In a world where one in five adults experiences a behavioral health condition, our payment models lack incentives for high value care and therefore our system is failing patients, providers, and payers. It's time for that to change. 

When tailored to the unique challenges of behavioral health care, value-based reimbursement models can transform outcomes, improve costs, and increase access to high quality treatment for millions of people.

The fee-for-service (FFS) model has long been the dominant paradigm in health care reimbursement, including behavioral health. Under this model, providers are paid for each service rendered, regardless of the necessity of the service or the outcome. While straightforward, this approach has significant limitations. 

The challenges of the FFS model are multifaceted and include:

• Misaligned incentives: FFS rewards volume over value, incentivizing providers to prioritize healthier patients with less complex illness who reliably present for treatment visits. 
  

• Fragmented care: Provider coordination within and outside of the health care system is not funded, leading to lack of coordination and safety concerns.
  

• No accountability for quality: Services are reimbursed regardless of quality or impact on patients’ health. 
  

These limitations have real consequences. People living with serious mental illness die 10-25 years earlier than the general population, mainly due to treatable medical conditions[1]. Moreover, the economic burden of mental illness in the United States is staggering, estimated at $225 billion annually[2].

Let’s also consider the case of untreated serious mental illness (SMI) and substance use disorders (SUD). Approximately half of adults living with SMI report an unmet need for mental health treatment [3]. These gaps in care lead to increased emergency department visits, hospitalizations, and incarceration rates, all of which drive up costs without improving outcomes. Value-based contracting (VBC) in behavioral health aims to align payment with patient outcomes, quality of care, and cost efficiency. 

The potential benefits of VBC in behavioral health are significant. For patients it promises improved outcomes, better care coordination, and enhanced overall experience. For providers, it means aligned incentives that reward quality care and reduce administrative burden. And for payers, it yields better cost management and predictability.

Early success stories for physical health VBC models provide a glimpse of the potential. For instance, the Medicare Shared Savings Program generated $1.66 billion in savings for Medicare in 2021 alone while improving quality metrics [4].

While most of us agree on the importance of driving toward value-based payment in behavioral health, the key challenges that often arise are related to quality measurement and outcome tracking, and provider and payer readiness.

Quality Measurement and Outcome Tracking

Person reported health outcome measures (PROMs) are our most robust tools as they are clinically validated, and are true health outcome measures as opposed to process measures, which focus on steps that should be followed to provide good care [5]. They also help define symptom burden and measure change over time during the course of treatment.  

Three concerns that are frequently cited for PROMs are as follows:

1. PROMs may be subjective based on patients’ experiences. My experience, however, both on the payer and provider side, is that individual subjectivity on a single PROM is balanced across a patient population. When taken in aggregate, with high quality treatment, population improvement measured by PROMs can be significant. Going forward, we also could develop a 360^o composite view of PROMs collected from the patient, caregivers, and providers. This aligns with how we as clinicians gather information on response to treatment and could provide a more balanced view of progress.

2.  There are a lot of PROMs. Patients often present with multiple co-occurring behavioral health conditions, such as substance use disorders and depression. There are specific PROMs that are clinically validated for each individual disorder, and so the number of PROMs applicable to each patient could be significant. This is addressed as follows:

• Payers and providers may develop a portfolio of PROMs, with 1-2 PROMs for each behavioral health condition. Payers and providers also define the frequency of measurement, including at baseline and follow up.
• During the course of treatment, the provider identifies the patient’s primary behavioral health condition and then uses the one PROM that corresponds to the primary condition to measure progress.
• Providers are accountable for clinically significant improvement, as it is defined by the relevant PROM. Providers are rewarded for the proportion of their total patient population that achieves improvement (usually defined as a 50% reduction in symptom burden) and remission (defined separately for each scale).

3. Providers may “cherry pick,” or select for healthier patients who perform well on PROMs and other measures. This is addressed in at least two ways: 

• Payers may apply risk adjustment to provide greater incentives for providers who treat patients with severe behavioral health conditions (defined using PROMs or ICD-10 codes) and those with evidence for a history of behavioral health crises (defined by CPT codes for ED and inpatient care). 
• Providers also can take accountability for population health outcomes and costs; this orients providers to focus on those patients with the greatest health vulnerability and likelihood of engaging in low-value and crisis-driven care.

Process and structural measures are important, but they only measure actions delivered within evidence-based treatment, as opposed to true health outcomes that matter for patients and families [6]. For behavioral health, the process measures with the most robust evidence for impacting health outcomes and costs are related to medication adherence. 

Patients living with schizophrenia who are adherent to antipsychotic medications - and patients with opioid use disorders who are adherent to medications to treat addiction - are more likely to enter remission from their behavioral health conditions and remain stable in the community, outside of hospitals, facilities, and other high cost locations of care. When implementing process measures, we should take the greatest care to choose those with the most robust evidence for driving improvement in total health and costs.

Provider Readiness for VBC

Behavioral health providers are an eclectic group. They range from solo practices to large multi-disciplinary health systems. Providers may specialize for a subset of behavioral health conditions, such as eating disorders or substance use. Providers also often focus on specific patient populations defined by age or payer type. They may offer one modality of treatment, such as therapy, medications or procedures like ECT.  Finding one value-based payment model that addresses all of these factors is challenging.

Behavioral health providers also vary in their ability to manage financial risk, with many providers being part of small practices that are operating with smaller margins and not able to accept financial accountability for patient outcomes. Additionally, many behavioral health providers were left out of HITECH and other incentive programs for EHR adoption; therefore, providers’ ability to measure and report on quality is limited [7].

Fortunately, value-based payment models exist on a continuum, with increasing accountability for population health outcomes and costs.  The spectrum is defined by the Health Care Payment Learning & Action Network (HCP-LAN) and is summarized as follows [8]:

1. Fee-for-service with link to quality and value (LAN category 2) allows for a simpler and financially safer entry into value-based payment for most providers. Payers may offer providers the infrastructure for PROMs measurement and reporting, and they may use process measures that are collected by claims to reduce providers’ administrative burden. Providers are guaranteed their FFS revenue, and they receive additional payment for improved patient outcomes.

2. Bundled payments and case rates for specific conditions (LAN category 3) allows providers to have greater flexibility in the services offered for specific behavioral health conditions, and these payments can be at up and downside risk based on outcomes.  The ASAM P-COAT model is a good example of bundled payments for SUD [9]. 

3. Population-based payment (LAN category 4) allows providers to receive a set per-member-per-month payment, adjusted for patient risk, with financial incentives that are tied to quality metrics. This model can work well for managing populations with chronic behavioral health conditions. It requires the full set of modalities of care - including all levels of care and types of treatment (therapy, medication management, case management, etc.) - either collected under one provider umbrella or achieved through clinical integration across multiple groups.
   

When offered this full spectrum of value-based payment options, then providers can enter at any stage depending on their readiness. Additionally, over time, payers can offer financial incentives and support for providers to build the infrastructure to move to greater accountability for population outcomes. Drawing on experience from the rest of health care - we will achieve the highest value from our behavioral health system when we move providers and payers toward alignment for population health and costs. We should orient our investments to helping providers along this spectrum as quickly and robustly as possible.

Payer Readiness for VBC

Value-based payment for services outside of health systems and primary care is relatively new, and payers have multiple issues to consider as follows:

1. Population definition: For VBC to drive value, we have to ensure a focus on the highest risk members. Additionally, we need to proactively identify and engage those members with high risk, as opposed to waiting for them to present for care, often in crisis. Several payers and provider groups have developed claims-based algorithms to identify patients with severe and complex behavioral health conditions. Risk stratification tools also are available, and work is underway to build models that more accurately predict rising risk. These models always need to be validated for equity.

2. Writing value-based payment models: Through the LAN framework, payers have a structure for delivering value-based payment models for behavioral health. Developing these models requires investments from payer actuarial, clinical and legal teams to define the opportunity, set goals, and write the contracts. Additionally, implementing methods for sharing risk and savings based on patient attribution status - i.e. with patients who may be attributed to primary care and other providers in other value-based payment arrangements - is a challenge for behavioral health and more broadly for any medical speciality. Aligning financial incentives across all providers and services that drive value for patients and families is key.
   

3. Insurance carve outs: High quality behavioral health care drives value for both physical and behavioral health. However, behavioral health carve out companies often are financially responsible only for the behavioral health services, and they do not benefit from value realized on the physical health side. Therefore, these companies have little incentive or funding to address the co-occurring physical health conditions that have high prevalence across our patient population, resulting in fragmentation and poor coordination of care.  Holding behavioral health carve outs accountable and allowing them to realize value delivered on the physical health side is critical for driving value holistically across the health care system. To achieve this, some physical health payers that use behavioral health carve out companies are changing their financial arrangements and performance guarantees to include physical health.
   

4. Claims systems: Most large payer claims systems are designed for FFS payment, and major investments are required to deliver adaptations to offer capitation payments, and especially those that combine payment for physical and behavioral health care. We need these investments in the infrastructure for payment in order to deliver robust value-based models. In lieu of that investment, payers may offer population-based payment models that are built on a FFS infrastructure (LAN category 3); in this way, providers continue to submit FFS claims, but the annual shared savings and risk are calculated based on the total cost of care.
   

Instead of counting services, the future of behavioral health care is about making every service count. By embracing value-based care, we will create a behavioral health system that truly serves the needs of patients, providers, and society. 

The journey is challenging and requires investment and alignment across the multiple sectors that impact patients and families. We cannot wait because the consequences are severe and are happening now - with hospital crowding due to lack of access to care in the community, rising suicide rates especially for youth and elderly individuals, and an inadequate provider workforce requiring us to use our limited resources as efficiently as possible.

The opportunities are huge – particularly for optimizing health care resources and improving lives - and they are clear and within reach.  Let’s join together and make it happen.

I look forward to discussing this topic in more detail at this year’s Behavioral Health Tech conference. I hope to see you there!

References

• [1] Fiorillo A. and Sartorious N., Annals of General Psychiatry, (2021).
• [2] Substance Abuse and Mental Health Services Administration. (2020). Behavioral Health Spending & Use Accounts 2006-2015. 
• [3] National Survey on Drug Use and Health, (2022).
• [4] Centers for Medicare & Medicaid Services. (2022). Medicare Shared Savings Program Saves Medicare More Than $1.6 Billion in 2021 and Continues to Deliver High-quality Care. 
• [5] American Psychiatric Association. (n.d.). PsychPRO PROMs Description Guide.
• [6] Agency for Healthcare Research and Quality. (n.d.). Types of Health Care Quality Measures.
• [7] National Center for Biotechnology Information. (n.d.). PMC6568002.
• [8] Health Care Payment Learning & Action Network. (n.d.). APM Framework Refresh White Paper.
• [9] American Society of Addiction Medicine & American Medical Association. (n.d.). ASAM-AMA P-COAT Final.