A Continuum-Based Framework as a Practice Assessment Tool for Integration of General Health in Behavioral Health Care

We performed a literature review of peer-reviewed articles published from 2001 to 2020 by searching the PubMed database for the following combined keywords: integrated general care and collaborative care and behavioral health clinics. This search yielded 60 articles that were relevant to GHI in behavioral health settings. We then identified 14 articles that either had an experimental design or were designed as large-scale quality improvement projects. These articles significantly informed the development of the GHI framework domain and subdomain stages (see the online supplement for a table that indicates the literature relevant to each subdomain).

After developing a draft framework, we recruited and engaged with 12 integration policy and practice experts for a diverse range of perspectives. These experts included five behavioral health clinician executives from New York State (NYS) with integration expertise, five policy and regulatory leaders with national and local expertise, and two behavioral health executives from NYS health plans. Each informant participated in a semistructured interview providing feedback on the framework domains and on current challenges and opportunities for integration. We also sought feedback on the early version from 12 psychiatric fellows working in diverse behavioral health clinics. A revised framework was then presented to 31 stakeholders at an in-person meeting (see the online supplement for a full list of stakeholders), including the 12 expert informants, behavioral health and primary care providers, consumer advocates, health plan representatives, and NYS regulators and policy makers. Consensus was achieved on the structure, content, and potential utility of the framework.

Eight domains with 15 subdomains are targeted by the GHI framework. Each domain and subdomain has a specific set of evidence-based interventions that are categorized into preliminary, intermediate 1, intermediate 2, and advanced stages of the domains specified (see online supplement). The major domains include screening and referral to care and evidence-based care for general medical conditions, ongoing care management, self-management supports, multidisciplinary teams, quality improvement, linkages to social service supports, and sustainability. The corresponding subdomains were screening, referral facilitation, evidence-based guidelines for preventive measures, evidence-based guidelines for general medical conditions, medication management, trauma-informed care, longitudinal clinical monitoring and engagement, tools to promote patient activation and recovery, care team, sharing treatment information, integrated care team training, use of quality metrics, linkages to housing and entitlement services, billing and outcome reporting, and regulatory supports. Using a strengths-based approach, we described preliminary stages as usual care, representing no or limited integration activity for the specified domain. Two intermediate stages were described to allow improved ability for sites to choose a lower stage to begin implementation from a preliminary stage (see online supplement for a list of key supporting evidence per domain). Advanced stages were those that described population-level interventions or required significantly more intensive resources than what most freestanding community behavioral clinics typically acquire. Some subdomains, such as trauma-informed care and sustainability, were not well studied for GHI, but stakeholders judged them to be important integration domains. The framework was designed to enable clinics to perform a baseline, team-based practice assessment and give them the flexibility to choose domains for focused advancement within a specific time frame (6–12 months) according to a clinic’s strategic priorities and available resources (15).

Conveying intentional and incremental progression of greater integration, the framework allowed clinics to initially identify their integration status for each domain (see online supplement). For example, a clinic was providing “usual care” if such care was most accurately described by the preliminary stages of a domain. Providing interventions or having workflows at the intermediate or advanced stage reflected a greater level of integrated care for a specific domain. Overall, the framework provided a roadmap for clinics to progressively invest in integration after an assessment of local needs and resources, such as, for example, investments in general medical health screening or training or developing collaborative partnerships with primary care.

Eleven licensed behavioral health clinics in New York City were recruited as a convenience sample through referrals from the stakeholders. Recruitment criteria included executive-level commitment to advance GHI, some recent or current experience in implementing GHI, and receiving most revenues through Medicaid and Medicare. All 11 clinics agreed to use the new framework as a practice assessment.

First, the project team provided a webinar to introduce the framework, describe the practice assessment survey, and define the rating scale for the stages within the domains. An online survey was sent to participating clinics to assess their status of general health integration based on the evidence-based framework for GHI (see online supplement). The survey questions were divided into three sections: baseline clinic characteristics, practice assessment to identify the stage within each domain or subdomain that was consistently performed at least 70% of the time (this benchmark was used in our previous framework so that practices could have confidence that a practice component was occurring much of the time even if it was not at a universal benchmark level [>90%]), and feedback on the experience of using the framework (16). Clinics were instructed to assemble a multidisciplinary team, comprising at least a senior behavioral health clinician, a team member providing general health support or education services, and an administrator or quality improvement specialist who was familiar with GHI workflows at the practice. The team assessed the current state of integration for all domains and 15 subdomains. The surveyed clinics also provided a user perspective on the logic and utility of the framework. All 11 clinics completed the survey, and their data were aggregated and deidentified for analysis. At least one technical assistance call was held with each participating clinic to help them confirm the stage that best described their integration interventions in certain domains. The project was reviewed by the Albert Einstein College of Medicine Institutional Review Board and was determined to be in the exempt category.

The 11 behavioral health clinics each served between 300 and 1,300 patients annually, predominantly insured by Medicaid managed care or Medicare fee for service. In total, 7,143 patients were served by all clinics. The clinics served significant numbers of patients belonging to racial-minority groups, with 45.0% (N=3,214) being Black or African American and 28.0% (N=2,000) Hispanic or Latinx. All clinics identified GHI support from nurse practitioners or registered nurses, with only one clinic reporting a formal collaboration with a primary care physician for consultation. Six clinics reported the use of certified peer counselors to support patient self-management for general health care. Clinics reported some experience with GHI initiatives, including use of electronic health records to extract data such as blood pressure readings and hemoglobin A1c.

All the practices reported that this was the first time a framework was used to assess their state of GHI. As expected, the current state of GHI among participating clinics varied by domain and subdomain (Figure 1). Most clinics had strengths (defined here as reporting subdomain stages at the intermediate-2 or advanced stage) in trauma-informed care, self-management supports, quality improvement, and social services linkages. Notably, all clinics reported being in the intermediate-2 and advanced stages for trauma-informed care and social service linkages. The activities related to trauma-informed care that most clinics reported as being at the intermediate-2 stage were routine staff education on trauma-informed care models and limited use of validated screening measures for trauma when indicated. Activities related to social linkages that almost all clinics were undertaking at the advanced stage included use of detailed psychosocial assessments that incorporated a broad range of social determinants of health needs to link patients to social service organizations and resources for improving appointment adherence (e.g., child care and transportation) and enhancing access to healthy food sources.

Opportunities for improvement (defined here as reporting domain or subdomain activities at the preliminary or intermediate-1 stage) were noted in 11 subdomains (see online supplement for a figure showing variation in the state of GHI among the participating clinics). For most of these subdomains, many of the clinics were mostly performing at the intermediate-1 stage. Notably, the subdomains of referral to care and billing sustainability were especially challenging, given that four sites in each category were performing at the preliminary stage. For example, routine referrals to primary care providers had no or only limited patient follow-up. For billing, these sites reported no or minimal attempts to bill for immunizations, screening, and treatment activities and were mainly supported by grants or other nonreimbursable sources.

Overall, the behavioral health clinics reported positive experiences using the framework. Ten clinics rated the framework as “easy to use.” Seven clinics could readily understand the meanings of the domains, subdomains, and stages within the GHI continuum and how it could assist in planning for stepwise implementation. Occasionally, clinics reported difficulties in deciding which stages to choose within a domain when they perceived that their processes fell between stages. In such cases, we asked them to conservatively choose the less advanced stage. Clarity was also desired on the expected types of screenings and treatment of general health conditions. Sustained funding, electronic health record capability, financial incentives, implementation support, and technical assistance were identified as key supports for advancing the level of integration. The most commonly cited barrier to advancing GHI was inadequate reimbursement for integration activities. Clinics reported that GHI staffing supports such as primary care providers (onsite or through telehealth) and nurses were needed to advance integration efforts and that many staff lacked GHI training or could not expand their scope of services without additional reimbursements or regulatory approval.