty, 2020; Goldshear et al., 2023; Mayer et al., 2024). This transition has also been linked to more frequent solitary drug use (Collins et al., 2020; Olding et al., 2023), which diminishes opportunities for bystander intervention in the event of an overdose (Rosen et al., 2023; Wojcicki, 2019). Studies have demonstrated that supportive housing residents experience substantially elevated risks of overdose death relative to non-residents, underscoring how the absence of responders/bystanders during overdoses and limited access to onsite naloxone propagates the fatal overdose risk environment in these congregate residential contexts (Davidson et al., 2003; Rowe et al., 2019). Expediting overdose detection and response in supportive housing is critical to preventing fatal overdoses among residents using drugs alone in their homes.

Overdose detection technologies (ODTs) are promising tools to mitigate risks of fatal overdose in supportive housing environments. ODTs encompass a variety of technologies, including wearable devices (e.g., naloxone auto-injectors), remote monitoring through mobile applications/hotlines, and fixed-location systems in high-traffic areas (e.g., contactless reverse-motion detectors, push-activated intercom systems) (Lombardi et al., 2023; Rioux, Marshall, & Ghosh, 2023). Fixed-location technologies have been installed in a variety of settings, from medical facilities to housing shelters, in North America, where they have been used to detect overdoses and other health-related emergencies (Bardwell et al., 2021; Beaugard et al., 2024; Buchheit et al., 2021). One fixed-location ODT, the Brave Button (Brave Technology Co-op, Vancouver, British Columbia, Canada), is a wall-mounted, wireless button designed for single-room occupancy (SRO) housing that can be pressed to activate alerts to designated responders in emergency situations, including overdoses. Brave Buttons facilitate rapid response to suspected overdoses and other emergencies while respecting end-user privacy and confidentiality (Lombardi et al., 2023); because designated responders are only alerted when Brave Buttons are activated (in most circumstances through an end-user pressing the technology), Brave Button utilization is voluntary, which can protect end-user privacy during responses (e.g., alerting building staff to nearby violence). While difficult to quantify the full impact of Brave Button installations, recent research suggests that the over 450 Brave Buttons implemented in SRO housing across North America have averted at least 100 overdose deaths to date (Welwean et al., 2023).

Despite the great promise of fixed-location ODTs like the Brave Button to address surging overdose deaths in housing environments, their relative novelty and subsequent unfamiliarity to residents and housing providers can pose challenges to their implementation. The acceptability and voluntary adoption of overdose detection technologies are of particular concern within supportive housing contexts. Brave Button activations prompt welfare checks from designated responders and can, therefore, result in unwarranted disclosure of residents’ drug use to staff and other residents, with potential adverse implications on residency contracts in some settings (Bardwell et al., 2019; Ivsins et al., 2022; Manson et al., 2022). Misconceptions of the Brave Buttons’ passive monitoring technology, specifically that the Buttons can record audio and video in resident rooms, could interface with broader concerns surrounding heightened surveillance of PWUD in supportive housing settings (Collins et al., 2019; Michaud et al., 2023)—further hindering the acceptability and adoption of these life-saving technologies.

In specific housing contexts, like SRO PSH buildings in San Francisco, overlapping social (e.g., mistrust between staff and residents), structural (e.g., punitive housing policies), and environmental (e.g., building infrastructure) forces may hinder Brave Button implementation efforts (Bardwell et al., 2019; Nowell et al., 2020). Overdose deaths in San Francisco have historically clustered in the city’s 500 SRO buildings, where nearly one-third of the city’s fatal overdoses occur (Thadani & Palomino, 2022). Additionally, substance use-related stigma, building policies governing onsite drug possession/use and naloxone distribution, gender-based violence, and poor living conditions amplify this overdose risk environment in supportive housing (Bardwell et al., 2019; Collins et al., 2018; Knight et al., 2014). These co-occurring factors may deter residents from voluntarily installing and using Brave Buttons for various harm reduction purposes, including overdose prevention/response and other safety concerns in PSH environments. Prior scholarship has identified enabling and constraining determinants of Brave Buttons in North American housing contexts (Bardwell et al., 2021; Ivsins et al., 2022), but this literature has investigated overdose detection strategies at a singular timepoint, either prospectively (before installation) or retrospectively (after an implementation run-in period). There is, thus, limited evidence regarding how housing providers can anticipate and effectively respond to contextual dimensions of the housing environment shaping the acceptability, uptake, and use(s) of ODTs as harm reduction tools.

1.1. Study setting and context

In this longitudinal study (May 2021 to February 2022), we used ethnographic methods (Gertner et al., 2021) to explore context-specific phenomena underpinning the anticipated and realized acceptability and adoption of the Brave Buttons prior to and after their installation in a 50-unit permanent supportive SRO building in San Francisco’s South of Market (SoMa) District, a neighborhood in which 18 % of the city’s 640 overdose deaths occurred in 2021 (Jung, 2021). Like elsewhere in the United States, fatal overdoses in San Francisco, which disproportionately affect Black and Hispanic PWUD, rapidly rose with the introduction of fentanyl and its synthetic analogs into the illicit drug supply (Coffin et al., 2022). Overdose deaths further surged with the onset of the COVID-19 pandemic, increasing at a weekly rate of 50 % between March and November 2020, coinciding with shelter-in-place ordinances that likely increased drug use frequency and solitary drug use events (Appa et al., 2021), in part to cope with socio-economic uncertainties and psychological distress during shelter-in-place ordinances (Compton et al., 2023; Frueh et al., 2023; MacKinnon et al., 2020; Schneider et al., 2023). The SRO building included in this study consisted of private rooms with shared bathrooms and exclusively housed individuals referred through the city’s Coordinated Entry System for homelessness services. The building was smaller and newer than most PSH buildings in SoMa, with most tenants referred from the same emergency shelter within the last two years (Olding et al., 2023). With oversight from the City of San Francisco’s Homeless Services, the building offers case management for mental health and substance use, as well as auxiliary services (e.g., food vouchers, wellness programming, community-building activities) for residents. Using mixed methods (i.e., naturalistic observation, serial in-depth interviews and surveys with PSH residents and staff, secondary data analyses of Brave Button installation and activation-response registers), the present study aimed to identify adaptive implementation strategies that promoted acceptability and adoption of the Brave Buttons in the PSH environment.

2. Methods

2.1. Brave button installation

Installation of Brave Buttons in the 50-unit SRO building coincided with the implementation of the SRO Project, a joint initiative of the San Francisco Department of Public Health and the Drug Overdose Prevention and Education (DOPE) Project to address overdoses in supportive housing (Olding et al., 2023). Briefly, the multi-phase SRO Project involved training building staff (i.e., case managers, front desk clerks, property managers, custodians) on overdose response and installing wall mounts stocked with nasal formulations of naloxone on each floor of the building (Phase 1, December 2020 to June 2021); recruiting and training building residents—prioritizing those networked with PWUD—to become compensated tenant overdose specialists, who would distribute naloxone and educate other residents on overdose response (Phase 2, June to August 2021); and scaling-up overdose awareness and response capacities onsite through tenant specialist outreach to residents, routine restocking of naloxone wall mounts, drop-in ‘office hours’ in building common areas, and on-demand naloxone trainings (Phase 3, August to December 2021). Specialists received $50 weekly in gift cards as compensation for their participation.

Brave Button installation began in September 2021, following SRO building staff and specialist trainings (see Fig. 1). Specialists served as Brave Button ambassadors, educating residents about the technology and promoting their uptake by disseminating advertisements (e.g., posters, flyers) to building residents and posting them throughout the building’s common areas. Building residents had to voluntarily request Brave Button installations in their rooms.

Fig. 1.

Brave Button demonstration and training materials. Notes: A) Instructions for building staff on responding to a Brave Button activation. B) Pamphlet of instructions for tenant use of Brave Buttons. C) Sample communications of activated Brave Button to responder phone. D) Example image illustrating the installation and size of the Brave Button. All materials are courtesy of Brave Technology Co-op.

2.2. Study design and procedures

Between May 2021 and February 2022, the research team conducted ethnographic fieldwork in the PSH building to generate nuanced understandings of the SRO Project’s implementation trajectory, including the installation of Brave Buttons (Olding et al., 2023). Ethnographic fieldwork consisted of three data collection modalities: (1) unstructured, naturalistic observation; (2) longitudinal semi-structured, in-depth interviews; (3) a structured survey administered in December 2021 to all SRO building residents eliciting perceptions and willingness to use Brave Buttons in a variety of scenarios; and (4) retrospective secondary administrative data analyses of Brave Button installation and response registers (September to December 2021), compiled and maintained by Brave Technology Co-op.

Ethnographers with experience conducting fieldwork in supportive housing settings and communities with lived/living substance use experience (MO, SC, KRK) conducted site visits throughout the SRO Project’s implementation, observing outreach and recruitment activities, overdose response trainings, tenant specialist recruitment and onboarding, and routine SRO Project meetings with housing staff and external project partners. Ethnographers informed building staff, residents, and other constituents of the study’s objectives and their role as observers in the spaces they entered and facilitated informal dialogues with building staff, tenant specialists, and other residents to document the evolution of the SRO Project’s implementation. Ethnographers developed extensive fieldnotes summarizing and interpreting their observations, yielding 100 pages of notes for 35 days of observation.

The research team conducted 30–60-min semi-structured, in-depth interviews with tenant specialists (n = 5) at two stages in the implementation process: (1) during specialist onboarding, after completing their first overdose response training [baseline]; and (2) three to six months after their onboarding as specialists [follow-up]. Baseline interviews explored tenant specialists’ housing histories, overdose intervention experiences, and anticipated constraints/opportunities related to SRO Project implementation. Follow-up interviews examined the evolution of the SRO Project’s implementation, including: emerging challenges and solutions identified, (mis)alignment of specialists’ anticipated and actual roles/responsibilities, recommendations for project optimization, and perceived sustainability of project implementation (Olding et al., 2023). Ethnographers identified and recruited tenant specialists during information sessions about the Brave Button and through staff referrals. SRO Project staff prioritized selection of residents who were already performing overdose response work informally within their buildings (e.g., distributing naloxone, responding to overdoses). Further details of the recruitment process for specialists are described elsewhere (Olding et al., 2023). Longitudinal in-depth interviewing of tenant specialists availed opportunities to characterize the trajectory of Brave Button installation and implementation in real time.

Ethnographers also conducted in-depth interviews with housing staff (n = 6) and project coordinators (n = 2), aided by semi-structured guides tailored to each of these staffing cadres. Conducted at baseline, these interviews explored the residential and overdose contexts within the building where Brave Buttons were later installed. These interviews focused on building and resident characteristics (e.g., services, policies, resident composition), experiences witnessing and responding to overdoses, and expectations for the SRO Project. Semi-structured guides also included questions about perceptions of Brave Button acceptability and adoption among residents, as well as anticipated benefits and constraints to Brave Button implementation and adoption. Staff and coordinators were recruited via email and purposively sampled to achieve representation across a variety of professional roles, including property managers, front desk clerks, case managers, and custodians. Ethnographers also intermittently engaged building staff and coordinators in informal interviews throughout the SRO Project’s implementation, eliciting perspectives and experiences with the SRO Project and Brave Button installation.

Interviewers also administered a questionnaire during baseline interviews with specialists and staff quantifying socio-demographic characteristics (e.g., age, gender, ethnicity/race, time living or working in building) and overdose response experiences (e.g., ever responded to an overdose, actions taken during most recent overdose responses, and confidence using naloxone) among specialists and staff. Due to resource constraints, participants only had the option of completing the survey in English.

Ethnographers conducted audio-recorded interviews face-to-face or virtually (via teleconferencing platform or telephone), which they transcribed verbatim and de-identified prior to analysis. Specialists and staff received a $25 prepaid Visa card as compensation for their participation in each semi-structured interview.

2.3. The research team

A medical anthropologist with over 25 years’ experience leading qualitative and ethnographic research with PWUD and persons experiencing housing instability in California (KRK) led the research project. Other team members involved in data collection and analysis included two ethnographers with expertise in peer-led overdose response interventions (MO) and housing services (SC), as well as a qualitative researcher with expertise in ethnographic study designs and analysis (NJ). Two mixed-methods substance use researchers with expertise in ODT implementation evaluations (JGR, JNP) led analyses for the present manuscript. Beginning in February 2021, the research team met approximately monthly with stakeholders at the DOPE Project (LG), the San Francisco Department of Public Health (EV), and pilot supportive housing sites (LH) to discuss research design, data collection, and interpretation of results.

2.4. Analysis

The research team analyzed qualitative data generated from fieldnotes and in-depth interviews using an inductive, iterative thematic analytic approach, guided by select principles (i.e., constant comparison, open and selective coding, memoing) of constructive grounded theory (Charmaz, 2014). Analysis occurred synchronously with ethnographic fieldwork, with the research team convening bimonthly to discuss emerging themes from observations and interviews. During bimonthly analysis meetings, the research team also developed a qualitative codebook capturing themes identified a priori (e.g., those related to the SRO Project’s specific aims) and other concepts emerging inductively from ethnographic observations and interviews. After finalizing and importing the codebook into Dedoose 9.0 (SocioCultural Research Consultants LLC, Manhattan Beach, CA), two ethnographers (MO, NJ) piloted the codebook to evaluate intercoder consistency, then applied codes to text segments in the remaining fieldnotes and interview transcripts.

The research team then abstracted coded text segments relevant to Brave Button installation and developed a more granular, focused set of thematic codes distilling emerging phenomena related to building and residential contexts preceding technology installation (implementation determinants), Button installation and early implementation experiences (implementation strategies), and perceived/realized impacts of Button installation on tenant safety and building residency culture (implementation mechanisms and implementation outcomes). The first author (JGR) applied these refined codes to pre-coded text segments from fieldnotes and interviews in Dedoose. Recoded text segments were then exported, and salient themes were synthesized and further interrogated through analytic memo-writing and continuous discussion among the research team. To analyze how contextual implementation determinants and strategies influenced acceptability and adoption of Brave Button technologies, the first author mapped emerging themes onto domains and constructs from three implementation science frameworks. First, the research team drew on domains from the Consolidated Framework for Implementation Research (CFIR) (Damschroder et al., 2022, 2009) to describe contextual factors within the PSH setting that enabled or constrained Brave Button implementation. Secondly, the research team used the Expert Recommendations for Implementation Change (ERIC) to characterize and categorize implementation strategies that supported Brave Button implementation (Powell et al., 2015). Third, following Proctor’s taxonomy of implementation outcomes (Proctor et al., 2011), the research team investigated how these contextual determinants and implementation strategies contributed to two key implementation outcomes: acceptability of Brave Buttons (as self-reported by residents and staff in interviews and surveys) and adoption of Brave Buttons (as indicated by installation rates and administrative data on use). (Dis) confirming case analyses (i.e., resolving ambiguities or discrepancies in the data through discussion and interpretation) enhanced confirmability of emerging findings (Patton, 1999). Lastly, across-case comparisons facilitated identification and synthesis of themes emerging distinctly across informant role and profession (Ayres et al., 2003).

To achieve a fully convergent mixed-methods approach (Creswell & Clark, 2011), the research team compared and triangulated insights from ethnographic interviews and observations with descriptive sample statistics from the resident survey and other administrative data summarizing Brave Button installation (i.e., number of Brave Buttons purchased/installed) and activation indicators (i.e., identified reason for Brave Button use) between September and December 2021. Automated data capture embedded into the technology passively recorded activations throughout the implementation period, and trained designated responders (i.e., building staff) documented response-related outcomes through a text-messaging interface developed by Brave Technology Co-op and deployed on responder telephones (Welwean et al., 2023).

2.5. Ethics

The University of California San Francisco Institutional Research Board (IRB) reviewed the research protocol and accompanying data collection instruments—deeming this evaluation to be a quality improvement project exempt from further IRB oversight (#21–33,357).

3. Results

Fig. 2 illustrates an Implementation Research Logic Model (IRLM) (Smith et al., 2020), which visualizes a causal pathway from identified CFIR-guided implementation contexts/determinants to Brave Button implementation outcomes (i.e., acceptability, adoption) vis-à-vis adaptive implementation strategies, framed by the ERIC strategy taxonomy, and proceeding implementation mechanisms. In the following sections, study findings are presented in alignment with a chronologically anchored installation/implementation trajectory of Brave Buttons in the PSH building: (1) pre-implementation contexts/determinants of Brave Button installation and implementation; (2) identification of implementation strategies to enhance acceptability and adoption of Brave Buttons; and (3) perceived/realized outcomes associated with strategy-enhanced implementation of Brave Buttons.

Fig. 2.

Implementation Research Logic Model (IRLM) of Brave Button installation and implementation at a permanent supportive housing building in San Francisco, California.

3.1. The technology is going to be feared”: Pre-implementation contexts and determinants of voluntary Brave Button installation

Specialists and staff identified potential benefits of Brave Button implementation in the SRO, emphasizing the technology’s ability to rapidly respond to a suspected overdose as a key strength. The greatest anticipated benefit of the Brave Buttons, as the tenant specialist below explained, was their ability to connect residents to building staff before they used drugs alone, prompting wellness checks (CFIR: Innovation Attributes):

The biggest advantage is going to be the ability for someone to hit the Brave Button and say, “I’m in my room. I’m about to get high. Could you just check on me in about 10 minutes?”

(Specialist)

Building staff also perceived that Brave Buttons could provide another layer of security to residents during solitary drug use events, creating a safer environment for unmonitored drug use in the privacy of resident rooms (CFIR: Individual Attributes):

A lot of tenants try to act like we don’t know they’re using drugs, so they try to act privately. They’re living a private life like we don’t know what’s going on. I think that [Brave Buttons] would be a good idea. That way, if they didn’t want to reach out anybody, they have that.

(Staff)

Despite the anticipated harm reduction benefits of the Brave Buttons, tenant specialists and staff initially communicated skepticism towards resident acceptability of the technologies, citing overlapping social, relational, and structural constraints to residents’ voluntarily installing and using Brave Buttons (see Fig. 2). Building staff and tenant specialists alike voiced concerns over contextual barriers to effective implementation of the Buttons. Among the most salient of these constraints was perceived resident mistrust of novel technologies with monitoring and surveillance properties (CFIR: Innovation Attributes). Existing mistrust of technologies in resident rooms, as the specialist below explained, could demotivate uptake of Brave Buttons by some residents, who could respond with suspicion to the Brave Buttons unless promoted by trusted peers:

There are so many people here that think that their microwaves are modifying their thought patterns, or the sprinkler is listening to them. But they know me, and I know they know that I’m not a part of whatever this monitoring governmental agency is that they’re worried about. The technology is going to be feared unless they know that someone like myself is the one that comes when they use it.

(Specialist)

Building staff—like the one below—further anticipated that some residents, notably those exhibiting pronounced fears of “gang stalking” (i.e., perceived surveillance by violent groups), would refuse to have Brave Buttons installed in their rooms (CFIR: Outer Setting):

I think that some people will be against it [installing Brave Buttons]. “I don’t want that in my room. What is that? I’m not going to be calling you. I don’t want to. I don’t need you guys to do anything for me.” It’ll be some of that. But I don’t think that it’ll be a lot of that. But then even someone who is suffering from the gang stalking, they take their smoke detectors down because the light is beeping.

(Staff)

Participants, especially tenant specialists, also expressed concerns about the building staff’s capabilities to respond effectively to Brave Button activations as an additional constraint to implementation (CFIR: Individual Attributes). Specialists explained how fragmented relationships between residents and staff, particularly temporary hires with more sporadic contact with residents covering swing (i.e., afternoon hours to midnight) and graveyard (i.e., evening to early morning hours) shifts, would likely demotivate staff from responding to Brave Button activations (CFIR: Inner Setting). As evidenced by the testimonies below, specialists described prior observations and interactions with building staff, which shaped their perceptions of staff capacity and willingness to respond to Brave Button activations:

But the swing shift and the graveyard shift, I think they would make it more difficult for us…because they don’t want to run up and down the stairs. They’re not going to want to answer that [responder] phone regardless of how many times it rings. That’s how they treat us [residents] when we need something.

(Specialist)

Some of the front desk people just prop both the doors open. They just put a stopper underneath the doors and leave them open. I saw this one guy prop both the doors open and then put up a piece of cardboard in front of the screen where he sits, so you couldn’t see him and was kicked back with his feet up and on his laptop watching something with headphones on. He couldn’t hear anything…Is that person going to hear the Brave Button alert? Are they going to care? Are they going to react? Are they going to even have training?

(Specialist)

Tenant specialists further expressed concerns over the increased overdose response burden on specialists and staff alike, particularly anticipated “scope creep” (CFIR: Inner Setting). Tenant specialists in particular expressed concerns about liability stemming from a formalized role in the overdose response apparatus vis-à-vis Brave Button installation. While not expected to serve as a designated responder for Brave Button activations, the tenant specialist below expressed concerns about being present and responding to an overdose during a Brave Button activation, fearing that responsibility would fall disproportionately on specialists:

I wouldn’t mind working with the staff as long as they didn’t put the sole responsibility on me…There are times where I walk my dog…or I have to go somewhere and do something, and I won’t be there for somebody. My concern would be, in working with the staff, all that responsibility would be put on me.

(Specialist)

Additionally, tenant specialists expressed concerns about downstream consequences of staff abandoning their posts when responding to frequent and, in some cases, inadvertent Brave Button activations. Reflecting on previous experiences with trespassing and vandalism, the specialist below expressed concerns about front desk clerks’ capacity to guard the building entrance while responding to repeated Brave Button alerts (CFIR: Inner Setting):

I’m concerned for the people that work here. How many times are they going to be buzzing that thing? That’s my main thing…The desk clerk has to go upstairs and just leaves the front door open for people to bring in guests that don’t live here because there’s no one at the front desk.

(Specialist)

Lastly, participants conveyed safety concerns related to responding to Button activations, particularly in the context of responding to emergencies involving harassment or violence (CFIR: Inner Setting). The specialist below, for example, explained how fears of an intervening in physical interactions could equally render her susceptible to harm, dissuading a potential response to Brave Button activation:

My boyfriend is always telling me, “Do not get into a dispute with a man… Don’t you get yourself hurt trying to step in on someone because you never know, whatever mindset they’re in, if they’re going to come back and hit you.”

(Specialist)

A senior staff member, likewise, conveyed how safeguards and formalized overdose response protocols must be established to keep both residents and staff safe when responding to Brave Button activations, especially in situations where responder safety could be compromised:

It shouldn’t be said to staff, “Put yourself in harm’s way when you go upstairs, and all bets are off around your safety to then provide safety to the residents.” There has to be an even balance… staff should feel safe, like if they don’t feel safe going upstairs because a resident is known to be violent or aggressive.

(Staff)

3.2. “This will help legitimize the work we’re doing”: Adaptive implementation strategies to promote Brave Button acceptability and adoption

Responding to the anticipated Brave Button implementation constraints communicated by tenant specialists and building staff, project staff deployed several strategies to optimize the success of forthcoming Brave Button installations (see Fig. 2). First, project coordinators engaged building management and leadership to develop structurally responsive policies around overdose response, including mandatory staff trainings in overdose response (ERIC strategy: conduct ongoing trainings) and naloxone provision onsite (ERIC strategy: change physical structure and equipment). To cultivate a more enabling environment for Brave Button implementation, one staff member emphasized the importance of securing buy-in and support from decision-makers (e.g., Department of Public Health) to enact responsive policy changes:

We built a good foundation with the Department of Homelessness and Supportive Housing…We wanted to make sure we had their backing… Each housing provider has to have an overdose policy in place and have Narcan onsite. This will help legitimize the work we’re doing…Some of the front desk staff at these hotels don’t want to get trained [in overdose response]. They’re going to have to get trained.

(Staff)

To address perceived resident suspicions and misconceptions of the Brave Buttons, DOPE and San Francisco Department of Public Health project coordinators also convened onsite “engagement sessions” with residents, which were co-facilitated by project coordinators and Brave Technology Co-op representatives (ERIC strategy: conduct educational meetings). As the staff member below describes, these sessions enabled project coordinators to disseminate information about Brave Buttons to residents, communicate the harm reduction philosophy in which the technologies were grounded (legitimized by the recently developed non-punitive overdose response protocols), and empowered residents to voice concerns and ask questions about the Buttons to implementers and manufacturers:

We’re holding these engagement sessions just to give basic information about the Brave Buttons. Luckily, someone from Brave [Technology Co-op] would Zoom in and run the meetings…And then from there, we had a list of people who were interested [in having Brave Buttons installed in their rooms].

(Staff)

Lastly, SRO Project staff purposefully mobilized and trained tenant specialists as ambassadors for the Brave Buttons (ERIC strategy: identify and prepare champions), leveraging their existing relationships with building residents to build trust in the technologies and the institutions installing them. Brave Button promotion by trusted peers, rather than building staff alone, helped to facilitate voluntary installation of the technologies in resident rooms. While building staff promoted the Brave Button more latently (e.g., posting Brave Button signs and flyers at the front desk) (ERIC strategy: distribute educational materials), specialists—like the one below—connected with residents individually to answer questions and allay concerns about the Brave Buttons:

Every time I passed someone in the hallway or was in someone else’s room hanging out, I would just ask everyone if they had heard of the Brave Button and if anyone is thinking about signing up to get one in their room because it could be a real game changer. I was talking to them in language that they are familiar with and receptive to, making the Button seem cool and shining a light on it…instead of it being in a newsletter that they got from the management or someone knocking on their door with a sign-up sheet saying, “Did you want a Button?”

(Specialist)

3.3. “It’s something we’re connected to”: Post-installation acceptability and voluntary adoption of Brave Buttons

Despite initial doubts about resident willingness to install and use Brave Buttons, implementation strategies proved successful in achieving high adoption and self-reported acceptability among residents. Between September and December 2021, Brave Buttons were voluntarily installed in 72 % (n = 34) of resident rooms. A total of 96 Button activations across 29 technologies were documented between September and December 2021. Responses to Button activations were primarily unclassified (73 %, n = 70) or attributed to an accidental activation (22 %, n = 21). However, some residents pressed their Brave Buttons proactively to request a welfare check before drug use (3 %, n = 3) and to alert staff to a suspected overdose (1 %, n = 1) or an unsafe guest in the SRO building (1 %, n = 1).

Tenant specialist and staff interviews reaffirmed the plurality of Brave Button uses, including and beyond overdose response. One specialist recalled pressing the Brave Button before using drugs alone in their room as a precaution, prompting building staff to respond proactively to verify the absence of an emergency. As the specialist below explains, this helped to foster trust in the technology through an empirical demonstration of its use, and adequate response, in a non-emergency situation:

I pressed it three times, and then I did some drugs. About 15 minutes after I pressed it, the front desk person was knocking on my door. I opened the door, and she said, “Everything okay? Are you alright?” And I’m like, “I’m perfect. Thank you so much for your response. Have a good day.” And I know that had I not answered the door, because the Brave Button was used, I think the next step would have been for them to use their master key and do a wellness check.

(Specialist)

Staff and specialists provided some insights into the “unclassified” activations of Brave Buttons, which they said some residents used to instigate a wellness checks on other residents or to alert staff of potential safety concerns (e.g., violence, verbal altercations). Critically, specialists—like the one below—asserted that Brave Buttons enabled residents to summon help from building staff without directly inserting themselves into potentially volatile situations:

Let’s say someone’s walking down the hallway screaming for 18 hours. You can press the Button in the comfort of your room and say, “Go get her.” There are all these different uses that were not drug use. I started to see that if people chose to have it their room, it had a lot of benefits.

(Specialist)

Among the 36 PSH building residents who completed the endline survey (90 % response rate), over two-thirds (67 %) indicated they have, or would like to have, a Brave Button installed in their room (see Fig. 3). Most residents endorsed willingness to use a Brave Button in the event of an overdose in their room (88 %) or their neighbor’s room (92 %), respectively, or for a safety concern (89 %). Residents universally endorsed Brave Buttons as a tool for making the PSH building safer (100 %).

Fig. 3.

Perceptions and willingness to use Brave Buttons among surveyed residents of a permanent supportive single-room occupancy housing building in San Francisco, California (N = 36).

Tenant specialists and staff also conveyed shifts in the relationships between residents and building staff ushered in by the installation of Brave Buttons. Specifically, specialists communicated how the Brave Buttons helped foster a culture of accountability of building staff to tenants. Because the Brave platform requested information from designated responders during alerts, the Brave Buttons actively captured data on staff response time and activations outcomes, which potentially served as extrinsic motivation for building staff to perform their new duties as designated responders. Specialists, like the one below, observed how the consistency and timeliness of staff response to Brave Button activations improved with time, allowing residents to witness the Brave platform functioning as advertised:

The more that we’ve used the system, I see the front desk being more consistent. I had a situation where I pressed the [Brave] Button, and it was on accident one day, and the front desk came up to the room…They sent the janitor…When he knocked on my door and said, “Are you okay?” I just said, “Yeah, I’m fine.” I think I was cutting my hair at the time, so I actually had the noise of the clippers on, so there wasn’t much dialogue… The guy at the front desk, because he didn’t get a visual, sent the guy back up again and had him open my door…And then I saw the guy at the front desk, and he says, “I’m glad you answered your door because I didn’t know whether to call the paramedics.”

(Specialist)

The growing contact between designated responders and residents during Brave Button activations fostered greater connection and trust between residents and building staff. Building staff, including this staff member below, framed their timely responses to Brave Button activations as acts of care, which they believed made residents feel acknowledged and valued:

There was one resident, and we did not get along. She didn’t like me, and I pretty much felt the same. But the Brave Button was activated, and I went up, and I knocked on her door. I was very humble, and I said, “I have activation of the Brave Button. Is everything okay?” And that turned things around. Now she even smiles at me. And she was fine. She didn’t even know she had the Brave Button. It meant something to her that I came up and checked.

(Staff)

A final critical downstream consequence of the installation of Brave Buttons was the formalization of previously informal systems of care among residents. Since residents had already established ad hoc monitoring systems to respond to suspected overdoses and other emergencies in the building, the Brave Buttons engendered a parallel system of care in which staff were now formally tasked with responding to overdoses and other emergencies in the SRO building—shifting some of the onus of overdose response away from residents. Nevertheless, specialists’ roles as Brave Button ambassadors illuminated the hidden labor of residents as overdose educators and responders, as the specialist below described:

It was something we’ve always been doing, but now it feels far more official—like we have an official capacity with this.

(Specialist)

This formalization of overdose response roles, coupled with specialists’ expressed optimism towards the Brave platform, motivated staff and residents to remain involved with Brave Button implementation, even if honoraria and other material resources availed through the SRO Project were discontinued. As the specialist below explained, the perceived novelty and scale-up promise of the Brave Buttons invited continued interest in, as well as sustained engagement with, the technology:

It means more to us than the stipends, and it’s something we’re connected to. We’re not going to stop…because we’re all committed…I’ve always loved the whole Brave platform, and I want to see it implemented everywhere. I just want to, maybe in a year or two, be able to say, “I was on the ground floor of this,” when it’s saving lives and implemented everywhere.

(Specialist)

4. Discussion

Drawing from ethnography and implementation science, this longitudinal, mixed-methods study sought to characterize determinants, strategies, and outcomes associated with the installation of wall-mounted ODTs in a 50-unit permanent supportive SRO building in San Francisco, California. Ultimately, the installation of Brave Buttons in the PSH building helped institutionalize systems of care while legitimizing the informal labor of residents as de facto overdose responders. The mobilization and training of building staff as designated responders to Brave Button activations created a formal mechanism for residents to summon support from building personnel, to which designated responders were accountable to fulfilling. Consequently, the increased frequency of contact between designated responders and residents during Brave Button activations helped foster connection between building staff and tenants. Furthermore, the institutionalization of overdose detection and response shifted the bystander burden away from other residents, who informally filled these roles prior to the installation of Brave Buttons. While scholarship has traditionally emphasized effectiveness outcomes (e.g., number of overdoses detected, deaths averted) for ODTs like the Brave Buttons (Rioux, Enns, et al., 2023; Roth et al., 2021; Schwartz et al., 2020; Viste et al., 2023; Welwean et al., 2023), these technologies can also powerfully reconfigure the social and relational landscapes of the spaces where they are installed, which carry unquantifiable benefits to the people and settings where they are implemented.

Despite low anticipated resident acceptability of the Brave Buttons by building staff prior to their installation, their uptake by SRO building residents emphasizes the potential of implementation strategies to effectively address and modify initial expectations/perceptions of this technology. Interviews revealed the centrality of tenant specialists, specifically, in efforts to increase acceptability and voluntary adoption of Brave Buttons, leveraging their pre-existing rapport and informal mutual aid networks with other residents to cultivate resident understanding and buy-in for ODTs. Beyond the advertisements displayed throughout the building, specialists actively dialogued with other residents about the technology, answering residents’ questions and responding to emerging concerns about the technology. Consistent with best-practice recommendations from implementation of harm reduction approaches (Kennedy et al., 2019; Owczarzak et al., 2020; Scow et al., 2023), recruiting and paying tenant specialists to promote Brave Buttons was critical to building trust in, and affinity with, these novel technologies, which to some residents intersected with the realities of surveillance and policing to which they were subjected as low-income, PWUD with minoritized racial identities (Beletsky et al., 2015; Collins et al., 2019; Urbanik et al., 2022). Early and purposeful integration of peers into installation and implementation plans for ODTs like the Brave Buttons, thus, facilitated their acceptability and adoption in this housing environment—a strategy that could be harnessed to support future ODT implementation efforts in other contexts.

Moreover, residents activated Brave Buttons in a variety of contexts beyond overdose emergencies, reaffirming the built environment’s contributions to the meaning and potential uses of novel ODTs. Administrative data demonstrated that most Button activations occurred in non-overdose-related scenarios. Specialists and staff described a plurality of scenarios in which residents used Brave Buttons, including to alert staff of altercations and other hazards occurring elsewhere in the building. While this empowered residents to notify building staff of potential emergencies from the safety and privacy of their rooms, this application also has potential to become a mechanism for surveilling and policing residents in SRO housing, which can reify pre-existing mistrust of ODTs by end-users (Rioux, Marshall, & Ghosh, 2023; Tsang et al., 2021; van Draanen et al., 2022). Separately, some residents, including interviewed tenant specialists, used the Brave Buttons for their intended purpose of requesting a staff wellness check prior to planned solitary drug use—an application of the technology made possible by the SRO building’s harm reduction-oriented policies, exemplified through the housing operator’s support of overdose response trainings for staff and specialists, efforts to make naloxone readily available onsite, and acceptance of the reality that substance use occurs in resident rooms (Enteen et al., 2010; Olding et al., 2023). Punitive policies towards substance use in other housing contexts may, therefore, constrain the therapeutic potential and promise of these technologies, which require an enabling, non-punitive environment for successful implementation.

Successful ODT implementation in supportive housing, however, also hinged upon a stable, capable workforce with adequate overdose response training and self-efficacy to respond to Brave Button activations, particularly during emergencies. Interviewed tenant specialists echoed concerns related to staff motivation and capacity, especially temporary staff manning the front desk during swing and graveyard shifts, to adequately respond to Brave Button alerts. While institutional reforms mandating universal overdose response trainings for staff helped assuage these early concerns, U.S. housing providers operate in a precarious funding environment, frequently relying on temporary, contracted labor for chronically vacant staff positions (Olding et al., 2022, 2023; Pixley et al., 2022). Universal training and workforce capacity development remain critical precursors to the viability and utility of Brave Buttons in supportive housing, but chronic staffing vacancies can easily undermine efforts to create a more enabling environment for ODT implementation—posing risks to the long-term sustainability of these technologies. Future research should interrogate strategies to sustain the implementation ecosystem for ODTs (e.g., timely response to alerts/activations) in the years proceeding their installation.

The present study fills critical gaps in the ODT scholarship, which to date have exclusively used cross-sectional designs that limit inquiry into the evolution and subsequent impacts of ODT implementation in heterogeneous housing contexts (Bardwell et al., 2021; Ivsins et al., 2022). Study findings, nevertheless, should be considered with several limitations in mind. This analysis of Brave Button implementation was an unfunded endeavor nested within an evaluation of the SRO Project, which limited the scope of data collection. Due to these budget constraints and other challenges presented by the COVID-19 pandemic, the recruitment plan for in-depth interviews did not include the broader sample of building residents, who might not have shared the technological optimism expressed by interviewed tenant specialists. Future research should elicit perspectives of Brave Button implementation from SRO PSH building residents. Second, data collection ceased roughly six months after Brave Buttons were first installed, limiting capacity to document longer-term implementation experiences with ODTs in the building—an important direction of inquiry for future scholarship. Third, despite the longitudinal nature of data collection, ethnographers conducted formal in-depth interviews with only 13 participants—potentially restricting efforts to achieve thematic saturation. Finally, as noted above, the SRO building’s characteristics (e.g., size, location, policies) relative to other housing operators indicate that study findings may not be transferable to other housing contexts. The present study’s focus on a single housing program in San Francisco, however, helped to uncover contextually specific constraints and opportunities for Brave Button implementation.

5. Conclusions

This longitudinal, mixed-methods study of the installation trajectory of fixed-location ODTs, Brave Buttons, in permanent supportive SRO housing in San Francisco identified multiple social, institutional, and structural benefits to technology implementation beyond fatal overdose prevention. Critically, early expectations of low acceptability, uptake, and use of Brave Buttons did not materialize—a testament to the effectiveness of adaptive strategic planning that was responsive to emerging sources of implementation opposition. High installation demands for Brave Buttons among residents, coupled with the frequency and plurality of Button activations, highlight how pre-installation technological reticence may not materialize into implementation resistance, especially when project leadership and staff commit to institutional reforms (e.g., onsite naloxone availability, mandatory overdose response trainings for all staff, formalizing non-punitive overdose response protocols) aligned with harm reduction principles.

List of abbreviations

CFIR

Consolidated Framework for Implementation Research

DOPE

Drug Overdose Prevention and Education

ERIC

Expert Recommendations for Implementing Change

IRLM

Implementation Research Logic Model

ODT

Overdose detection technology

PSH

Permanent supportive housing

PWUD

People who use drugs

SoMA

South of Market

SRO

Single-room occupancy