Behavioral Health Service Utilization Research Articles

Parent SMART: Effects of residential treatment and an adjunctive parenting intervention on behavioral health services utilization

IntroductionScant research has examined the impact of residential treatment on adolescent behavioral healthcare utilization post-discharge, even though behavioral healthcare utilization is major driver of healthcare costs. In the primary analyses of a pilot randomized trial, Parent SMART – a technology-assisted intervention for parents of adolescents in residential treatment – was found to improve parental monitoring and parent-adolescent communication, reduce adolescent drinking, and reduce adolescent school-related problems, relative to residential treatment as usual (TAU). The goal of this secondary analysis of the pilot randomized trial was to assess the effects of residential treatment and the adjunctive Parent SMART intervention on both the amount and type of subsequent behavioral healthcare utilization. MethodThe study randomized sixty-one parent-adolescent dyads to residential TAU (n = 31) or residential TAU plus Parent SMART (n = 30). Of the 61 dyads, 37 were recruited from a short-term residential facility and 24 were recruited from a long-term facility. Adolescents completed a structured clinical interview and self-reported their behavioral health-related visits to the emergency department, nights in residential/inpatient, and outpatient visits over the past 90 days, at baseline, 12-, and 24-weeks post-discharge. Generalized linear mixed models (GLMMs) examined both linear and non-linear (pre- to post- residential treatment) trends, pooled, and stratified by residential facility to examine behavioral health service utilization. ResultsBoth the linear and pre-post GLMMs revealed that behavioral health-related emergency department visits and residential/inpatient nights decreased across both residential facilities. GLMMs estimating change from the pre- to post period indicated that outpatient visits increased across both facilities. There were no significant effects of the Parent SMART adjunctive intervention in GLMMs, though bivariate tests and the direction of effects signaled that Parent SMART was associated with more nights of residential/inpatient utilization. ConclusionResidential substance use treatment may reduce adolescents' subsequent utilization of costly behavioral healthcare services such as emergency department visits and residential/inpatient nights, while increasing utilization of outpatient services. Parent SMART was not associated with significant changes in behavioral healthcare utilization, but the pattern of results was consistent with prior literature suggesting that stronger parenting skills are associated with greater utilization of non-emergency services.

Associations Between Patient Comfort with a Primary Care Provider and Three Measures of Behavioral Health Services Utilization.

Behavioral health services (BHS) can help improve and treat mental and emotional health problems. Yet, attitudinal and/or structural barriers often prevent individuals from accessing and benefiting from these services. Positive provider-patient interactions in healthcare, encompassing patient comfort with a primary care provider (PCP), which is often enhanced by shared decision-making, may mitigate the stigma associated with seeing a mental health professional; this may improve BHS utilization among patients who need these services. However, few studies have examined how patient comfort with a PCP, often through shared decision-making, may influence patients' BHS utilization in the real world. This study sought to address this gap in practice. Multivariable regression analyses, using weighted data from an internet panel survey of Los Angeles County adults (n = 749), were carried out to examine the associations between patient comfort with a PCP and three measures of BHS utilization. Subsequent analyses were conducted to explore the extent to which shared decision-making moderated these associations. Participants who reported an intermediate or high comfort level with a provider had higher odds of reporting that they were likely to see (aOR = 2.10 and 3.84, respectively) and get advice (aOR = 2.75 and 4.76, respectively) from a mental health professional compared to participants who reported a low comfort level. Although shared decision-making influenced participants' likelihood of seeing and getting advice from a mental health professional, it was not a statistically significant moderator in these associations. Building stronger relationships with patients may improve BHS utilization, a provider practice that is likely underutilized.

A classification system for youth outpatient behavioral health services billed to medicaid.

Rates of youth behavioral health concerns have been steadily rising. Administrative data can be used to study behavioral health service utilization among youth, but current methods that rely on identifying an associated behavioral health diagnosis or provider specialty are limited. We reviewed all procedure codes billed to Medicaid for youth in one U.S. county over a 10-year period. We identified 158 outpatient behavioral health procedure codes and classified them according to service type. This classification system can be used by health services researchers to better characterize youth behavioral health service utilization.

Increasing Access to and Utilization of Behavioral Health Care Through Integrated Primary Care.

To evaluate changes in access to and utilization of behavioral health (BH) services after the integration of psychologists into primary care clinics compared with clinics without integrated psychologists. We integrated 4 of 12 primary care clinics within our academic health system. We used the median wait time for BH services as a proxy for changes in access and defined BH utilization as the percentage of primary care visits that resulted in contact with a BH clinician within 180 days. We compared changes in access and utilization from the year before integration (September 2015 to September 2016) with the 2 years after integration (October 2016 to October 2018) within integrated clinics and between integrated and nonintegrated clinics. We used difference-in-difference analysis to test the association of study outcomes with the presence of integrated psychologists. Access and utilization were similar across all practices before integration. After integration, BH utilization increased by 143% in integrated clinics compared with 12% in nonintegrated clinics. The utilization of BH services outside of the medical home (ie, specialty BH service) decreased for integrated clinics only. In clinics with integrated psychologists, 93% of initial BH visits happened on the same day as a need was identified. The median wait time for the 7% in integrated clinics who were not seen on the same day was 11.4 days (interquartile range = 5.3-17.7) compared with 48.3 days (interquartile range = 20.4-93.6) for nonintegrated clinics. A team-based integration model increased BH utilization and access.

Linking Medicare-Medicaid Claims for Patient-Centered Outcomes Research Among Dual-Eligible Beneficiaries.

Evaluation of Medicare-Medicaid integration models' effects on patient-centered outcomes and costs requires multiple data sources and validated processes for linkage and reconciliation. To describe the opportunities and limitations of linking state-specific Medicaid and Centers for Medicare & Medicaid Services administrative claims data to measure patient-centered outcomes for North Carolina dual-eligible beneficiaries. We developed systematic processes to (1) validate the beneficiary ID linkage using sex and date of birth in a beneficiary ID crosswalk, (2) verify dates of dual enrollment, and (3) reconcile Medicare-Medicaid claims data to support the development and use of patient-centered outcomes in linked data. North Carolina Medicaid beneficiaries with full Medicaid benefits and concurrent Medicare enrollment (FBDE) between 2014 and 2017. We identified need-based subgroups based on service use and eligibility program requirements. We calculated utilization and costs for Medicaid and Medicare, matched Medicaid claims to Medicare service categories where possible, and reported outcomes by the payer. Some services were covered only by Medicaid or Medicare, including Medicaid-only covered home and community-based services (HCBS). Of 498,030 potential dual enrollees, we verified the linkage and FBDE eligibility of 425,664 (85.5%) beneficiaries, including 281,174 adults enrolled in Medicaid and Medicare fee-for-service. The most common need-based subgroups were intensive behavioral health service users (26.2%) and HCBS users (10.8%) for adults under age 65, and HCBS users (20.6%) and nursing home residents (12.4%) for adults age 65 and over. Medicaid funded 42% and 49% of spending for adults under 65 and adults 65 and older, respectively. Adults under 65 had greater behavioral health service utilization but less skilled nursing facility, HCBS, and home health utilization compared with adults 65 and older. Linkage of Medicare-Medicaid data improves understanding of patient-centered outcomes among FBDE by combining Medicare-funded acute and ambulatory services with Medicaid-funded HCBS. Using linked Medicare-Medicaid data illustrates the diverse patient experience within FBDE beneficiaries, which is key to informing patient-centered outcomes, developing and evaluating integrated Medicare and Medicaid programs, and promoting health equity.

Predicting mental and behavioral health service utilization among child welfare-involved caregivers: A machine learning approach

Caregiver substance misuse (SM) and mental illness (MI) are risk factors for child abuse and neglect and are associated with more intensive CPS involvement including increased risk of foster care placement and multiple re-reports. This study examines the prevalence of SM and MI among 929 CPS-involved caregivers during the early phases of CPS involvement and explores the extent to which family and CPS-case characteristics predict referral and service receipt. We used a machine learning approach to identify the strongest predictors of SM and MI service receipt by comparing the predictive strength of random forest and logistic regression models. Results indicate a high prevalence of self-reported need for SM (13%) and MI (34%) services among caregivers. Nearly one-quarter (23.5%) of caregivers with SM needs and 34% of caregivers with MI needs did not receive needed services. Frequent contact with CPS workers, adverse experiences in adulthood, and court involvement were strong predictors of both SM and MI service uptake. Findings suggest the need for consistent screening for SM and MI among primary caregivers at the early stages of CPS-involvement, as well as enhanced referral practices. Machine learning applications for applied social science researchers are also discussed.

Impact of COVID-19 on Behavioral Health Services Use Among Medicaid Enrollees with Chronic Behavioral Needs by Race and Ethnicity.

The COVID-19 pandemic may widen the disparities in access to behavioral health (BH) services among groups that have been historically marginalized. However, the rapid expansion of telehealth presents an opportunity to reduce these disparities. The objective was to assess the impact of COVID-19 on BH visits, including in-person and telehealth, and BH treatments by different race and ethnicity groups. This was a retrospective, observational study using administrative claim data. Two cohorts were created: a before-COVID-19 group and a during-COVID-19 group. A difference-in-differences analysis was conducted to assess the access to BH-related visits between the 2 groups by different race and ethnicity groups. The study sample included 90,268 patients aged 18 to 64 years with repeated BH diagnoses in baseline periods and continuous medical and pharmacy enrollment. During the pandemic, BH telehealth visits surged, whereas the overall utilization of BH services, mental health medication, and counseling declined among all racial groups as the BH telehealth increase did not fully compensate for the reduction of in-person visits. Latino patients had a higher likelihood of using BH telehealth visits than White patients. However, Black patients had a lower likelihood of using substance use disorder (SUD) treatment than their White counterparts. Our results also suggested that care continuation and pre-established care-seeking behaviors are associated with increasing BH visits and treatments. As policy makers and payers are expanding offerings of telehealth visits, it is imperative to do so through a health equity lens and center the needs of groups that have been economically and socially marginalized to advance equitable adoption of telehealth.

Exploring Disparities in Behavioral Health Service Use in the Early Stages of the COVID-19 Pandemic.

Previous studies examined the impact of the COVID-19 pandemic on mental health, but few studies looked at associations between the pandemic and state-funded behavioral health service utilization. We aimed to examine behavioral health service utilization during the early phase of COVID-19 among individuals with psychiatric disorders (PD), substance use disorders (SUD), and co-occurring disorders (COD). Using the last Adult Needs and Strengths Assessment (ANSA) completed in 2019 and 2020 in a Midwestern state, a column proportion test and Poisson regression model examined the associations of the pandemic year, age, gender, race/ethnicity, diagnostic type, and behavioral health needs. Between 2019 and 2020, the number of adults newly involved in behavioral health services increased from 11,882 to 17,385. The number of total actionable items (TAI) differed by gender and age group. Adults who were Black or American Indian were more likely to have a significantly higher number of needs that interfered with functioning (β = 0.08; CI [0.06, 0.09]), (β = 0.16; CI [0.08, 0.23]), respectively, than White peers. Individuals with COD showed the highest number of needs (β = 0.27; CI [0.26, 0.28]) when compared to the needs of individuals with psychiatric disorders, after controlling for year, age, gender, and race/ethnicity. Additional research is needed to better understand the intersections of age, gender identity, race/ethnicity, the complexity of needs, and useful strengths. The involvement of practitioners, service organizations, researchers, and policymakers will be required to provide accessible, effective behavioral health services with cultural and developmental adaptations to support recovery.

RWD153 Investigating Trends of Behavioral Healthcare Resource Utilization Burden in Children and Adolescents with Suicidal Symptoms in the United States: An Analysis of Electronic Health Record Data

The increasing rates of mental health disorders and suicide among children and adolescents have highlighted the importance of understanding the burden of healthcare resource utilization for this population. This study aims to investigate trends in the utilization of behavioral health services by minors with suicidal or self-harm symptoms in the United States.

Exploring the importance of predisposing, enabling, and need factors for promoting Veteran engagement in mental health therapy for post-traumatic stress: a multiple methods study

PurposeThis study explored Veteran and family member perspectives on factors that drive post-traumatic stress disorder (PTSD) therapy engagement within constructs of the Andersen model of behavioral health service utilization. Despite efforts by the Department of Veterans Affairs (VA) to increase mental health care access, the proportion of Veterans with PTSD who engage in PTSD therapy remains low. Support for therapy from family members and friends could improve Veteran therapy use.MethodsWe applied a multiple methods approach using data from VA administrative data and semi-structured individual interviews with Veterans and their support partners who applied to the VA Caregiver Support Program. We integrated findings from a machine learning analysis of quantitative data with findings from a qualitative analysis of the semi-structured interviews.ResultsIn quantitative models, Veteran medical need for health care use most influenced treatment initiation and retention. However, qualitative data suggested mental health symptoms combined with positive Veteran and support partner treatment attitudes motivated treatment engagement. Veterans indicated their motivation to seek treatment increased when family members perceived treatment to be of high value. Veterans who experienced poor continuity of VA care, group, and virtual treatment modalities expressed less care satisfaction. Prior marital therapy use emerged as a potentially new facilitator of PTSD treatment engagement that warrants more exploration.ConclusionsOur multiple methods findings represent Veteran and support partner perspectives and show that amid Veteran and organizational barriers to care, attitudes and support of family members and friends still matter. Family-oriented services and intervention could be a gateway to increase Veteran PTSD therapy engagement.

A Comprehensive Examination of Pediatric Behavioral Health Service Demand and Utilization in a Large, Academic Health System from 2019 to 2021.

Studies of the effects of COVID-19 on youth suggest a worsening in mental health globally. We performed a retrospective analysis of data from January 2019-November 2021 for all outpatient referrals, as well as outpatient, inpatient, and emergency department (ED) encounters for behavioral health (BH) reasons in children aged < 18 in a large academic health system in the United States. Mean weekly rates of outpatient psychiatry referrals, outpatient psychiatry visits, ED visits, and inpatient admissions for BH reasons were compared between pre-pandemic and pandemic periods. The average weekly rate of ambulatory referrals (8.0 ± 0.33 to 9.4 ± 0.31) and completed appointments (194.2 ± 0.72 to 213.1 ± 0.71) significantly increased during the pandemic, driven largely by teenagers. The weekly average of ED pediatric encounters for BH did not increase during the pandemic, although the percentage of all pediatric ED encounters that were for BH did increase from 2.6 to 4.1% (p < 0.001). Length of stay for pediatric BH ED patients increased from 1.59 ± 0.09 days pre-pandemic to 1.91 ± 0.11 days post-pandemic (p < 0.0001). Inpatient admissions for BH reasons overall decreased during the pandemic, due to a decrease in inpatient psychiatric bed capacity. However, the weekly percentage of inpatient hospitalizations for BH reasons that occurred on medical units increased during the pandemic (15.2% ± 2.8-24.6% ± 4.1% (p = 0.0006)). Taken together, our data suggest the COVID-19 pandemic had varying degrees of impact, depending on the setting of care.

Examining Patterns of Psychotherapy Service Utilization Among Medicaid-Enrolled Adolescents.

Adolescents with behavioral health disorders (i.e., mental health disorders and substance use) often experience frequent recurrence of symptoms, suggesting a need for an ongoing behavioral health intervention, rather than a single course of treatment. However, little is known about mental health care service use among adolescents over longer periods. The authors examined longitudinal patterns of outpatient behavioral health service utilization in a large sample of adolescents. Medicaid claims for 8,197 adolescents (ages 10.0-13.9 years, mean±SD=11.5±1.2; 61% male) from one Indiana county between 2006 and 2017 were examined, with a focus on outpatient psychotherapy visits. Latent class analysis (LCA) was used to detect clusters of longitudinal patterns of outpatient psychotherapy visits across 5 years, beginning with an adolescent's first behavioral health visit. A five-class LCA model emerged with unique classes of service use based on duration and level of engagement (frequency) of monthly outpatient psychotherapy visits. Most adolescents fell in the nonuse class (38.7% of the sample). Additional classes were defined as late-onset low engagement (17.1%), early-onset high engagement (15.5%), early-onset moderate engagement (16.7%), and continuously high engagement (11.9%). Statistically significant differences were found across the classes in average duration and frequency of involvement (p<0.001), as well as in demographic characteristics (race, age, gender, and ethnicity) and behavioral health diagnoses (p<0.001). These findings confirm that adolescents with behavioral health diagnoses do not follow a uniform pattern of psychotherapy utilization. The distinct patterns of service use point toward the need to identify appropriate long-term service recommendations for adolescents.

Latent Classes of Incarcerated People’s Behavioral Health: Differences by Childhood Adversity and Criminal History

ABSTRACT The present study identified latent classes of incarcerated people based on indicators of their substance use and mental health. Data were used from the nationally representative 2016 Survey of Prison Inmates. Two classes of incarcerated people were identified: Non-clinical class and Clinical class. Compared to the Non-clinical class who had limited history of mental illness and limited substance use, the Clinical class had an extensive history of mental illness, substance use, and behavioral health services utilization. Moreover, the Clinical class had significantly higher prevalence of childhood homelessness, foster care, and parental incarceration and increased number of prior total arrests.

15.2 Psychotropic Medication and Quality Improvement Opportunities Within a Medicaid Behavioral Health Managed Care Organization

Managed care organizations (MCOs) have the responsibility to monitor utilization of behavioral health services and psychotropic medications, along with quality and safety of care, and can use their data to develop quality improvement activities (QIAs). Community Care Behavioral Health Organization, a public behavioral health MCO (BH-MCO), developed a QIA regarding pediatric polypharmacy, because the use of concurrent psychotropics has limited evidence raising quality and safety concerns.

Comparing Pre- to Post-COVID-19 Health Disparities Between Black and White Female Connecticut Medicaid Beneficiaries in Behavioral Health Utilization

Background Broad health disparities (HDs) persist in Connecticut and the United States between minority and White patients, especially in access to and the utilization of behavioral telehealth services. Objective We aimed to determine the geographic distribution of HDs in Connecticut between Black female and White female adults in Medicaid behavioral telehealth utilization in 2019 and 2020. Methods We used the following spatial Connecticut data: (1) behavioral health utilization from Medicaid claims, from the Connecticut Department of Social Services, for the third quarters of 2019 and 2020; (2) mental health and drug and alcohol treatment facilities and ZCTA (ZIP Code Tabulation Area)-level descriptors from PolicyMap; and (3) Connecticut ZIP-to-ZCTA crosswalk data. Data were joined spatially, merged, and analyzed using spatial autoregressive models in Stata 17 (with outcome, predictors, and errors spatial lags). We computed ZCTA-level HDs comparing Black and White adult female Medicaid beneficiaries’ rates of face-to-face and telehealth behavioral services utilization. Spatial regressions were used to test spatial effects, which are extensions of classic regressions, that add neighbors’ effects to covariates. Results Distances to nearest treatment facility vary quite widely in Connecticut by ZCTAs, from 0.06 mile to 13.4 miles—3.5 miles on average. The overall White female versus Black female HDs in behavioral health care utilization were impacted by the distance to the nearest facility, such that ZCTAs farther away from the nearest facility display larger Black versus White HDs—nearly statistically significant effect in 2019 and significant effect in 2020. In 2020, in ZCTAs situated farther away from treatment facilities, both White female and Black female Medicaid patients had higher telebehavioral health utilization (spatial effects +1.3% points and +2.0% points, respectively, for 1 more mile farther away). The differential Black versus White female HDs in telebehavioral health care utilization were not impacted by distance to nearest facility, according to the total effect (direct and indirect through neighboring ZCTAs; P=.26). Conclusions Quantitative analyses indicate broad differences in Medicaid enrollment and the utilization of behavioral health services among Black and White female Medicaid recipients in Connecticut and that these differences were rather stable between 2019 and 2020. It appears that the expansion of telebehavioral health services in 2020 enhanced the access to treatment among residents who were located furthest away from providing facilities. Conflicts of Interest None declared.

Family functioning screening, referral, and behavioral health utilization in a family medicine setting.

The objective of this study was to determine the preliminary feasibility and acceptability of screening for family functioning in a family medicine setting, and secondarily to explore differences in parent-child dyad interest in behavioral health service utilization by demographics, pediatric behavioral symptoms, and pediatric QOL. The McMaster Family Assessment Device General Functioning subscale was used to assess family functioning among 58 parent-child (ages 11-26) dyads in family medicine. Feasibility and acceptability were assessed through study interest and participation and interest and attendance in behavioral health services. Associations with interest in services, Child Behavior Checklist, Pediatric QOL Inventory, and select demographics were conducted using independent samples t-tests and Mann-Whitney tests. Fifty-eight parent-child dyads participated in the study. Close to half of dyads who expressed interest in the survey completed the assessment (46%). Dyads who completed the assessment had a family functioning impairment rate of 45%, and 54% of dyads with impairment expressed interest in services. Interest in services was significantly associated with several domains of childrens' behavioral health symptoms and pediatric QOL, but no demographics. Dyads experienced high rates of clinically significant family functioning impairment, without interest in receiving colocated behavioral health services. Future research should continue to explore how family functioning assessment can be utilized to identify child behavioral health symptoms broadly and engage families in colocated behavioral health services. (PsycInfo Database Record (c) 2022 APA, all rights reserved).

Examining School Based Health Service Utilization Among Marginalized Youth in Connecticut During COVID.

PurposeThis study examines how the coronavirus disease 2019 (COVID-19) is affecting utilization of medical and behavioral health services through school based health centers (SBHC s).MethodsWe leveraged the electronic health records from one of the largest sponsors of SBHCs in the country, and tested differences in SBHC utilization with chi-square tests one year prior to the pandemic (pre-pandemic: March 2019-February 2020) compared to one year into the pandemic period (March 2020-February 2021).ResultsA significant difference in SBHC utilization was found between pre-pandemic and pandemic periods (P <.001). 63.5% of SBHC patients accessed medical services pre-pandemic compared to 51.2% during the pandemic. In contrast, 36.5% of SBHC patients accessed behavioral services pre-pandemic compared to 48.8% during the pandemic, representing a 12.3% increase in SBHC behavioral service utilization since the pandemic.ConclusionsSBHCs may serve as an invaluable means of ensuring youth, particularly those from disadvantaged communities, have access to needed behavioral health services during the current public health crisis.

Behavioral Health, Social Engagement, and Long-Term Care Services Use Among Community Older Adults: USA vs. Taiwan

As healthcare advances, older adults are living longer. While 90% of older adults prefer aging in their own homes and communities, it is important to examine key factors influencing healthy aging-in-community and community-based long-term care (LTC) services available in different countries. This symposium examines behavioral health, social engagement, and LTC services utilization among community-dwelling older adults in the USA and Taiwan. Lessons learned from older adults across countries will provide insights for tailored community-based LTC services and program development. Dr. Hou from The University of Central Florida (UCF) will highlight similarities and differences in behavioral health profiles and the topics that most interest community-dwelling older Americans participating in three aging-in-community programs in Central Florida. Dr. Wang from Case Western Reserve University will examine the impact of neighborhood social cohesion on mobility among community-dwelling older Americans aged 65 and older from the national Health and Retirement Study. Dr. Liu from National Cheng-Kung University in Taiwan will share results of healthy lifestyle on quality of life among community-dwelling older adults in southern Taiwan. Dr. Young from State University of New York at Albany will compare long-term care use among community-dwelling older adults with and without dementia in Central Taiwan. Finally, Drs. Cao and Hou from UCF will analyze home and community-based services in the USA versus Taiwan. This symposium will further discuss similarities and differences of key factors related to healthy aging-in-community, along with practical recommendations and lessons learned across countries and cultural environments to improve community-based long-term care services and programs.

Employer-Sponsored Health Centers Provide Access to Integrated Care via a Hybrid of Virtual and In-Person Visits.

Background: Since the explosion of telemedicine resulting from the SARS-CoV2 pandemic, employers have been particularly interested in virtual primary care as a novel means of expanding primary care services. The purpose of this study is to describe a model of integrated care delivered both in-person and virtually at employer-sponsored health centers nationwide. The key outcomes of this analysis were the proportion of all care delivered in-person and virtually by clinical discipline, the types of care and member satisfaction for care delivered in-person and virtually, and a description of the use of multiple clinical disciplines by the employee population.Methods: Retrospective observational study comparing health services utilization of primary care, behavioral health, and physical medicine services both in-person and virtually in employer-sponsored clinics between January 1, 2020 and June 30, 2021.Results: Of the 331,967 visits with employer-sponsored health center staff, 63% were in-person and 37% were delivered virtually. Most visits were for primary care services (59.5%), with physical medicine visits and behavioral health visits accounting for 25.1% and 15.4%, respectively. Whereas the preponderance of behavioral health visits were virtual visits (72.5%), less than a quarter (18.2%) of physical medicine visits were delivered virtually. 19.6% of patients were seen by more than two clinical disciplines and 2.6% were seen by three different disciplines. Overall, patients were highly likely to recommend the health center across both modalities (Net Promoter Score 89.1 for in-person care and 88.4 for virtual care).Discussion: The future of employer-sponsored integrated team-based care may require a hybrid approach that can lean heavily on virtual visits but requires the infrastructure necessary for in-person care.

The Effects of Integrating Behavioral Health into Pediatric Primary Care at Federally Qualified Health Centers: An All Payer Analysis

Research ObjectiveApproximately 1 in 5 US children have a mental health (MH) disorder. Children with MH disorders, particularly those that are under‐diagnosed or under‐treated, have higher rates of avoidable utilization and health care costs. Despite the availability of evidence‐based treatments for child MH conditions, there are many systemic barriers to receiving adequate MH care, especially for low‐income and racial/ethnic minority populations. There is also substantial unmet need. As such, starting in mid‐2016, three Boston‐based community health centers (CHC) began implementing TEAM UP—a complete behavioral health integration model for low‐income children. Our objective was to examine the impact of TEAM UP on rates of health care utilization in children.Study DesignOur primary data source was the 2014–2017 Massachusetts All Payer Claims Data (APCD). Our primary utilization outcomes included inpatient admissions, emergency department visits, primary care visits, other professional and outpatient visits, and use of behavioral health services (any services, intake/evaluation, psychotherapy, group therapy, psychiatric medication management, family consultation, screening, testing, other therapeutic services, family training and counseling, other outpatient services).Our unit of analysis was the person‐quarter. A difference‐in‐differences approach was used to estimate the effect of the intervention on intervention‐site patients, relative to a comparison group of similar non‐intervention site patients. Utilization outcomes were estimated using generalized estimating equations (GEE) with a negative binomial distribution and log link. For all models, outcome variable Y iq was indexed to patient i in quarter q. Independent variables included a dummy for whether a patient was attributed to an intervention site, a dummy for the pre‐ (2014q1‐2016q2) versus post‐period (2016q3‐2017q4), an interaction term between intervention status and post‐period, quarter, number of eligible member months in quarter q for patient i, a vector member‐level covariates (age, sex, payer type, clinical indicators, zip code‐level covariates), and site fixed effects, with errors clustered at the site‐level and using robust standard errors to account for repeated patient measures. All results are reported as marginal effects.Population StudiedChildren ages 3–21 who were attributed to one of three intervention site CHCs or to one of six geographically proximal non‐intervention site CHCs. This included a final sample of 325,675 person‐quarters representing 31,626 unique children, after exclusions; we excluded the first 6 months before and after implementation due to differential ramp‐up.Principal FindingsAfter 1.5 years of implementation time, TEAM UP was associated with increases in behavioral health service utilization, especially for other therapeutic MH services (difference‐in‐difference: 108.6 visits/1000 patients/quarter, 95% CI: 95.2, 122.0) and family training and counseling (difference‐in‐difference: 78.5 visits/1000 patients/quarter, 95% CI: 69.0, 88.1). Effects were greatest in Medicaid‐enrolled children. We did not observe any short‐term effects on other utilization measures.ConclusionsTEAM UP was associated with increased utilization of pediatric behavioral health services. Additional implementation time is necessary to determine if this will translate into reductions in avoidable utilization.Implications for Policy or PracticeThe TEAM UP model may hold promise in linking low‐income children to behavioral health services.Primary Funding SourceSmith Family Foundation; Klarman Family Foundation.