Barriers to Increasing Access to Brief Pediatric Mental Health Treatment From Primary Care

More than 25% of pediatric primary care patients present to care with a psychosocial problem (1), yet less than one-third of children referred to mental health treatment by their primary care provider (PCP) complete an outpatient visit (2, 3). Although most parents report interest in receiving child behavioral treatments through primary care (4), existing referral and handoff processes to mental health care are insufficient to engage the majority of families. Additionally, evidence-based treatments (EBTs) for pediatric mental health problems can be lengthy, cost-intensive, and burdensome. Common barriers to engagement include lack of trained providers, limited treatment capacity, and logistical and transportation problems among patients. Thus, there is a need to improve integration of mental health care with primary care through consultation or team-based processes and to develop briefer and more targeted EBTs to increase treatment uptake, retention, and reach (5).

In response to a state initiative to improve access to behavioral treatments in remote areas with minimal uptake of and demand for primary care–embedded mental health services, we conducted a quality improvement process to provide rapid access to brief behavioral treatment for children and adolescents in an underserved region of Washington State. The process created pathways from primary care clinics and schools to regional mental health clinicians trained in brief EBTs. We evaluated the feasibility of implementing the program, acceptability of the model, and preliminary clinical outcomes. We hypothesized that increasing availability of and access to brief EBTs would increase service uptake for pediatric primary care patients. The 2-year project period for working within existing community care systems provided opportunities to understand system barriers and to test solutions designed to improve care delivery and quality.

The state-funded quality improvement process was conducted in Benton and Franklin counties, located in South Central Washington State, with limited specialized health services and without integrated services in primary care. Over 29 months, we sought referrals from primary care (and later from schools) of children with disruptive behavior problems and adolescents with depression, all with Medicaid insurance.

We developed brief treatments to enhance system capacity and family engagement. Two four-session First Approach Skills Training (FAST) treatment manuals were adapted from full-length EBTs and reviewed by child clinical psychologists, psychiatrists, and community therapists. Both manuals are available for free (https://www.seattlechildrens.org/healthcare-professionals/access-services/partnership-access-line/pal-plus). The programs were designed for patients with mild-to-moderate acuity without immediate safety risks or with a different primary treatment need. We supported patients with higher acuity and needs in accessing other services through the same community mental health centers where FAST programs were delivered.

The FAST-Behavior (FAST-B) program was designed for children ages 4–12 with a primary disruptive behavior problem, including oppositional behavior, attention-deficit hyperactivity disorder, parent-child relational problems, and/or adjustment problems. Content was adapted from the Defiant Children manual (6) and included skills training for one-on-one play time, labeled praise, planned ignoring, incentives, and time-out. Children referred to more intensive community services included those with autism spectrum disorder, primary posttraumatic stress or depression, and open child welfare investigations.

The FAST-Depression (FAST-D) program was developed for adolescents ages 12–17 with mild-to-moderate depressive symptoms. The protocol was adapted from the Behavioral Activation for Adolescent Depression manual (7) and included psychoeducation on depression, sleep hygiene, goal setting, and activities planning. Adolescents with primary anxiety disorders, posttraumatic stress, substance abuse, eating disorders, bipolar disorder, ongoing self-injury, or active suicidality were referred to higher-intensity care.

We remotely trained three mental health clinicians from community mental health agencies to deliver the programs. Training was manual based and consisted of 4–6 hours of training, via videoconferencing, with clinical psychologists specializing in EBTs as well as weekly phone consultation.

To improve the referral and handoff process to behavioral treatment, we created a one-step phone or fax referral pipeline directly to clinicians. We advertised the program to PCPs through a regional medical conference, an e-mail registry, and recruitment visits to primary care offices. When program capacity remained after 14 months, we invited school-based referrals through phone calls and school staff trainings.

After being referred to the program, families received up to three outreach phone calls and one letter from the clinician within 2 weeks. Responding families were phone screened for eligibility and then invited to attend an in-person screening at the clinician’s clinic. Enrolled families were offered four free 1-hour FAST treatment sessions. Clinicians sent “faxbacks” to PCPs describing the referral outcome and sent treatment summaries if the families had enrolled. Families that were ineligible or declined the programs were connected to other local services. Because FAST clinicians were located within community mental health centers, they could provide immediate access to full-length treatments in that setting and, in many cases, could provide the higher-level intervention themselves or offer a direct handoff to a colleague.

For FAST-B, we received 140 referrals (N=104 from PCPs), and 84 families responded to contact. Of these, 47 families attended the screening (34% of referred), 46 enrolled, and 26 completed the program. Mean±SD age was 7.3±2.27, and 77% (N=36) were male. Of those who attended the screening, 28% (N=13) identified as White, 9% (N=4) as mixed race, and 30% (N=14) as Hispanic; nine participants spoke primarily Spanish. The primary reasons for declining the program were caregiver’s preference for individual child therapy and family scheduling barriers. Several children were screened out during the initial phone call because of a primary mood disorder, suicidality, high-risk aggression, or trauma-related problems and referred to traditional community mental health treatment services. Of those who attended the screening, 43% (N=20) completed all sessions.

For FAST-D, we received 80 referrals (N=45 from PCPs), and 58 families responded to contact. Of these, 38 adolescents attended screening (48% of referred), 21 initiated treatment, and 15 completed the program. Mean age was 13.6±1.53, and 55% (N=24) were female. Of those who attended the screening, 19% (N=7) identified as White, 5% (N=2) as mixed race, and 10% (N=4) as Hispanic; two participants spoke primarily Spanish. Notably, PCPs referred fewer patients to FAST-D than to FAST-B and tended to refer individuals with more severe and complex cases. We received nearly an equal number of referrals to FAST-D from schools as from PCPs and in only half the time. A majority of youths screened for the program showed severe depressive symptoms, suicidal ideation, or other risk factors necessitating more intensive community treatment, which was facilitated by the clinician. Of those who attended the screening, 71% (N=15) completed all sessions.

We administered family acceptability questionnaires adapted from existing surveys (8) privately after the final session. Questionnaires were completed by 23 FAST-B caregivers, 15 FAST-D caregivers, and 13 adolescents who attended FAST-D. All respondents reported that the program was helpful, that they would recommend it, and that they were satisfied overall. Most caregivers in both tracks reported that the program met most or all of their needs. Most FAST-B caregivers (N=20 of 23) agreed that there were enough sessions, whereas nine of 15 FAST-D caregivers and nine of 13 of adolescents agreed. Most adolescents (N=7 of 13) reported they would have been “not at all likely” to seek mental health treatment if not offered FAST-D.

Study clinicians completed an acceptability rating scale after seeing several patients and again after the project ended. Initial clinician acceptability for FAST-B (N=3 clinicians) was very high for ease of use and comfort with the manual; high for training, user-friendliness, consultation, and appropriateness of content; and moderate for flexibility and length of the program. FAST-D acceptability (N=2 clinicians) was very high for user-friendliness, ease of implementation, training, and consultation and high for flexibility, appropriateness, and length. After the pilot, FAST-B clinicians (N=2) rated all acceptability items highly, except for split responses (moderate/high) on fit of the program. The FAST-D clinician (N=1) rated all aspects as highly acceptable.

PCPs referring at least two patients received a digital satisfaction questionnaire, and five of the 13 who completed it indicated that the program was easy to refer to, was a valuable additional service, and addressed patients’ needs. Four of five reported that communication with the clinician was easy. PCPs were split (three agreeing, two disagreeing) on whether they could now better meet patients’ mental health needs.

FAST-B parents rated child behavior problems on the Home Situations Questionnaire (HSQ) at each session. Adolescents attending FAST-D completed the Patient Health Questionnaire–9 at each session; their caregivers completed the Short Moods and Feelings Questionnaire at baseline and the final session. All caregivers completed select subscales of the Weiss Functional Impairment Rating Scale–Parent (WFIRS-P) at baseline and the final session. We estimated preliminary effect sizes of clinical outcomes for families who completed at least one treatment session by using paired t tests and the formula for Cohen’s d_av (9). For FAST-B, there was significant improvement on the HSQ (t=2.79, df=45, p=0.008; d_av=0.40) and WFIRS-P (t=4.39, df=17, p<0.001; d_av=0.81). FAST-D had a smaller sample and small but nonsignificant effect size for adolescent-reported depressive symptoms and functional impairment and a moderate but nonsignificant effect size for parent-reported depressive symptoms.

We originally hypothesized that creating brief EBTs with direct referral pathways from primary care would increase pediatric mental health treatment uptake in an underserved region. However, our service model was insufficient to accomplish this goal and required improvements along the way. Our project succeeded in increasing availability of brief EBTs; acceptability was high for participating families, most of whose needs were met by brief treatment. Program completion rates were comparable to mental health care generally and better than therapy in community mental health settings (10). However, our approach did not substantially increase treatment uptake. Rates of in-person session attendance for referred patients (34% [N=47] for those with disruptive behavior and 48% [N=38] for those with depression) exceeded the 30% threshold observed in previous studies (2, 3), but most referrals still did not initiate treatment. Our program highlights continued service barriers and potential solutions to improve treatment access. The introduction of a novel provider and location outside of the familiar primary care environment and a time gap of several days since leaving clinic likely constituted barriers to access and engagement. Over time, not being colocated and integrated within the clinic likely also diminished our program’s visibility to PCPs, who are notoriously busy. Inappropriate referrals received support in accessing alternative services, but nonenrollment in our program may have discouraged referrers. The lack of routine behavioral health screenings in local practices and PCP bandwidth for in-depth mental health assessment may have prevented milder cases from being identified. Relative to PCPs, school staff appeared better able to identify adolescents with depression, highlighting benefits of including schools in primary care and mental health care collaborations.

Our project demonstrated the need for integrated mental health care to provide a “warm handoff” in a comfortable and familiar primary care setting. Routine pediatric mental health screening measures allow PCPs to identify patients with mild-to-moderate acuity cases and make immediate treatment recommendations. After the initial project period, we transitioned to provide FAST training and weekly videoconferencing consultation to mental health clinicians integrated within pediatric primary care practices across Washington State and incorporated their feedback to improve FAST usability in collocated service settings. We have observed that referrals and handoffs are more effective in this context. By request from PCPs, we also developed a FAST pediatric anxiety manual. Additionally, telehealth delivery during the COVID-19 pandemic has facilitated program access and should remain a delivery format for brief treatments. We also learned that intensive community treatments are often unavailable, and our brief programs likely constitute an appropriate first step for families waiting to initiate additional care. Referring families with higher acuity to a separate program creates additional care barriers for them. One benefit of offering FAST in community mental health centers was that no additional contact was needed for most referred families to initiate more intensive treatment. However, integrated mental health care should incorporate stepped-care models, in which lower and higher acuity services are available with a single entry point or completion of a lower-intensity treatment leads directly to a higher level of care, when needed.

Our project was limited in scope. Only those who completed the program rated acceptability, and those who initiated the treatment reported clinical outcomes; families with more hardship and barriers were underrepresented by our data. A strength of the program was its deployment focus, meaning that our model could be replicated in other communities with traditional health care infrastructure. Our program could be appropriate for remote areas with insufficient demand for embedded mental health care because a single regional program can serve many clinics. Furthermore, implementing this project in a “real-world” setting allowed us to observe and respond flexibly to barriers, for example, by increasing communication with PCPs, adding school referrals, and partnering with primary care–embedded clinicians during a second phase. Our findings demonstrate that brief behavioral treatments can expand system capacity and meet the needs of lower-acuity families, but do not substantially increase service uptake. Rather than replicating our program model, future efforts should focus on integration with primary care, where families have greater access to and comfort with treatment and care teams can work collaboratively.