Financing Early Psychosis Intervention Programs: Provider Organization Perspectives

We used purposeful sampling to recruit professionals (program directors, team leaders, and other administrators) from CSC provider organizations across the United States. We initially sent a recruitment e-mail in June 2019 to recipients of a national e-mail Listserv (from the Prodrome and Early Psychosis Program Network) serving the CSC provider community. We subsequently supplemented such recruitment with snowball sampling on the basis of recommendations from the informants. Participants provided oral consent before beginning the interview. Each CSC program participating in the study received a $50 American Express gift card. This study was conducted in accordance with a protocol approved by the institutional review board of Weill Cornell Medicine.

We developed a semistructured interview guide to cover three major domains: current financing approaches adopted by the CSC program, perceived sustainability of the current approaches, and perceived ideal payment models. Open-ended questions allowed informants to elaborate on issues they considered most salient. Additionally, we collected information on each CSC program, including organizational affiliation, client eligibility criteria, referral sources, team caseload, and CSC team staffing. The interview guide is available in an online supplement to this article.

Two investigators (Y.B. and M.A.P.) conducted the interviews via Zoom teleconferencing software between July 2019 and February 2020. Interviews typically were held with the program director or team lead of the CSC program and occasionally with an additional staff member. Three of the interviews were with individuals who had spearheaded CSC implementation in their state and who now had consulting and other leadership roles. Two informants held dual roles as program directors at their own site and trainers or coordinators for CSC implementation at multiple sites in their state. Interviews lasted on average 1 hour. The interviews were audio recorded and were transcribed verbatim by one investigator (M.A.P.). Notes taken during and immediately after the interview were incorporated into the data analysis (10).

The qualitative data were analyzed by using an integrated approach, which included the principles of inductive reasoning and grounded theory (11) and a deductive framework for code organization (12). In this approach, researchers begin with and iteratively revise an organizing framework, attributing labels (codes) to salient text within transcripts and documenting recurring themes that emerge. In following accepted, qualitative data-coding approaches (13, 14), two investigators (Y.B. and M.A.P.) independently coded each transcript and met to discuss the code definitions and negotiate a consensus. A third investigator (R.L.) independently coded two transcripts and participated in coding meetings to provide an outside perspective and to help in reaching consensus. The codebook and themes were iteratively developed throughout this process. (The final codebook is available from the authors on request.) The qualitative analysis software NVivo, version 12, was used to facilitate text and code management and development of themes.

All 19 programs reported a mix of funding sources. Eight (42%) of the programs reported insurance billing as their primary source of funding, supplemented by MHBG, other federal and state funding, or research funding. Of these eight programs, four had robust Medicaid fee-for-service (FFS) coverage for SEE, peer specialist, or care or case management services (in addition to the clinical services typically covered by Medicaid). These four programs also billed private insurance, which provided much more limited coverage of CSC activities, especially for services delivered by nonclinical staff or in the community. Three programs received bundled payments from Medicaid. One of these three programs received a cost-based daily rate through participation in the Certified Community Behavioral Health Clinics demonstration program that comprehensively covers CSC costs. The daily rate was determined by dividing actual annual CSC program operation costs by the total volume of visits to the program. The other two programs received a cost-based, monthly bundled payment from the local Medicaid-managed care organization, which used the “in lieu of services” provision (i.e., that CSC, because of its medical necessity and cost effectiveness, is provided in lieu of other covered services) (16). For both of these programs, actuarial studies of costs of bundled CSC services provided the basis for the monthly rate, although the bundled payment received by one of the programs did not cover SEE services. One of the 19 programs (5%) derived its revenue entirely from billable services to Medicaid and private insurance; the program was not eligible for MHBG, according to rules of its state, and had to refer patients to a nearby community mental health center for CSC services that were not reimbursable. For 10 of the 19 programs (53%), MHBG, state funding, and other federal funding (e.g., grants from the Substance Abuse and Mental Health Services Administration [SAMHSA]) were primary funding sources. Of these 10 programs, three did not bill insurance (public or private) at all. Abundance of non-MHBG state funding varied substantially across states, ranging from no funding to 4–5 times the amount of the MHBG set-aside. When distributing funding to CSC programs, states typically combined MHBG and any state funding. However, how CSC programs could tap into funding varied substantially, ranging from a lump-sum grant based on estimated population needs in the program’s catchment area, to “net-deficit” funding where state grant funds supplemented program costs not covered by insurance billing, to staff time–based billing to the state for services than could not be reimbursed by other sources.

Existing payment mechanisms involving insurance billing typically do not cover several essential and integral CSC elements, such as SEE services, peer services, and care or case management. Team meetings, supervision, and community outreach activities are CSC linchpins, but because they do not directly involve enrolled clients, they are not covered by insurance. As a result, teams that were ineligible for MHBG or state funding, or whose state contract required service-by-service billing, heavily staffed their teams with licensed professionals eligible to bill. They also sometimes chose not to provide the uncovered services and instead referred clients to partner organizations that had the means to provide them, leading to potential care fragmentation. In addition, several teams we interviewed expressed concerns that the limited insurance coverage of CSC services encouraged insured clients and families to pick and choose what services they wanted to receive and to refuse services that were not covered or were inadequately reimbursed by their insurance, thus compromising fidelity to CSC.

For covered services (e.g., medication management and individual and group therapies), the standard insurance billing rate was deemed inadequate for the intensity of services required for CSC. Several informants expressed concern that, if they had to rely on insurance billing, they could not deliver the program effectively. In addition to forgoing critical services like SEE, they would have to increase provider caseload to meet productivity targets. One CSC program was grateful for receiving enhanced FFS rates from a managed care plan in recognition of the high intensity of the services provided.

The restrictive conditions associated with FFS insurance billing were further misaligned with CSC goals. With a few exceptions in Medicaid, FFS coverage applies to in-clinic service delivery only. Several informants emphasized how meeting the needs of CSC clients required flexible and innovative service delivery, for example, by meeting at a coffee shop in the community or by addressing a crisis over the phone. However, by doing so, CSC programs risked not being reimbursed for such services, including travel time. Another prominent issue was reimbursement for services involving multiple providers (e.g., a therapy session where multiple therapists are present or receipt of services from multiple providers during a single client visit to the clinic). Because of restrictive payment rules in such circumstances, several provider teams reported having to devise strategies to maximize the billing yield.

Finally, FFS insurance billing consumed substantial staff time and resources. CSC programs engaged in insurance billing expressed frustration with claim submission, rejection, resubmission, and required paperwork. One informant indicated that, although insurance was estimated to cover 40% of the operating costs of their program, they could recover only about 10%. A staff member from another CSC program estimated that they “bring in about 30 cents on the dollar of what’s billed out.” The high cost of billing and low capture rate is especially unsustainable for small, independent provider organizations that neither benefit from the billing infrastructure of large organizations nor receive intraorganization subsidies.

Overall, CSC programs in our study considered MHBG and state funding relatively stable funding sources, especially if the funding was legislatively earmarked and if a few years remained in the legislative cycle. However, several CSC informants expressed concern about the grant nature of these sources and the inherent unsustainability, as in the following quote: “I have a fear every day as we get closer to June that if the grant funding goes away, we simply are not sustainable.”

When asked what would be an ideal way of paying for CSC, informants from 15 of the 19 programs (79%) endorsed the idea of a cost-based bundled payment and used terms such as “program fee,” “per diem,” “per month,” and “capitated arrangements.” The informants believed that such a payment model would shift the focus from productivity (in terms of billable encounters) to holistic problem-solving and innovation, with the explicit goal of improving client engagement, outcomes, and experience. A staff member from one program already receiving a Medicaid bundled payment for CSC said that the bundled payment “shifts the focus . . . to the actual participants and what their experience is” with respect to the services that can be delivered under a bundled-payment model.

In addition to a bundled payment, the vast majority of informants also endorsed holding CSC programs accountable for client outcomes in the form of a value-based payment. There was high consensus opinion regarding the types of client outcomes to incentivize, including engagement in employment or school and stable housing and avoidance of criminal justice involvement and hospitalization or emergency department visits for behavioral health issues. Although none of the 19 programs had their funding or payment explicitly tied to client outcomes at the time of the interview, nearly half (N=9, 47%) routinely collected client outcomes data and reported them regularly. Staff from several CSC programs expressed concern, however, about having hospitalization or emergency department visits as a performance measure, citing perverse incentives for denying clients hospitalization when they need it.