Financial sustainability of payment models for office-based opioid treatment in outpatient clinics

Among provider practices that offer OBOT, some have the physician conduct visits without staff assistance, while others designate a clinic staff member (often a nurse or social worker) to coordinate patient follow-up, buprenorphine prescription refills, drug testing, and prescription drug monitoring program queries as well as other functions [34]. In some models a nurse care manager conducts most of the visits, interspersed with less frequent physician visits. Coordinated care management is viewed as an efficient way for the prescribing provider to manage more patients [35], and can also facilitate connections for patients who need medical care or additional mental health services, such as counseling [27].

An additional important component is the participation in technical assistance by practices adopting OBOT, including both training sessions and provider-to-provider consultations. In several states, the technical assistance has been delivered using adaptations of Project ECHO (Extension for Community Healthcare Outcomes) [36], and the Providers Clinical Support System [37].

In a number of states, OBOT clinics can affiliate with a ‘hub and spoke’ network. As originally designed in Vermont, this approach creates ‘hubs’, typically at OTPs, for patients with more severe OUD, and ‘spokes’, which are primary care practices that offer OBOT for more stable cases of OUD as well as delivering psychosocial services [38, 39]. Patients typically receive induction and stabilization at the hub, and are then referred back to the primary care spoke for maintenance, and return to the hub if they destabilize. The hub and spoke model has also been adopted and in some cases modified elsewhere, including in Washington state [40] and California [41], sometimes using sites other than OTPs as hubs. In Vermont, spokes must be staffed by at least one nurse and one mental health counselor per 100 patients [42]. At each spoke, care is coordinated by a registered nurse clinician case manager and/or care connector (peer or behavioral health specialist). In addition, hub clinicians provide consultative services to the spokes, and are available to manage clinically complex patients or support tapering of MOUD [34].

A number of different models have emerged for delivery of OBOT, which differ in how they organize the above-listed components. Further details are provided in a recent report to the Agency for Healthcare Quality and Research [35].

For a clinic considering adopting OBOT, the best scenario (but not the usual one) is when it can actually bill for the novel services involved. For example, in Massachusetts, community health centers that are federally qualified (FQHCs) are able to bill for NCM visits at a comparable rate as for other licensed clinical providers. This opportunity creates the potential to support the NCM salary out of fee-for-service billing of OBOT services [43]. The minimum caseload needed to achieve this is further discussed below.

A more limited version of this ‘billability’ occurs at hospital-based clinics which are allowed to charge a facility fee for nurse services, even though the professional services are not billable.

If a payer enhances fees for billable services, the clinic can use that extra revenue to support its non-billable activities (cross-subsidization). In 2017, Virginia’s state Medicaid program started offering waivered office-based providers enhanced fees for MOUD treatment and care coordination services, as well as exemption from prior authorization requirements for those services. These benefits are available to waivered clinicians who apply to become “preferred OBOT” providers, as long as their site also has behavioral health clinicians [44]. By 2020 the state had more than 100 preferred OBOT providers. One could interpret the higher fees as a cross-subsidy to help the clinic to cover its cost of non-billable services, not unlike the grant funding used in Massachusetts. However, Virginia also established a monthly bundled rate ($243) specifically for SUD-specific care coordination, a more direct subsidy to nonbillable services (as discussed more in Sect. “Funding from alternative payment models”).

A third possible model is to fund the non-billable services using the margins obtained on additional billable activity, assuming that these are positive. Positive margins depend on productivity being high enough to keep the unit cost per OBOT visit below the reimbursement rate. In several of the case studies reviewed below, financial viability depended on assumptions about staff productivity. For example, in a simulation study Fried et al. used assumptions about physician hours and NCM hours provided per patient per year in each of 4 delivery models (e.g., 2.0 and 4.0 respectively in a nurse-led model)[45].

Similarly, in one health system in western North Carolina, Farrar et al. [46] report that adoption of OBOT resulted in a net increase of 1.93 medical visits per patient per month among OUD patients. Under a series of assumptions reviewed below, they concluded that treating 1 existing patient for 1 year would generate $1439, and treating 1 new patient would generate $1677. Across multiple patients, the additional revenues could be sufficient to support a NCM position.

This funding model is better suited to larger clinics, which can generate enough additional revenue to fund the new services. It is also better suited to more integrated health care systems, which may be able to capture additional revenues generated for services such as pharmacy, in addition to the OBOT revenue, to offset their costs of delivering nonbillable services. For example, the setting studied by Farrar et al. was a community-based family medicine residency clinic with integrated behavioral medicine and clinical pharmacy. More generally, OBOT may generate additional pharmacy billing, but that revenue will not be captured by a clinic that has no pharmacy (as in many traditional primary care clinics), or a clinic that has one but most patients fill their prescriptions elsewhere. Similar considerations apply to other types of revenue generated outside the clinic. If the clinic is part of a hospital system, the question becomes whether system administrators recognize the additional revenue that the OBOT clinic is generating for it, making them more willing to sustain that investment. Without that recognition, system administrators will likely view the OBOT clinic purely as a cost center and be correspondingly less willing to invest. For example, one clinic computed how many patients used the hospital’s pharmacy on the same day they had an OBOT visit, as a measure indicating possible revenue gains from OBOT.

An additional concern would apply if some of the additional visits generated are delivered via telemedicine, as is common during the current Covid-19 pandemic. Some insurers pay less for telemedicine than for in-person visits, which would reduce the revenue gained from any extra visits.

Some states fund nonbillable OBOT services through grants they receive from SAMHSA under the State Targeted Response (STR) and State Opioid Response (SOR) programs. Another growing resource for states are Section 1115 demonstration Medicaid waivers of Institution for Mental Disorder (IMD) exclusion from the Centers for Medicare and Medicaid Services, which are intended to improve care for substance use disorder (SUD), serious mental illness and severe emotional disturbance[47]. With these waivers, CMS allows states to bill Medicaid for care in previously ineligible inpatient and residential settings, enabling states to use some of the funding thus freed up from the federal match to expand community-based treatment. CMS has made state investment in evidence-based care for SUD and co-occurring disorders, including expansion of MOUD, a condition of obtaining such waivers. So far, 32 states have Medicaid 1115 waivers for IMD exclusion for SUD treatment approved, and 3 more have waivers pending [48].

Some health centers have been able to tap external (non-fee-for-service) funding to pay for nonbillable services, such as the NCM staff position. In Massachusetts, the state-initiated rollout of OBOT at community health centers included state block grants to cover nonbillable services. These included the salary of a full-time registered nurse, staff training, and technical support on implementation of protocols for each program. More recently the state switched to funding those nonbillable services with a ‘unit rate’. In most other states, clinics do not have access to such funding. However, some clinics have been able to access funds from the Health Resources and Services Administration (HRSA). Similarly in Vermont, for every 100 Medicaid patients receiving MOUD at a ‘spoke’ (OBOT) practice, the state Department of Health Access pays the cost of one nurse and one licensed mental health/addiction counselor to support the prescribing providers [49].

Expert consultation to OBOT providers, being similarly nonbillable, has also been implemented using grant funding from SAMHSA to states (as in Massachusetts and California [50]), or from universities and foundations (e.g., in North Carolina [51] and West Virginia [52]). In Massachusetts, a network of expert faculty is paid a retainer by the Medicaid carve-out to take calls from primary care providers regarding MOUD and other opioid prescribing. Again, this is supported from Medicaid 1115 waiver funding.

Another approach to funding OBOT would be for payers to provide a preset amount to spend on treating OUD, which is not conditioned on the number of specific services provided. This approach is most often used by managed care payers.

One example of this type of ‘alternative payment model’ is bundled payment (a preset amount per treatment episode). Bundled payment is used by some state Medicaid programs (e.g., Rhode Island) to reimburse OBOT [34], and similar payment models for OBOT have been developed by the American Society for Addiction Medicine [53], and by the Innovation Accelerator Project at the Centers for Medicare and Medicaid Services (CMS) [54]. A major advance occurred in January 2020, when Medicare introduced a new monthly bundled payment option for OBOT, supported by three new procedure billing codes. Code G2086 covers “office-based treatment for OUD, including development of a treatment plan, care coordination, individual therapy and group therapy and counseling; at least 70 min in the first calendar month.” Code G2087 covers at least 60 min in a subsequent calendar month, for the same services (except for development of a treatment plan). Code G2088 covers the same services as G2087 for each additional 30 min beyond the first 120 min. Notably, these codes do not cover payment for the OUD medications or for related toxicology testing, which continue to be billed separately. The codes are also not usable for patients concurrently seen at an OTP, as CMS set up a separate bundled payment rate for OTPs. Interestingly, the codes are also usable for visits delivered as telemedicine. Responding to earlier comments on its proposed rule, CMS said it would consider later refining the codes to adjust for different levels of patient need [55]. In addition, various states are reportedly considering applying a Medicare-type bundled payment approach in their Medicaid programs.

Another alternative payment model is capitation, where the provider is paid a preset amount per plan member per month, regardless of actual service use. Several states (including Maine, Maryland, and Vermont) have used funding from Medicaid’s health home program to encourage providers to set up ‘opioid health homes’ paid on a capitated basis, as a way to cover non-billable services such as care management and care coordination. Maine’s Medicaid program pays opioid health homes capitation amounts that range from around $400 to over $2200 per member per month, depending on the level of care a patient needs and whether the health home also provides coordinated case management [56]. A key component of this model is who is identified as a health home member, with some states (e.g., Maine) asking patients with OUD to opt in, while others auto-enroll them with an option to opt out [56].

If a clinic’s only revenue source is capitation, the distinction between billable and non-billable services disappears, so the clinic has greater flexibility in its staffing decisions. Novel services like NCM are then on an equal footing with traditional services, and the clinic’s decision can focus instead on whether or not hiring an NCM will reduce its other costs. And similarly for physician consultation and other services that were non-billable under a fee-for-service model. Of course, if some payers continue to pay fee-for-service, the clinic cannot so readily abandon the billable/non-billable distinction. Capitation offers the provider greater flexibility than bundled payment.

From the clinic’s perspective, a key question about funding under alternative-payment models is the adequacy of the preset payment amount, and what level of service provision it assumes. For example, in setting the payment rates for its Medicare OBOT bundled payments, CMS assumed an average of two individual psychotherapy sessions per month and four group therapy sessions per month, while recognizing there will be variability from patient to patient and over time. Similarly, state Medicaid programs use prospective (daily or monthly) rates to pay both FQHCs and CCBHCs for OBOT, and the rates are reportedly sometimes more generous for CCBHCs. In the case of capitation, revenue is tied to enrollment, and clinics face lost revenue on patients who are periodically disenrolled from Medicaid.

The ‘Massachusetts model’ was developed in order to expand Boston Medical Center’s OBOT Collaborative Care Model to community health centers (CHCs), from its initial base at an academic medical center. The model involves hiring nurse care managers (NCMs) at each CHC to provide waivered providers with clinical support to manage patients with OUD on buprenorphine [43]. Among other roles, the NCM (a) performs the initial screening, for patients seeking OBOT treatment, after which the patient sees the waivered provider; (b) assesses the patient for withdrawal symptoms following a protocol; and (c) supports the patient through the induction process under the orders of the waivered provider. Initially, patients are required to see the NCM weekly for follow-up visits with drug screening and verification of behavioral health counseling, although this frequency can decrease depending on progress [43].

An analysis of the Massachusetts FQHC OBOT program concluded that having a NCM provide OBOT is sustainable and viable in a Federally Qualified Community Health Center, at least in that state [43]. The analysis found that it would take approximately 40 cases per year, at 27 visits per patient per year to fund a full-time NCM position, after adjusting for efficiency and administrative cost. A typical NCM’s caseload was initially 100 patients, more than twice the number of cases required [43]. The results depend on assuming 90% utilization for a caseload with varying frequency of visits. The authors noted that even in Massachusetts, not all CHCs could use that funding model, because only FQHCs were paid prospectively by Medicaid and were able to bill individual medical visits delivered by nurses at the same rate as those delivered by physicians [43]. Thus, this analysis would not apply in the many states and contexts where NCM services are non-billable.

A later report noted that 7 of the Massachusetts FQHCs were able to expand beyond grant funding and add an additional NCM, which appeared to support the earlier conclusions about financial sustainability [57].

In the North Carolina study, Farrar et al. concluded that the additional billable revenue generated by delivery of OBOT had created positions for 2 licensed clinical addiction specialists and 1 peer support specialist [46]. They estimated that delivering OBOT to 100 existing OUD patients increased the clinic’s revenue by $143,861, through increased delivery of billable medical visits. For 200 patients, the revenue would be double that amount. Their model included assumptions that (a) all visits were reimbursed using the North Carolina Medicaid Physician Fee Schedule (presumably lower rates than other payers); and (b) the costs to be covered included 20% ‘cost of business’, e.g., central business office, facilities etc. A weakness of the design was lack of a control group, implying that some of the increase in visits that they observed could have resulted from influences other than delivery of OBOT, e.g., the patients’ course of disease, or policy changes besides OBOT [46].

In 2017, an OBAT program was launched at Hennepin County Medical Center, the largest safety-net hospital in Minnesota, led by one of the authors (GB). This clinic is mostly paid fee-for-service, most patients are uninsured or publicly insured, and NCM services are non-billable, although payers reimburse a facility fee for nurse services provided in hospital-based ambulatory care clinics. In the planning stages, medical center staff analyzed the financial impact of investment in an OBOT clinic. They projected that in the first year, operating surplus would not cover the overheads, but thereafter it would exceed overhead, resulting in a net surplus. This improvement did occur. In the model, financial improvement resulted from an assumed growth in physician productivity over time, as the program ramped up. (Beverlee Shellum, personal communication).

Key assumptions in this Minnesota modeling included the number of annual visits each physician could provide (840 for new patients and 2100, later 5040, for established patients); how many annual visits each nurse could provide (1932); the facility fees for physician and nurse services; the percent of billed charges actually collected (37.5%); and the overhead charged by the parent entity (about 8% on top of actual expenses). Various of these assumptions could be adapted by other organizations considering OBOT adoption, if they conduct similar analyses.

Fried et al. recently published a study simulating the effect of OBOT adoption that was based on the experience of 20 primary care practice leaders engaged in buprenorphine prescribing in the United States [45]. They considered the effects of 4 model types (with varying division of labor among clinicians) in 4 different settings (varying by rurality, poverty and whether a CHC). For many parameters they used averages of values supplied by the study respondents.

Fried et al. concluded that any of the 4 approaches they considered would be expected to show an overall increase in net patient revenue, relative to the status quo of referring patients to other providers for addiction care. However, providers would need to maintain at least 9 patients in treatment and no-show rates of < 34%. The savings are expected to be larger with a nurse-care manager model than with physician-led approaches. This is probably because the NCM model achieves higher productivity, reducing the unit cost per visit relative to the reimbursement rate.

The preceding case studies can also be classified according to the various funding approaches presented in Sect. “Funding the new services needed for OBOT”. The Massachusetts example used grant funding and the billability of NCM visits. Lacking that advantage, the North Carolina and Minnesota clinics appear to have succeeded by generating additional billable services at a low enough unit cost to allow them to cross-subsidize nonbillable services. In the Fried study, models with more reliance on nurse care had more billable services to cover, but also looked more financially viable. The Minnesota clinic was the only one of these examples affiliated with a hub-and-spoke model (it was a hub).