The first author conducted ethnographic observations during biweekly court review sessions from July 2022 to September 2023, including the transition from video conferences due to the COVID-19 pandemic to in-person sessions. The observations averaged over 1 hour and 40 minutes per session. The lead author observed 23 of the 28 sessions from July 2022 through Sept 2023 (missing 5 sessions for personal or professional reasons). Observations were concluded when data saturation was reached, and no new insights were gleaned from the final 3 observation sessions [37]. AKS recorded jottings and later summarized them in field notes during observations [38]. A review of the field notes was conducted to confirm that no additional novel insights emerged in the final observations.

Observations focused on the technology used in the courtrooms that enabled or created barriers to access. Additionally, special attention was paid to both consumer-facing and provider-facing technologies that impacted each consumer’s mental health care access and retention. This included “social services” and “other services” related to mental health recovery, including housing, substance use treatment, food benefits, vocational support, transportation, and educational services [39].

AKS conducted semistructured interviews with consumers who were receiving care from CMH and with criminal justice histories (n=14) and health and service providers (HSPs) affiliated with the MHTC (n=14). Prior to conducting the first interview, AKS piloted the interview with a colleague to clarify language and estimate the timing of the interview, with consumer interviews taking 90 minutes and HSP interviews taking 60 minutes. Interviews were conducted using an interview guide. All providers were current or former employees or volunteers who worked with consumers in court. HSP interviews occurred between January 2022 and November 2023, and consumer interviews between October 2023 and May 2024. For all interviews, no one else was present besides the participants and AKS. AKS transcribed the interviews verbatim or summarized the interviews (upon participant request). Interviews for both consumers and providers continued until after the data were saturated and there were no new emergent themes [37]. Subsequently, the final 3 transcripts from both samples (consumers and HSPs) were reviewed to confirm that there were no new emergent findings and that no new inductive themes emerged from the data. No transcripts were returned to participants for comment or correction.

Providers and consumers completed a demographic survey after the interviews, with the option to answer survey questions over the phone. Both groups were asked questions about sex, age, race/ethnicity, and the highest level of education. Consumer surveys also asked about patient health insurance and years of technology use.

All qualitative coding was conducted using NVivo, and iterative review of the transcripts began when data collection started. The coding mixed deductive and inductive approaches. Deductively, AKS and TCV developed codes based on the Levesque et al [16] model, in which each process-oriented concept in the model became a process code. A round of coding was conducted to group various actions and processes across the Levesque framework, specifically coding interviews and transcripts into the following groups: (1) perception or identification of health care needs, (2) health care seeking, (3) health care reaching, (4) health care usage, and (5) health care consequences. While coding the access stages, AKS identified that “staying in care” was an additional stage not accounted for and subsequently added that code, reviewing all transcripts for this additional inductive code. An associated coding matrix can be found in Multimedia Appendix 1 [16]. Furthermore, inductive coding involved AKS and TCV, both researchers with mixed methods health informatics research experience, independently reading 3 randomly selected consumer interviews, 3 randomly selected provider interviews, and 5 randomly selected field notes to develop emergent codes via process coding [40]. To gain a nuanced understanding, AKS triangulated data across the sets of interviews and ethnographic field notes, applying the same codes to all data sources to enable comparison across sources. AKS subsequently developed analytical memos for individual codes to synthesize emergent results and map the inductive codes to the access model.

AKS calculated descriptive statistics for demographic variables, including frequencies, means, and SDs.

While the Levesque et al [16] model described the “Perceptions of needs and desire for care,” all study participants had already been diagnosed with SMI and were largely using some form of mental health care. Thus, a more apt description of this stage for participants is “Identifying care needs.” Additionally, unlike the Levesque et al [16] model, which categorizes identifying care needs as one discrete stage, there were two need identification stages for study participants. The first stage involved the initial identification of care needs, and the second stage involved the ongoing identification of care needs (Figure 1).

Within this sample, the first stage of initial identification of care needs often occurred decades prior, typically in young adulthood. As one consumer (C012) stated, “I’ve been getting help with mental health since I was younger…I didn’t do that on my own.” Others initially identified needs because of an acute mental health incident in adulthood. For example, C010 stated that “up until the time I cracked, I never needed any [mental health care].”

Some consumers pursued ongoing identification of care needs to become aware of any new or increased needs (Figure 1). Hence, monitoring symptoms with technology was a key part of this stage. For instance, 2 consumers used smartphone mobile apps and wearables in ways that identified symptom changes. As discussed below, technology-mediated communication with friends and family also served this purpose.

For the “Health care seeking” stage of the Levesque et al [16] model, care seeking is depicted for only one care setting at once. However, low-SES consumers with SMI frequently sought several different sources of care—some of them related to recovery (eg, housing) but not necessarily providing treatment—often simultaneously. Hence, in Figure 1, we have changed this category to “Seeking Care.” Specifically, for the population under study, this involved therapy and psychiatric services, as well as “other services” and “social services.” Case managers also helped people connect with “other services” based on each consumer’s needs, including substance use treatment, food access, educational services, vocational support, group therapies, and assisted living. This could involve legal services such as probation officers or MHTC staff, depending on the situation. For example, one MHTC participant describes:

The “social services” that consumers with SMI in this sample regularly sought included: (1) housing resources (ie, Section 8 vouchers or low-income housing), (2) health insurance if needed (eg, Medicaid/Medicare), (3) transportation supports (eg, free bus passes, medical transportation services), (4) assistance with job seeking (vocational support), and (5) food benefits. Typically, “other services” and “social services” were offered by CMH partner organizations. After being connected with these services, consumers could navigate them with varied levels of support. C008 uses CMH for therapy, psychiatry, and case management. Beyond that, he describes some of the different services that CMH helped coordinate, like “$300 towards fixing my car,” securing a “housing voucher,” and obtaining a federally subsidized “food card.”

Another departure from the Levesque et al [16] model was an initial seeking care step: case management. Most consumers experienced this step when they enrolled as their local CMH. When this happened, they automatically received a case manager to facilitate their further care-seeking and linkage. As Figure 1 shows, case management initially facilitated access to psychotherapy, psychiatric services, and necessary “other services” and “social services” that support mental health recovery. CMH typically provides psychotherapy and psychiatric services directly. However, consumers sometimes sought therapists outside the CMH-provided options due to provider shortages, related wait times, and schedule limitations at CMH.

An extension of the technology-mediated care context is locating and acquiring resources with technology. As mentioned, some consumers searched for potential therapists via web-based databases or web searches. Additionally, consumers made inquiries or completed web-based forms on their devices. One consumer described completing a web-based form for multiple resources:

An additional necessary modification of the seeking care stage is the addition of emergency service acquisition, as demonstrated in the revised model (Figure 1); emergency care-seeking circumvented several care access stages when required. This occurred when consumers arrived at psychiatric emergency services of their own volition (calling 911, going to the emergency room [ER]) or due to others in the community calling 911 if they believed that the consumer was a threat to themselves or others. Calling 911 due to a psychiatric emergency may have occurred at either the initial or ongoing needs identification stage. In recognition of the use of technology to acquire emergency services, the category of technology-mediated contact was added. Notably, through this emergency service acquisition, it is also possible that consumers will have contact with the criminal justice system. For example, C003 describes witnessing this experience by saying,

Calling the police works to keep CMH consumers safe while in crisis, but it often compensates for a lack of resources for low-SES CMH consumers.

Depending on the situation, this step may also result in legal trouble that could lead to participation in the MHTC diversion court.

Starting at the care-seeking stage and continuing through reaching care, a new cross-stage category is needed: “playing the waiting game” (Figure 1). This was a common experience of low-SES consumers with SMI, who often faced prolonged waits for appointments, new provider assignments, housing acquisition, and medical transport. Waiting typically occurred with little guidance or communication of updates from HSPs. For example, a fieldnote from observations at court in April 2023 described the following scenario:

Notably, there was typically a lack of technological systems through which consumers could follow the status of applications for housing or other services. Instead, a consumer or case manager sought any updates via email or phone.

For the health care reaching stage of Levesque et al [16], we again see low-SES consumers or the HSPs who work with them contacting multiple forms of care, often simultaneously. Furthermore, it is critical to note that the care that was reached was delivered either in-person or in a technology-mediated manner (ie, logging into a computer for a telehealth appointment). Hence, the name given to this stage of the model was altered to say, “Reaching in-person and technology-mediated care.”

Although some care was delivered in person, coordinating transportation with technology for in-person visits was often necessary. Some consumers used transportation apps on a smartphone or booked nonemergency medical transportation. For instance, one consumer (C006) used a bus app to arrive at their appointments on time, stating, “I have maps on my phone, so I use that a lot to tell me when the bus is coming.”

Preparing for technology-mediated care was necessary here as well. This involved consumers acquiring and maintaining technologies, often with the help of HSPs or clubhouse peers. It also consisted of following various procedures and using different technologies driven by those used by the HSP in question. Consumers constantly renegotiate reaching care, partly due to technology-related challenges related to having technology and adapting to various HSP systems. For example, even if a consumer knew how to use a platform (like Zoom), variable settings across providers created complexity.

C005 also argued that HSPs make “too many assumptions that everyone should be using computers or have the knowledge or capabilities to access everything… one assumption that is that everyone must have patience and endurance to contend with computer glitches, free ups, correction to typos, tech compatibility, and trying to figure out how to use the different apps and how to use the different things accordingly.”

Playing the waiting game also persisted into the health care reaching stage because consumers who lacked personal transportation had to plan their medical transportation far in advance. They also spent much time waiting for transportation providers to arrive. Another participant described how she used her phone to establish medical transportation:

On the day of the appointment, the transportation service provider often sends a digital notification or reminder:

The Levesque et al [16] model maintains a singular box to represent “health care usage,” with primary and secondary access subcategories. However, this distinction does not entirely fit SMI care for low-SES patients. Rather, here again, there is a need to maintain recognition of the use of multiple services simultaneously: therapy, psychiatric services, other services, and social services (Figure 1). These services were often provided via technology-mediated delivery, such as therapy appointments or court review sessions via telehealth. One benefit of this approach was that it reduced consumers’ travel times by using telehealth calls, making it possible to fit more into their schedules, allowing for increased schedule flexibility. For example, service providers (eg, HSP03) identified such a benefit for MHTC participants (“other services” box in Figure 1), stating that Zoom is “easier for the participants as far as transportation and the ability to get to the court.”

At this stage, there is also a need to recognize the continued pathway of emergency service acquisition. Such direct service occurs via the ER or inpatient care if the hospital admits a consumer. This typically resulted in hospitals discharging patients into CMH care. However, there were situations in which consumers could be court-ordered to adhere to treatment recommendations, often when they posed a risk to themselves or others or if their judgment was deemed to be impaired by their SMI. C003 described her experience with this:

This pathway primarily stabilized consumers so they could again engage in care through usual routes or make care engagement decisions with sound judgment.

A “staying in care” stage was added to the original framework to accommodate SMI care’s complex access cycle. The importance of this category of access was evident in the experiences of the study participants. Most study consumers received psychiatric services 2‐24 times a year and received therapy as often as every week. The presence of staying in care is an explicit step in the framework to emphasize that consumers benefit from sustained service use and that retention helps produce favorable mental health outcomes from participants’ perspectives. In Figure 1, arrows to the right of the use category indicate that consumers must keep using these services regularly, thus staying in care.

At this stage, managing time with technology is a key adaptation for a technology-mediated care context. This involved accomplishing time-sensitive activities such as scheduling, rescheduling, and remembering health care appointments using various technologies. For example, as discussed subsequently, many consumers used calendaring apps to keep track of their upcoming health care appointments and stay in care. Health care organizations also helped consumers stay in care through reminders in the form of automated phone calls, live calls, text messages, emails, or appointment cards.

As for the health care consequences stage, the Levesque et al [16] model outlines the economic, satisfaction, and health-related consequences of health care access. While each of these was evident in study results, data concerning the experiences of low-SES consumers suggests the need to modify the framework to additionally reflect the psychosocial and emotional consequences as part of health consequences. Indeed, there were negative consequences in each category above for those who could not figure out how to access technology-mediated care. As we elaborate below, consequences included experiences of frustration, anger, and lowered self-efficacy, suggesting extensions to the existing model. For instance, C004 described how she felt when she could not get her telehealth video call working; she said, “Frustrated. We expect technology to work a hundred percent of the time.”

Consumer technology use falls into 3 emergent categories that recur across multiple stages of the access framework: keeping track, figuring it out, and leaning on others. In this study, keeping track entailed consumers monitoring their symptoms, mental health status, medications, schedules, and appointments. This allowed them to make necessary adjustments and clearly understand their present and future care needs. Figuring it out refers to the continuous, individualized efforts of low-SES consumers with SMI to comprehend and navigate their care across multiple services and amid complex and often unclear processes. This involved performing tasks to prevent care disruptions and adapting to new tools used for accessing care. As C006 noted, “Technology in some ways has been good, but it has also been very trying to figure it out.” Leaning on others refers to the external support consumers seek or receive from friends, family, and peers with SMI or HSPs to facilitate access to technologically mediated health care. Consumers leaned on others routinely, such as regularly discussing new symptoms, or on an ad hoc basis, like when troubleshooting spotty Wi-Fi. Having introduced these categories, their presence in the context of each Levesque et al [16] model stage is now outlined.

As mentioned, most low-SES consumers with SMI were primarily engaged in the ongoing identification of care needs rather than in recognizing initial symptoms. At this stage, monitoring symptoms with technology was performed as part of consumers’ larger practices of keeping track. Here, consumers were often leaning on others, particularly loved ones or peers, by digitally communicating their SMI symptoms, using tools to identify new or changing symptoms collaboratively. Consumers usually call or text friends and family to “check in” on their mental health. For some, technology-mediated “check-ins” with a loved one increased symptom awareness and allowed consumers to verify their own mental health self-assessments and determine the necessary next steps.

Regular technology-mediated communication with friends also supported symptom monitoring, allowing individuals to lean on each other for awareness of mental health episodes.

These mechanisms for technology-mediated communication supported consumers’ needs identification while providing them with social support.

Two consumers also used smartphone mobile apps and wearables in ways that identify symptom changes. Specifically, C009 used mood-tracking features on the MyStrength mobile app designed to self-monitor her symptoms. C008 used a Fitbit wearable and its mobile app for biometric feedback, alerting her to changes in her stress and anxiety: “This Fitbit [that’s] hooked up to my phone… it’ll beep and say, ‘okay, your stress level is at blah, blah, blah’,…it’ll buzz me.” For C008, this app helped her identify her needs, adding that the app helped to make her “more aware.”

Consumers were sometimes proactive in seeking care by figuring out what services were available and how to access them with technological tools. We called this access process “locating and acquiring resources with technology” in the modified framework above. Figuring it out for some consumers involved searching for potential therapists via web-based databases, web search engines, or perusing service-related websites to locate resources. Web-based applications were typically necessary to acquire resources. Some consumers also independently contributed to figuring out how to apply for social services while using technology, such as housing or food stamps, and then did so. Nevertheless, many consumers faced obstacles when figuring out how to access services alone. For instance, issues with insurance compatibility and navigating these websites frequently complicate these processes. Such challenges led consumers towards leaning on others while seeking services, especially HSPs, who guided consumers through the technology-mediated process of applying for services. Importantly, case managers provided a second set of eyes on important paperwork before a consumer submitted it, as one consumer stated:

Leaning on others when using technology to seek health care also included others submitting documents related to consumers’ mental health needs. For example, C013 used resources at a mental health clubhouse to fill out and submit a housing application, stating, “The clubhouse staff, they would help me fax applications to different apartments or email them to print them off.” Similarly, C009 stated, “I filled out my paperwork and I think [CLUBHOUSE STAFF] emailed it to the landlord.” By leaning on case managers and clubhouse staff to review and send documents via technology, consumers could more confidently navigate the complex process of seeking social services in technology-mediated ways.

Sometimes, HSPs also offered in-person social support to those who struggled to navigate technology-mediated applications for social services. In turn, this assisted with understanding and emotional regulation.

C005 described how in-person case manager support prevented her from becoming too emotionally overwhelmed at the computer: “In-person, if they say, 'Okay, can I save this? Can we break it down to doing it probably 20 minutes at a time.’” By transitioning from an independent digital submission to a supportive in-person approach, consumers found it easier to manage the lengthy and complex form submission processes in the care-seeking stage.

As mentioned, technology was involved in 2 parts of the reaching care stage: coordinating transportation with technology for in-person visits and preparing for technologically mediated care. For the former, consumers set up their medical transportation through mobile apps or phone calls before their appointment. For example, C006 described reaching in-person services, stating, “I’ve got my [TRANSIT] app and… you schedule rides there.” For some, this coordination included medical transportation services (such as local ride-sharing services for people with disabilities) and ride-hailing services (such as Uber or Lyft).

Figuring out how to coordinate transportation meant that participants had to use various apps or platforms to get to in-person care. Consumers who lived near accessible public transportation often used apps on their cell phones to plan their travel and physically navigate their way to those appointments. For example, one consumer used a bus app to arrive at her appointments on time and also described how she used a website provided by her local transit:

As for preparing for technologically mediated care, consumers spent time figuring it out in relation to having the necessary technology for visits. Indeed, consumers often faced financial challenges that restricted their ability to use telehealth as they balanced their health care needs with other expenses. To address this challenge, leaning on others was common. HSPs like case managers and clubhouse staff played a crucial role, providing the technology consumers needed to connect to technology-mediated care. They ensured that consumers had access to the right digital tools and often had private spaces appropriate for conducting telehealth calls.

Preparing for technology-mediated care also meant acquiring technology skills, another place in which consumers reported leaning on others. For instance, one consumer who had a peer support role at a Clubhouse also described her role in teaching her peers how to use novel technologies:

Furthermore, as mentioned, it was common for each consumer to have multiple providers with whom they interacted regularly. Often, the consumer had to figure it out in relation to interacting differently with each provider, either via different technologies or in person. Due to variable mechanisms for engaging with other HSPs, it was difficult for consumers to know what was appropriate and what was not, necessitating figuring it out. As one consumer described, mechanisms for accessing health information were not streamlined. This puts consumers in a position to try multiple digital methods (eg, calling and portal use) to change or reschedule an appointment. C006 explained how she would get a new appointment scheduled, expressing frustration with the process:

Leaning on others also taught consumers how to use technology-mediated communication to interact with service providers. A peer support specialist and consumer (C004) stated, “I can show people how to email their case manager or get on the phone and teach them proper etiquette.” This effort needed to figure it out, which added complexity to reaching care.

Health care typically involves some technology-mediated care delivery. For example, C012 described the range of services he used:

Notably, for this consumer and others, the provider typically decided whether an appointment would be in-person or technologically mediated, often without much input from the consumer. When it came time for telehealth visits, consumers reported leaning on others, specifically clubhouse staff and members, to make visits work. In clubhouses, there was an environment in which learning to use or access technologies for care was a collaborative effort. One consumer shared her experience, stating:

At times, clubhouse members also walked other members through using a novel device to connect with care.

Keeping track of medications and appointments was a goal of some SMI consumers, and it was especially important for consumers involved in the MHTC since therapy was part of probation conditions. However, this was difficult for most consumers. Regardless of MHTC involvement, there was a lack of sufficient consumer-facing technologies. Indeed, most consumers in this study did not have patient portals for their CMH care. This stymied consumers. As C004 mentioned, “I use the patient portal for [HEALTHCARE SYSTEM], so I should probably use it for CMH too.” Limited or no access to their records made it burdensome for consumers to obtain information about diagnoses, medications, and appointments. Hence, consumers were trying to figure out how to manage their medications without efficient technological systems for medication management and how to keep track of appointments without such technologies. Consumers saw a need for improved technological systems that would enable them to track and manage their care easily. C002 described the potential benefits of having an integrated portal:

As a result of inadequate access to personal health information, consumers often had to lean on others. Specifically, they relied on case managers as intermediaries to keep track of their appointments and medications. C002 explained how she would obtain her medication list:

A major part of staying in care is managing time effectively to remember and attend appointments. Service providers highlighted that some consumers face significant adjustments when beginning or reinitiating care. An MHTC HSP (HSP08) noted that consumers struggled to establish “an understanding of all that is expected to them and time management and being able to comply with [APPOINTMENTS].” Additionally, HSP04 stated:

Thus, some consumers had to figure out how to manage time with technology to stay in care.

From those figuring it out to those who were more accustomed, consumers in the sample used calendaring apps to keep track of their upcoming health care appointments to stay in care. One consumer (C004) said, “I use the calendar app a lot because if I don’t put it in there and get a reminder, I’ll forget about.” Knowing the different upcoming commitments ensured that they could attend and prepare themselves for these appointments. To keep track of her medications, one consumer (C008) also used a “reminder list on my phone that says, ‘okay, take this medication with you today.’”

While digital calendars worked for some, other consumers used physical calendars to keep track of appointments and other commitments. Still, others used both digital and physical calendars to reinforce their memory by recording upcoming appointments in multiple places. When asked how she used her paper calendar, one consumer stated

Consumers with challenges attending appointments also used physical calendars to keep track, as indicated by this MHTC fieldnote:

This consumer was reticent to adopt a digital calendar but ended up incorporating both versions for keeping track.

Consumers reported leaning on others through digital and paper reminders provided by health care organizations and the court to help them stay in care. For instance, health care organizations help consumers stay in care through reminders in the form of automated phone calls, live calls, text messages, or emails. This resulted in consumers keeping track of appointments more easily. One consumer described her experience with phone call reminders:

This kind of reminder from CMH helped jog consumers’ memory, alerting them that they needed to attend an appointment—often the next day. Another consumer noted they are less likely to miss care when they receive reminders:

In other settings, including the MHTC, consumers were also given a paper card or sheet indicating their upcoming appointments to help them keep track of appointments. This is analogous to paper reminder cards that consumers receive in person after health care appointments:

One service provider also described how case managers often gave such physical reminders:

Consumers who could not figure it out with respect to managing time with technology risked missing care-related appointments, even if an MHTC mandated them. An example of this in the legal context can be seen in MHTC fieldnotes:

The original Levesque model highlights the economic consequences of care, which, for our purposes, encompasses the time and travel costs related to health care visits. For technology-mediated care, consumers highlighted positive impacts on time costs. C004 explained:

Reduced transportation challenges were also evident from consumers’ perspectives. C014 said that technology-mediated connections remove additional time needed to get to care: “I don’t even have to think about traveling or taking the bus, it’s easier to do it virtually.” HSPs identified a similar benefit for MHTC participants, stating that Zoom is:

Satisfaction is also a health care consequence in the Levesque et al [16] model. Most consumers advocated for maintaining hybrid care options to increase access, recognizing both the benefits and trade-offs of technologically mediated care and improving overall satisfaction. Here, it was recognized that technology-mediated care could make people more comfortable as an aspect of satisfaction. For example, HSP05 mentioned that when conducting court sessions over Zoom, consumers are “more comfortable. It’s not as intimidating.” HSP02 also reflected this sentiment:

However, less-satisfied consumers and service providers raised concerns about negative impacts on social connectedness, with C003 stating, “It just kind of makes me feel lonely seeing people in a computer.”

The Levesque et al [16] model also highlights the health consequences of care, including emotional and psychosocial health, for this study. Here, the impacts of technology-mediated care could be negative. Specifically, failing to figure out how to facilitate technology-mediated care delivery caused emotional distress, as one peer described:

This may lead to frustration and anger, as a consumer noted: