Pre-implementation determinants for digital mental health integration in Chicago pediatric primary care Open Access

Abstract

Objective

Pediatric primary care (PPC) is a common treatment site for pediatric mental health, but it is currently unable to meet the needs of all teen patients, particularly those with minoritized identities and/or marginalized experiences. Digital mental health (DMH) low-intensity treatments (LITs) can increase mental health screening and care capacity in PPC, but how this is done successfully without burdening providers, patients, or families is unclear. This paper presents a pre-implementation study aimed at understanding the implementation context (PPCs in Chicago, IL) for a specific DMH LIT.

Method

Using a mixed-methods design, quantitative data from an online survey of providers assessed current DMH practices in PPC, and qualitative interviews with Pediatricians and Pediatric Psychologists examined implementation determinants for a specific DMH LIT. Quantitative data were analyzed using descriptive statistics, and interviews were analyzed using rapid qualitative assessment.

Results

Survey reports (n = 105) and interviews (n = 6) indicated low current use of DMH. Providers in PPC clinics voiced multiple reasons for low usage and low perceived feasibility, including: Consolidated Framework for Implementation Research (CFIR) Inner Setting Domain (PPC clinic workflow, responsibility and ethical considerations, patient privacy and confidentiality), CFIR Outer Setting Domain (hospital and healthcare system factors), CFIR Innovation Domain (DMH design), and a cross-cutting theme of safety.

Conclusions

Provider-reported low feasibility for integrating DMH in PPC is a call to action to partner with interdisciplinary colleagues and identify how such settings can ethically and seamlessly deliver digital evidence-based and accessible screening and care prior to implementation.

Teen mental health needs have amplified as a result of multiple and overlapping endemics (Murthy, 2022). Meeting these needs is challenging, as most youth who qualify for a mental health diagnosis do not receive treatment (Olfson et al., 2024), and this treatment gap is even more marked for minoritized youth (Chang & Slopen, 2024). These unmet needs are in part driven by limited access and provider shortages. As such, identifying alternative approaches to address adolescent mental health needs is essential. One possibility is leveraging pediatric primary care (PPC) visits, which nearly all children in the United States (U.S.) receive annually (National Center for Health Statistics, 2024). PPC is a low-stigma setting, and increasing numbers of caregivers and teens are receptive to regular mental health screenings and referrals during PPC visits (Kass et al., 2023; Stiles-Shields et al., 2024).

Yet, pediatricians report multiple and growing barriers to assessing mental health needs and connecting teen patients to care (O’Brien et al., 2016). Further, teens with intersecting minoritized racial, ethnic, and gender identities are confronted with disparities in PPC settings, with reduced detection of mental health disorders and lower behavioral healthcare quality and service use, both in-person and via telehealth (Chakawa et al., 2021; Prichett et al., 2023; Slopen et al., 2024). Despite these barriers, PPC is often a frequent treatment site for pediatric mental health (Duong et al., 2021; Shahidullah et al., 2023), because patients bring their concerns (physical and emotional) to their pediatricians. Hence, for PPC to better serve the mental health needs of minoritized youth who may have difficulty accessing safe services elsewhere (Shahidullah et al., 2023), it is necessary to understand current practices and contextual factors impacting the implementation of mental health supports in PPC.

Digital approaches are increasingly considered viable to improve capacity for comprehensive care for teens in PPC settings. Indeed, integrated digital mental health (DMH) care has been proposed as a strategy to integrate mental health services within primary care settings (Lim et al., 2024). DMH tools and interventions have demonstrated efficacy with pediatric populations (e.g., Fedele et al., 2017), and teen patients express a preference for DMH screening in healthcare settings (Martel et al., 2021). Multiple barriers prohibit recommended universal depression and anxiety screenings for teens in PPC (e.g., time, discomfort with topic; Acri et al., 2018; Bose et al., 2021; Kenny et al., 2021), and DMH may be particularly poised to bridge this gap as a low-intensity treatment (LIT). LITs are patient-facing tools with demonstrated efficacy to provide brief screening, psychoeducation, and/or resources—serving as a promising early step for care environments, such as PPC, experiencing difficulty managing the screening and patient care volumes (Lorenzo-Luaces & Fite, 2024).

Unfortunately, several factors are implicated in the research-to-practice gap for DMH LITs in healthcare settings. Pediatric DMH has often overlooked implementation planning at the start of design (Psihogios et al., 2022), and has therefore not included the perspectives of key players (teens, caregivers, providers) into how a digital tool could fit into their lived experience. In healthcare environments, DMH LITs differ from traditional services and workflows—leading to poor uptake, engagement, and, often, abandonment of the DMH (Graham et al., 2020). Further, because most DMH research has focused on adult healthcare settings, and time periods and workflows existing prior to 2020, little is known about current DMH implementation in PPC.

The EPIS Framework and the Stages of implementation completion (SIC) both divide implementation into phases: pre-implementation (exploration, preparation), implementation, and sustainment (Aarons et al., 2011; Chamberlain et al., 2011). Drawing from these frameworks, pre-implementation activities provide a critical foundation for successful integration of a new practice or intervention into new contexts (i.e., uptake, acceptability, sustainment of DMH into PPC; Alley et al., 2023). During pre-implementation, there is an assessment of engagement, feasibility considerations, and readiness planning (Chamberlain et al., 2011). These activities are particularly important to promote healthcare equity through implementation science and establishing partnerships with community and provider colleagues in their spaces (i.e., ensure an intervention has partner input based on their needs, lived experience and/or values; Baumann et al., 2023). The Consolidated Framework for Implementation Research (CFIR; Damschroder et al., 2009, 2022) provides a model for understanding the multilevel factors (i.e., individual, clinic, organizational, and intervention) that impact the implementation of evidence-based practices in everyday settings—and has been commonly applied in pre-implementation activities for adolescent healthcare settings (Zolfaghari et al., 2022). Pre-implementation assessment and planning informed through frameworks like the CFIR better ensures that DMH integration into PPC is contextually relevant, feasible, and responsive to needs of the local community.

Pre-implementation for a digital mental health low-intensity treatment in pediatric primary care

The current study sought to understand the implementation context of Chicagoland PPCs, and support the integration of a specific DMH LIT, the Teen Assess, Check, and Heal (TeACH) System, aimed at serving the mental health needs of communities that are systemically excluded from accessible mental health resources and support (i.e., the West Side of Chicago, IL; Lynch et al., 2020). Shaped by input from teens and caregivers, the TeACH System provides: (1) brief anxiety screening (via the Kiddie-Computerized Adaptive Tests [K-CAT] anxiety module; Gibbons et al., 2019); (2) brief feedback for teens scoring with mild or greater anxiety (e.g., anxiety prevalence, psychoeducation [e.g., “Teens from your neighborhoods have described this experience as ranging from ‘Just feelings of worry or uneasiness. Just worrying about the future.’ to ‘Anxiety is the feeling of being constantly worried, stressed, feeling anxious about everything every day in life, about little stuff as well.’”]); and (3) freely accessible mental health resources teens may access on their own (e.g., apps, TikTok reels/YouTube videos grounded in evidence-based strategies). The TeACH System is one DMH tool among LITs that are growing in prevalence but have not been integrated into routine care settings (Lorenzo-Luaces & Fite, 2024). As such, findings from the pre-implementation activities for the TeACH System may inform integration of DMH LITs into PPCs, including in urban environments that have built systemic disparities to impact specific communities and residential areas (Lynch et al., 2020).

The current study aimed to address two primary pre-implementation questions. First, what are the current needs and practices of clinics serving teens across the Chicagoland area? To address this, a necessary first step was to characterize current mental health screening and referral practices, as well as the current DMH implementation landscape in Chicagoland PPC and medical clinics serving teens (e.g., types of DMH tools currently leveraged, the extent of their usage, and providers’ perspectives on the feasibility of these models). Second, what are the pre-implementation determinants relevant to the integration of a DMH LIT (i.e., the TeACH System) into specific PPCs serving teens who experience multiple and overlapping minoritized identities and/or experiences? To this end, CFIR was used to elucidate relevant pre-implementation determinants for the TeACH System and similar DMH LIT, broadly. Because previous studies of DMH determinants have been primarily focused on (1) adult healthcare settings, and (2) time periods and workflows existing prior to the onset of the COVID-19 pandemic, current social justice movements, and declaration of a youth mental health crisis (Graham et al., 2020; Murthy, 2021, 2022), the present study was exploratory in nature. The study findings were intended as a first step to guide future implementation planning for the TeACH System, and serve as an example to inform broader DMH LIT-PPC integration, with the ultimate goal of increasing access to mental healthcare resources for PPC patients.

Methods

Participants

Providers participated in an online survey characterizing mental health, screening/referral practices, and DMH landscape in PPCs, or key informant interviews to provide perspectives on both the TeACH System and similar DMH LIT. Survey providers were recruited from emails circulated through Departmental listservs and the Illinois Chapter of the American Academy of Pediatrics. Consistent with pre-implementation activities aimed at understanding the local implementation context, eligibility criteria included treating adolescent patients in a medical clinic located in the Chicagoland area. Interview providers were recruited via email invitations based on recommendations to reach out to key opinion leaders in their settings (i.e., two PPC clinics housed in academic medical centers on the West Side of Chicago; Malterud et al., 2016). These locations were selected as the initial integration sites for the TeACH System, and as part of a larger body of research focusing on adapting DMH LITs for and with teens living in communities disproportionately burdened by health disparities, violence exposure, and higher economic hardship (e.g., Stiles-Shields et al., 2023).

Procedure

All procedures were approved by Rush University Medical Center and the University of Illinois Chicago Institutional Review Boards (IRBs). Survey providers received an electronic link to a study consent form. After giving digital consent, survey providers answered screening items to ensure eligibility and then the survey items. To promote engagement while maintaining anonymity, the final page of the survey provided a link that enabled providers to provide contact information. The first 30 respondents who provided their information received a $10 egift code.

Following consent, interview providers were scheduled for individual interviews conducted via a HIPAA-compliant zoom account. One-hour interviews were conducted by the lead author and were audio-recorded. Interview providers received a $10 online gift code for their time.

Measures

Provider survey: landscape for mental health screening, referral, and DMH practices

Delivered via REDCap (Harris et al., 2009), survey items were organized around three domains. First, clinic-level factors included: clinic type (e.g., PPC, school-based health center), setting (e.g., medical center), zip code, primary languages in clinic, use of the interpreter line, populations served, use of universal vs. targeted validated screeners, administration processes for screeners, use of DMH in clinic (i.e., use for screening or referral in past year, preferred workflows for screening DMH in clinic, Wi-Fi access in clinic, tablet presence in clinic) and individual-level characteristics (i.e., discipline). Clinic zip code data were used as a proxy for community factors, including: (1) Distressed Communities Index (DCI); and (2) percentage of neighborhood households with Broadband Internet Access. While clinics typically have larger catchment areas than their immediate zip code, this variable was selected to provide a characterization of the clinic’s community setting while minimizing participant burden. The DCI is a tool derived from the U.S. Census Bureau data to create a ranking by zip code to account for seven metrics (i.e., unemployment, education level, poverty rate, median income, business establishments, job growth, housing vacancies). Total scores range from no distress (0) to severe distress (100), with quintiles of well-being (prosperous, comfortable, mid-tier, at risk, distressed; Kesler, 2022). Percentage of neighborhood households with Broadband Internet Access was calculated using the interactive map created and maintained by the University of Chicago’s Data Science Institute’s Internet Equity Initiative (Map—DSI Internet Equity Initiative, n.d.)

Next, teens’ mental and behavioral health factors included: top concerns and reasons for referral, and how often teens requested confidentiality. Finally, provider-level factors included their perceived feasibility of integrating DMH in clinic (via the Provider version of the Mental Health Implementation Science Tools [mhIST]; Aldridge et al., 2022). The mhIST assesses mental health interventions by implementation domains outlined by Proctor et al. (2011). Providers were prompted to “Please consider the use of Digital Mental Health (DMH) to provide brief mental health screenings and referrals to online tools for mental health symptoms for teen patients in medical settings.” The mhIST was used to assess the Appropriateness (Self-Perception of Effectiveness items; e.g., “Is DMH a good way to address your clients’ problems?”), Feasibility (Skills, Time, Resources items; e.g., “Do you have enough time to regularly provide DMH to those who need it?”), and Accessibility/Reach (e.g., “Would parents or other caretakers seek DMH services for their children if needed?”). The mhIST has demonstrated validity and reliability for providers in international contexts (Aldridge et al., 2022) and demonstrated acceptable reliability for the current sample (αs > .70).

Provider interviews: determinants of the TeACH System, other DMH LITs

Interviews began with an overview of the proposed design of the TeACH System. Providers commented on both the TeACH System and similar DMH LIT integration into their PPCs. Questions were structured based on CFIR constructs (Damschroder et al., 2009, 2022): Compatibility (e.g., “How well does the intervention fit with existing work processes and practices in your setting?”), Relative Priority (e.g., “To what extent might the implementation take a backseat to other high-priority initiatives going on now?”), Available Resources (e.g. “Do you expect to have sufficient resources to implement and administer the intervention?”), Self-Efficacy (e.g., “How confident are you that you will be able to successfully implement the intervention?”), and Champions (“Other than the formal implementation leader, are there people in your organization who are likely to champion the intervention?”) in their clinics (Damschroder et al., 2009, 2022).

Data analysis

Analytic approach

An exploratory sequential design was used (Creswell & Plano Clark, 2023), with quantitative survey data characterizing the general practice landscape of teen mental health support and DMH in PPC, with qualitative data elaborating upon the determinants driving current DMH usage practices and perceived feasibility in two specific PPCs. We used a connecting process for data integration, wherein quantitative and qualitative data were analyzed independently, and findings were integrated at the interpretation stage to elaborate upon one another (Palinkas et al., 2011).

Qualitative analysis. This study drew from a pragmatist orientation where multiple understandings of reality drive practical actions to serve the interests of the community (Cornish & Gillespie, 2009). Recognizing that positionality influences the design, methods, and analysis of a study and is connected to the researchers’ personal and philosophical views, we include positionality statements in the Supplementary Materials.

We used rapid qualitative assessment analysis (RQA) to identify emergent themes from the interviews (Hamilton, 2020). RQA is time-efficient and has demonstrated similar outcomes as thematic and in-depth qualitative analyses (Gale et al., 2019; Taylor et al., 2018). RQA is particularly appropriate for questions related to implementation science (i.e., examining implementation determinants), because it is rigorous and time-sensitive, thereby allowing for more responsive action to real-world implementation challenges (Hamilton & Finley, 2019). The first step of using RQA in the current study was to create a summary template with domains generated deductively based on the interview questions/CFIR and inductively based on transcript review. Consistent with RQA methods (Hamilton, 2013), a practice transcript was coded to establish consistency in use of the summary template, extracting key excerpts to enter into the template. The coders met regularly to minimize drift and resolve questions. The coders independently reviewed all data in the summary template, summarizing key points and noting emergent themes. They then discussed their notes to identify and refine themes.

Quantitative analysis. Survey data were examined via descriptive analyses.

Results

Participants

The survey was completed by 105 Chicago providers serving teens in medical clinics. Table 1 and Supplementary Tables detail the clinic, respondent, and community characteristics. Survey providers worked in clinics situated in communities with a high variability of well-being, ranging from 4.2 (0 indicates no distress) to 99.1 (100 indicates severe distress; M  = 57.33 ± 32.71). On average, 71.98% (SD = 16.82) of the homes immediately surrounding the clinics had broadband internet access (range: 46.10%–94.0%). Clinics were described as general pediatrics/primary care (38.1%; n = 40) or a pediatric subspecialty clinic (e.g., endocrinology; 26.7%; n = 28). Nearly half of clinics were described as serving specialty populations (e.g., LGBTQI+, refugee; 48.6%; n = 51). Providers were most often pediatricians (50.5%; n = 53), psychologists (14.3%; n = 15), or advanced practice registered nurses (14.3%; n = 15).

Four pediatricians (66.7%) and two pediatric psychologists (33.3%), who serve teen patients in two PPC clinics located on the West Side of Chicago, were interviewed. These PPCs are located in communities considered to be at risk (distress score of 75.7) and 55.6% of homes in these communities had broadband internet access. The sample size is justified by having: (1) narrowly defined objectives (Hennink & Kaiser, 2022), and (2) recruited participants with “information power” (Malterud et al., 2016) via purposive sampling (Hamilton & Finley, 2019).

Mixed-methods results

Quantitative and qualitative findings were integrated at the interpretation stage after independent analyses. As such, we present quantitative and qualitative results alongside one another to facilitate their combined interpretation. Further, we indicate throughout the results whether data reflects the TeACH System specifically, or DMH LIT broadly.

RQA of the interviews resulted in six themes describing different CFIR-related implementation determinants (Damschroder et al., 2009, 2022): (1) PPC clinic workflow, (2) hospital & healthcare system factors, (3) responsibility and ethical considerations, (4) patient privacy and confidentiality, (5) DMH design, and (6) a cross-cutting theme of safety. While guided by CFIR, category and theme names are based on the content of providers’ statements, and hence do not reflect formal CFIR construct names. Instead, we note the specific CFIR constructs corresponding to themes in Table 2. Hence, safety topics are highlighted across results for each of these themes. Quotes are labeled below by provider type (i.e., Ped or Psych) and interview number (e.g., “Ped 1”).

Pediatric primary care clinic workflow

The PPC clinic workflow theme refers to clinic processes and infrastructure logistics related to the patient and provider experience from waiting room to post-appointment. Providers noted several factors in the PPC workflow that would impact current use and the feasibility of DMH LIT (including the TeACH System) deployment in this setting, described as follows.

Survey providers indicated that DMH LITs were typically not part of their clinic workflows. Administration of common self-report mental health screeners (see Supplementary Table S1) most commonly occurred using paper and pencil (41.0%; n = 43) or the electronic medical record (EMR)/“MyChart” (patient-facing portal for EMR; 19.0%; n = 20). In the past year, most providers reported not using DMH to screen teens for mental health symptoms (63.8%; n = 67) and about half had not referred teen patients to any type of DMH LIT (e.g., trying out the mindfulness app, “Calm”; 45.7%).

Providers consistently characterized clinics as fast-paced, making it hard to plan when to administer the TeACH System. Yet, they also recognized that there can be frequent, but fragmented, downtime for patients (e.g., “Especially in patients that are seeing residents, because it’s just sort of, like, creating the system so that you can monopolize on that downtime, which is not necessarily all in one chunk.” [Ped 1]). Some survey (28.6%) and interview providers recommended conducting DMH mental health assessment or LIT at home, before the appointment (e.g., “If you’re giving the questionnaires afterward, that means that the provider has to go back and revisit because they may have not even been there for that visit, or that reason, and how to deal with it” [Ped 3]).

The waiting room was cited by survey providers (30.5% preferred waiting room for DMH screening) and interview providers as a likely spot in the workflow when teens could engage with DMH LITs. Yet, concerns were raised about the limited privacy in this mixed space: “Your issue is PHI. You don’t have any privacy [… and] you already have the stigma associated with mental health…you won’t have people want to engage, even if they’re interested.” (Ped 3). Further, the waiting rooms can mix with specialty clinics, calling into question if specific DMH-driven screenings, such as the K-CAT administration in the TeACH System, is appropriate for all patients in this setting.

Interview providers also noted limited staff capacity, questioning who would be responsible for patient follow-ups if teens screened positive for mental health symptoms. There was disagreement on whether the responsibility fell on the pediatrician or other staff if the TeACH System or another DMH LIT assessed for mental health symptoms outside of typical clinic assessments (e.g., PHQ-2). Providers questioned how their workflow could accommodate the need for referrals to care: “I don’t know if the pediatricians can be asked to do much more.” (Psych 2). Further, it was noted that safety issues can already complicate the clinic workflow at present, and would need to be clarified for any DMH integration:

And, they’re like, “Oh, shoot!” they checked they are suicidal. And they just walked out the door and they didn’t tell me that but they checked it on this paper! So, there’s certainly downsides, the way we currently do it. But, that’s the biggest thing: when you’re asking these questions, you need to know what the resources are going to be able to be. If you identify these things, we have so few resources in our clinic, even for anxiety, depression. (Ped 2)

There was also a lack of consensus on specific “Champions,” or those who would go “above and beyond” to integrate the TeACH System into PPC. This may have been driven by mixed enthusiasm for DMH LIT integration, broadly. Survey providers rated using DMH in their clinics as “A little bit” Appropriate (i.e., mhIST self-perception of effectiveness; 1.65 ± .70) and for Accessibility/Reach (1.16±.77), and “Not at all” for Feasibility (i.e., skills, time, resources; .88±.86). While interview providers indicated general enthusiasm (e.g., “I think adolescent providers would be super excited for it” [Ped 2]), they were unable to identify Champions in their clinics. Some mentioned the broad possibility of engaging residents, social workers, and/or staff to rally as champions, but that funding to support time would influence their enthusiasm.

Lastly, PPC technology infrastructure was also cited as a potential implementation barrier for the TeACH System. EMR integration was supported by both survey providers (preference for MyChart link for DMH screeners, 14.3%) and interview providers, noting that this would influence the feasibility of use: “But that’s my biggest concern. It’s like they endorse some really concerning things, or have a big flag, and then that never gets to a clinician.” (Ped 2). Further, providers noted potential challenges related to Wi-Fi service and difficulties of tracking devices across a potentially interrupted workflow.

Hospital and healthcare system factors

This theme refers to the broader healthcare and hospital ecosystem in which PPC clinics are situated, and how this system impacts DMH LIT use in PPC via resources and policies. Survey providers reported anxiety (86.7%), depression (82.9%), and behavioral problems (58.1%) as teen patients’ primary concerns—and the top reasons for behavioral health referrals. These data were included in this theme, as respondents noted that after identifying mental health needs, identifying referrals is difficult due to a “severe shortage of practitioners, including practitioners who accept Medicaid and who are located in the patient’s catchment areas” (survey provider). Interview providers underscored the lack of resources in the broader healthcare system to address these behavioral health concerns:

It’s because of the healthcare system. We don’t have enough time to see patients. We don’t have enough mental health. We don’t have enough social workers. We’re stretched and we continue to take on different things. So, it’s not that [TeACH System] isn’t appropriate but it just makes more stress for the provider and burnout as well. …The reason why you’re having to do this intervention is because primary care providers are having to take up the fight in terms of behavioral health where there should be behavioral health providers doing it. (Ped 2)

Providers went on to describe feeling overburdened in this environment, and constrained in their ability to respond to mental health needs given limited resources: “What do we do when it’s positive, and we just don’t have the… resources…?” (Psych 2). Other providers acknowledged some existing resources built into the hospital infrastructure (e.g., collaborative care), but expressed concerns about further burdening the few existing services if more patients were identified as needing care. Psych 1 noted:

But, just an anxious kid, they’re not going to pop on the PHQ-9, and let’s say they’re not like, disclosing in the appointment … And now they’re like, kind of like screening positive [because of TeACH System use] … adding to the potential referral burden on collaborative care. Umm … needing resources to manage that.

The overburdened perspective may lend insight into low targeted and universal mental health screening reported by survey providers. Despite recommendations for universal depression and anxiety screenings for all teen patients in primary care (Walter et al., 2020; Zuckerbrot et al., 2018), universal administration of the Patient Health Questionnaire-9 (PHQ-9) and Generalized Anxiety Disorder-7 (GAD-7) was endorsed by only 52.4% and 22.9% of the sample, respectively.

Additionally, interview providers identified hospital rules and policies that may inhibit how integration might work for teen patients, who might prefer communicating through text message. For example: “It’s hard enough even to send an appointment reminder via text…The hospital is very, they don’t budge on anything” (Ped 2). Further, hospital regulatory systems require a clear protocol for safety concerns that may arise in any type of mental health screening, including through digital methods: “Your person who’s giving it has to be trained to deal with that situation, so that would have to be an LCSW, LCPC, MD, or something, or the IRB is gonna be like: ‘So what are you gonna do?’ And then you have to have a protocol in terms of if that child is positive: What are your next steps?” (Ped 3).

Responsibility and ethical considerations

In this theme, providers reflected on their responsibility and ethical concerns in terms of how they respond to patients’ identified mental health needs in the context of digital assessment. Providers spoke to a sense of ethical responsibility to respond to patient-reported mental health needs: “This is my office so, I feel personal responsibility.” (Ped 4). Perhaps reflecting this feeling of responsibility, about a third of survey providers reported that they verbally administer mental health screeners directly to their teen patients (32.4%; n = 34). Interview providers stated that mental health endorsements require an immediate response, particularly for safety concerns related to depression:

You can’t ask that question without addressing it. I mean, we specifically don’t put our PHQ-2’s on MyChart. We have those only in the office. And we, the physicians, ask the questions. We don’t even allow the MAs to do it. We feel like, if a teen is gonna divulge that kind of information, they need an immediate response. And what if my MA did it, and then I missed it, and the teen goes off to football practice? (Ped 4)

Providers’ sense of responsibility to respond to patient-reported mental health needs intersected with their concerns about broader mental health resources (see Hospital & Healthcare System Factors theme), as the lack of available resources limits their ability to respond to patient needs.

A related ethical consideration was around providers’ potential responsibility to alert caregivers to their child’s mental health needs: “There’s sort of this moral question of, like, you get information on any of those diagnoses which, with maybe the exception of ADHD, and I feel like maybe somebody should be alerted, like a parent.” (Ped 1). However, about half of survey providers (46.7%) reported that their teen patients have requested that their mental/behavioral health symptoms not be shared with their primary caregiver. How to manage disclosure was an unanswered question for providers, but it was noted as something that would need to be addressed to deploy the TeACH System in PPC.

Teen patient rights, privacy, and confidentiality

This theme referred to data sharing, privacy, confidentiality, and broadly, what a teen might want control over in terms of sharing mental health information. Providers raised questions about the potentiality of teens not wanting their DMH data to be shared with providers and/or guardians (as noted above), and challenges to maintaining teen confidentiality. MyChart was one example of how something intended to be only for a teen patient may become accessible to caregivers (e.g., “All the time, I have parents who have access inappropriately, and then, I will you know, take away their access and then they’re pissed” [Ped 1]). Similarly, interview providers wondered about confidentiality from caregivers in the waiting room: “The parent’s sitting right next to the kid? Looking over their shoulder?” (Ped 4).

While wanting to honor teen’s privacy boundaries, interview providers also noted the potential benefits of sharing screening results with guardians: “it is helpful to keep them in the loop in certain ways that help them understand anxiety, and how they can help support their kid.” (Psych 1). Further, the question of caregiver involvement was underscored in the context of reported suicidality or substance use: “Sometimes they may need to know, right? Sometimes, if there’s a real safety concern, with suicide or with substance use.” [Peds 1]).

Another consideration raised was around how data would be managed for wards of the state and children under the care of the Department of Children and Family Services or “people who come to the doctor without their primary caregiver, so they don’t want their aunt or their uncle or their foster care, and all that.” (Ped 3). Thus, the line between pediatric patient autonomy and caregiver/guardian involvement was acknowledged as quickly becoming complicated in teen healthcare; this is an additional implementation determinant that will be necessary to address for all DMH LIT integration, including the TeACH System.

TeACH System design

The DMH Design theme encompassed preferences and concerns around the design, usability, and characteristics of the TeACH System (i.e., screening, feedback, referral). Design considerations were shaped by factors related to modality, language, and understanding.

Providers emphasized that teens are most likely to engage with content presented via social media. Consistent with the design they were presented, providers recommended social media referral options grounded in evidence-based skills (e.g., a TikTok reel portraying progressive muscle relaxation): “I like the idea of being able to give them resources that are vetted. So, I like the idea of having a set group of resources to give kids. And we know they use social media. So, of course, I like those” (Ped 4). They contrasted this with teen’s poor engagement with apps recommended in clinic (e.g., “… and I say, let’s download it, and they, like, don’t, even though they’re with me.” [Psych 1]) They also noted that teens would want to see themselves represented in graphics, with a focus on “mak[ing] it more diverse” (Ped 3).

Providers raised questions about design considerations related to: (1) linguistic diversity (i.e., whether available in English and Spanish; “10% or less of our patients and families speak Mandarin or Cantonese” [Ped 1]); and (2) what teen patients would learn and understand from the TeACH System’s screening and resources. When queried about commonly assessed symptoms on computerized adaptive tests (i.e., the full K-CAT yields severity scores for anxiety, mania/hypomania, oppositional defiant disorder, attention-deficit/hyperactivity disorder, depression, conduct disorder, suicidality, substance use disorder; Gibbons et al., 2019), it was noted that: “I don’t think some of the patients would know what some of these things [diagnoses] are. So, you would have to use more [lay] language” (Ped 1). Providers also expressed concern about mental health labels being placed on teens without learning their broader context: “I think we’re both aware about the risk of getting that label put on a teen. It carries a lot of potential, like, longer term impact that I just wouldn’t want. I would hate for somebody who doesn’t have the ability to, like, really make that diagnosis, to see the screener and, like, throw it on their chart, or something like that” (Ped 2). Further, providers wanted more information on whether the K-CAT (screener for the TeACH System) could: (1) differentiate between active and passive suicidal ideation; and (2) include other risk assessments, such as non-suicidal self-injury or disordered eating.

Discussion

This study was situated at the pre-implementation phase (Chamberlain et al., 2011; Cornish & Gillespie, 2009) and used a mixed-methods approach to better understand the implementation context for a DMH LIT (i.e., integrating the TeACH System into PPCs serving teens on the West Side of Chicago). Quantitative results highlighted clinic catchment areas characterized by a high need for behavioral health services, inconsistent use of universally recommended mental health screeners, low usage rates of DMH in Chicagoland medical clinics serving teens, and ultimately low perceived feasibility of DMH LIT integration. In compliment, qualitative findings highlighted consistent mention of the high need for teen behavioral health services. Qualitative results elucidated multiple inner and outer setting domain-related determinants uptake and integration, likely underlying the low DMH usage and feasibility ratings from the quantitative findings. Namely, providers were skeptical of TeACH System integration in their clinics due to providers’ high workload and hospital/clinic administrative practices that impede integration. Further, providers noted that if the TeACH System increased symptom identification in teen patients, they worry about how to support them given limited referral options outside of the TeACH System’s own referrals. These concerns also tied to questions of ethical dilemmas and complications around teen confidentiality that need to be addressed for any DMH LIT integration. In sum, the findings indicate that providers are eager for more behavioral healthcare options for teens, but recognize multiple barriers that would impede DMH in successfully providing these services in their setting.

The TeACH System

CFIR-related determinants established through pre-implementation activities directly informed the TeACH System’s subsequent iterative design and implementation planning for PPC clinics serving teens on the West Side of Chicago. For example, the theme of DMH Design indicated concerns about digital screenings generating mental health labels in the absence of the broader diagnostic and psychosocial context. Further, the theme of Hospital & Healthcare System Factors reflected concerns about screening for multiple mental health disorders in the absence of broader system infrastructure to address the identified needs. As such, the use of the K-CAT as a screening tool in the TeACH System was limited to the anxiety module only (rather than multiple modules assessing additional mental health domains; Gibbons et al., 2019). A focus solely on anxiety while the System and its integration are optimized during pilot feasibility testing: (1) aligns with recommended anxiety screening for all teens in PPC (Mangione et al., 2022); (2) decreases the amount of time required for engagement across busy PPC workflows (i.e., the full K-CAT administration is ∼8 min, while the K-CAT anxiety module administration is ∼1 min); and (3) removes the possibility that the TeACH System will uncover safety-related items (e.g., suicidality, substance use) that were of most concern to providers, even with proposed TeACH System protocols and support team presence. This design change illustrates how the assessment of pre-implementation determinants is a critical step for future implementation feasibility. Without provider feedback, the TeACH System would have been delivered with screening modules that would have posed significant barriers for providers to respond to, which in turn would limit its usability and long-term sustainability. This study provides an exemplar for future deployment of DMH LITs in PPC, highlighting the necessity of understanding the unique contextual factors relevant to specific PPC settings to promote DMH fit (Chamberlain et al., 2011; Mohr et al., 2017).

Broad DMH LIT integration

While the pre-implementation activities presented in the current study were aimed at informing TeACH System integration in specific PPCs, the design and purpose of the System is similar to other DMH LITs in that it provides a patient-facing tool for screening, psychoeducation, and referrals to brief interventions that may be accessed and/or completed at the discretion of and best time for the patient (whether in or outside of the healthcare setting). As such, the findings provide a case example of pre-implementation factors relevant for DMH LITs, and may also provide a model for approaching the integration of DMH LITs in other urban pediatric care settings. Indeed, there have been repeated and growing calls to improve the reach, uptake, and tailoring of DMH LITs for teen patients—particularly teens with intersectional and/or minoritized identities (Bounds et al., 2022). Methods grounded in social justice and centering lived experience and promoting and amplifying diverse voices in design, clinical, and research frameworks and spaces (Figueroa et al., 2021, 2022; Friis-Healy et al., 2021; McCall et al., 2022; Ramos et al., 2021), are integral to creating and disseminating more equitable and effective DMH LITs. However, a lack of consideration of the implementation of DMH has plagued pediatric DMH broadly (Psihogios et al., 2022), indicating that the field has not only been neglecting input from teens and caregivers, but also overlooking the expertise of providers. All providers in the current study were acutely aware of the systemic barriers that limit effective teen behavioral healthcare. Traditionally noted barriers to carrying out recommendations for universal mental health screening, such as time (Acri et al., 2018), were not identified as directly prohibiting administration to teens. In fact, providers strongly voiced a feeling of responsibility to take the time to assess mental health themselves during teen patient sessions. The driving barrier for the providers in this study was a reluctance to expand mental health screening through digital delivery without adequate additional resources to support teens who need care beyond DMH. As such, introducing DMH LITs as a first line of care for primary care patients will not be acceptable or feasible to providers without also knowing that there are accessible and timely second and third lines of care for patients with higher needs. Coverage for patients who require more care than a DMH LIT may offer is likely a key determinant for integration in PPC moving forward.

While this study has multiple strengths, it should also be considered in light of specific limitations. First, while it was necessary for the goals of the study to focus on the implementation context of PPCs in Chicago, the findings may not generalize across diverse geographical contexts. However, the specificity of the current study activities might serve as a case example for pre-implementation planning for other LITs in pediatric healthcare settings. Second, to broadly characterize the landscape of behavioral healthcare for teens in medical clinics, survey provider participants included clinicians outside of PPC (e.g., subspecialty clinics). This expansive perspective may therefore implicate practices less pertinent to PPC. Further, given the anonymous nature of the provider survey, it is unknown if there was overlap across survey provider and interview provider participation. Third, interview providers were shown design details for the TeACH System early in their interviews, which may have influenced their approach to considering broader DMH LIT integration. Finally, while purposive sampling occurred to recruit providers with “information power” (Hennink & Kaiser, 2022), the sample of interview providers was relatively small and limited to Pediatricians and Pediatric Psychologists. While this study focused on the perspectives of providers in specific Chicago PPCs, implementation research must also be informed by the lived expertise of teens, caregivers, and other PPC providers (e.g., nurses, MAs, administrators). As such, future research should partner with and examine the experience of these other groups when assessing the feasibility of DMH in PPC.

DMH LITs stand as a potential low-cost and scalable way to expand mental health screening and connection to resources for teens in PPC settings. Yet, a focus on design to the neglect of implementation has plagued pediatric DMH (Psihogios et al., 2022). To join calls to consider implementation from DMH design to sustainment (Graham et al., 2020; Mohr et al., 2017), the current pre-implementation study assessed determinants relevant to integrating a specific DMH LIT, the TeACH System, into PPCs serving teens on the West Side of Chicago. Low perceived feasibility for integration is a call to action, rather than a deterrent for the use of digital tools and services for teens in PPC. For the TeACH System, responsive design and implementation planning changes occurred to address provider concerns for upcoming evaluation and optimization phases (Mohr et al., 2017). Yet, to best support diverse teen needs in PPC settings broadly, pediatric DMH LITs will need to repeatedly assess and address pertinent questions relating to workflow integration, ethics, and ability to streamline access to evidence-based, accessible care–even beyond the DMH, itself.

Supplementary material

Supplementary material is available online at Journal of Pediatric Psychology (https://dbpia.nl.go.kr/jpepsy/).

Conflicts of interest: None declared.

Funding

Research reported in this publication was supported in part by a grant from the National Institute of Mental Health of the National Institutes of Health (K08 MH125069). The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health. The authors have no financial or other conflicts of interest to report.

Author contributions

Colleen Stiles-Shields (Conceptualization [lead], Data curation [lead], Formal analysis [equal], Funding acquisition [lead], Investigation [lead], Methodology [equal], Project administration [lead]), Erika Gustafson (Formal analysis [equal], Methodology [equal], Writing—Original draft [equal], Writing—Review & editing [equal]), Paulina S. Lim (Methodology-supporting, Writing—Original draft [supporting], Writing—Review & editing [supporting]), Gabriella Bobadilla (Project administration [supporting], Writing—Original draft [supporting], Writing—Review & editing [supporting]), Dillon Thorpe (Project administration [supporting], Writing—Original draft [supporting], Writing—Review & editing [supporting]), Faith C. Summersett Williams (Writing—Original draft [supporting], Writing—Review & editing [supporting]), Geri Donenberg (Methodology [supporting], Writing—Original draft [supporting], Writing—Review & editing [supporting]), Wrenetha Julion (Methodology-supporting, Writing—Original draft [supporting], Writing—Review & editing [supporting]), Niranjan Karnik (Conceptualization [supporting], Resources [supporting], Writing—Original draft [supporting], Writing—Review & editing [supporting]).

Data availability

The data underlying this article will be shared on reasonable request to the corresponding author.

References