Introduction
Value-based payment (VBP) is designed to support collaborative partnerships between patients and physicians, improve the quality of care and reduce health care spending. To achieve these aims, VBP for primary care must support the four key functions of primary care (i.e., first-contact access, comprehensiveness, coordination and continuity), which are essential to meeting the goals of improved quality and reduced spending.1 The success of VBP is highly dependent on alignment across payers and unlikely to work if only a small subset of a practice's patient population is included. Increased investment in primary care across public and private payers using VBP models designed for primary care will contribute significantly to improving health, eliminating inequities, reducing the per capita cost of care over time and improving the well-being of the care team.
To help facilitate the transition away from fee-for-service (FFS) payment and toward VBP arrangements that sustainably support the kind of robust primary care that is essential to a high-performing health care system, the American Academy of Family Physicians (AAFP) has established a set of guiding principles to describe the ideal design for key components of VBP models for primary care.1
This paper and others in the series aim to translate the guiding principles into actionable steps that key stakeholders can use to implement VBP models that sustainably support primary care. The AAFP will provide background information, industry context and stakeholder recommendations, along with our expertise, to provide a broad understanding of the issues unique to primary care in VBP implementation.
This call-to-action brief focuses on the following principles1:
Methodologies used to determine the patients for which physicians and care teams are held accountable must prioritize existing patient-physician relationships over less reliable claims or geographic methods while ensuring physicians and primary care teams have reliable, timely information about the patients for whom they are held accountable.
Rationale for Action
Continuity of care is one of the most important aspects of primary care that distinguishes it from other medical specialties. The relationship between high-quality primary care and better patient outcomes is well documented.2-4 The evidence suggests that patients with a continuous source of primary care receive higher quality care overall, use the emergency department and are hospitalized at lower rates and have fewer unmet health-related social needs.5,6 In addition to these meaningful patient-level impacts, sustained, continuous care through physician-patient relationships benefits the broader health care system with reductions to overall health care spending and improvements to patient and physician satisfaction.
Despite the many known benefits of maintaining continuous primary care relationships, the current structure of the U.S. health care system presents many challenges for patients, purchasers and primary care organizations seeking to establish or facilitate strong, uninterrupted physician-patient connections. One of the most significant structural challenges is a fragmented insurance market with multiple payers covering different populations in different ways. In this fragmented system, a patient likely will receive coverage through several different payers throughout their lifetime, often disrupting primary care relationships. One 2016 survey estimates that almost half of insured individuals in the United States had a relationship with their primary care physician for fewer than three years and changed their health plan in the last three years.7
While the disruptive nature of a multi-payer system is particularly challenging for those with employer-sponsored insurance whose coverage is tied to employment, it also impacts many covered by Medicaid and Medicare as those public programs increasingly rely on private payers through Medicare Advantage (MA) plans or Medicaid managed care organizations (MCOs).
In this multi-payer insurance market, where access to care is tied to each payer’s unique network, patients can face challenges in maintaining consistent primary care coverage. Furthermore, care delivery organizations are neither well supported in establishing continuous, longitudinal care relationships nor are they rewarded for the many financial and quality gains that accrue from long-term physician-patient relationships based on the annual insurance renewal cycle that values short-term return on investment over long-term benefits. Due to these and other factors that impede primary care access, the number of individuals in the United States reporting a usual source of care continues to decline—dropping steadily from 84% in 2000 to 74% in 2019—with younger adults especially unlikely to report a usual source of care and even less likely to designate a particular physician within a practice.8,9
VBP models represent a promising vehicle for codifying and rewarding longitudinal care relationships. Under these models, physicians and practices are accountable for the cost and quality outcomes of their patients. Unlike FFS, which promotes reactive, limited interactions, VBP offers enhanced payments and expanded responsibility for physicians and practices who agree to assume responsibility for the care needs of patients and populations. VBP rewards care delivery organizations for providing appropriate clinical care while also managing total costs. This enables physicians and non-physician clinicians to proactively identify opportunities to prevent the escalation of care needs, offering high-touch, coordinated care and allows them to work on upstream factors to address patients’ holistic drivers of health. To facilitate this type of proactive, longitudinal care, physicians engaging in value-based contracts must know in advance which patients they will ultimately be responsible for managing.
The process of identifying physician-patient relationships to assign accountability under a VBP model is often called attribution. The Health Care Payment Learning and Action Network (HCP-LAN) defines attribution as the method of matching patients to their preferred primary care physician (or other clinician) to determine which clinician and provider group are accountable for a patient’s total care and costs.10 Just as there are many different types of VBP models—as explored in the AAFP’s Value-Based Primary Care Payment Models Call-to-Action Brief—there are numerous approaches to attributing patients for VBP.
Recognizing the importance of consistent physician-patient relationships, as well as the power of VBP to help codify and reward these relationships by assigning accountability, the industry is increasingly moving away from FFS and towards VBP arrangements. To accelerate the adoption of VBP, the Centers for Medicare & Medicaid Services (CMS) has introduced a goal to have 100% of all Medicare beneficiaries and the vast majority of Medicaid enrollees in an accountable care relationship by 2030 and plans to use its authority to drive broader industry adoption of value.11 Achieving these goals requires stakeholders from across the health care industry to come together to solidify and align around the optimal approaches for attributing patients to accountable primary care physicians. As adoption accelerates, it will be increasingly important that the public and private payers who design and administer VBP programs utilize transparent and accurate alignment methodologies that prioritize existing physician-patient relationships while also ensuring that physicians and care teams are provided reliable and timely information about the patients for whom they will be held accountable.
Establishing a system that values continuous primary care relationships—the kind of care that leads to improved health outcomes for patients and populations—requires unified action across stakeholders. Key stakeholders include purchasers (e.g., employers and/or union trusts who purchase health care on behalf of their workforce), public and private payers (e.g., Medicare and commercial health insurance companies), policymakers (e.g., lawmakers and regulators), physicians and the organizations that employ family and other primary care physicians.
Considering the potential promise and current shortcomings of methods to attribute patients to physicians for establishing accountability under value-based primary care payment, the AAFP recommends the following actions in Table 1 for various stakeholder entities.
Table 1. Actions to Establish Accountability in VBP
Current State
The desire to identify established physician-patient relationships is not a new concept in primary care. Primary care practices have long relied on the process of “empanelment”—where primary care practices identify a roster or “panel” of patients for whom physicians are accountable. This is done to achieve the desired level of access for patients and to balance workloads across physicians and other team members. Empanelment also supports equitable payment distributions and performance evaluations within the practice. Effective panel management translates into better patient access, a more satisfied care team and is considered a best practice in an efficiently managed primary care practice. Payers, especially Medicaid plans, have historically utilized “assignment” to identify the primary care physician or practice accountable for an individual to ensure that all Medicaid enrollees have access to a viable primary care source. As part of the transition to VBP, new methods of establishing physician or practice accountability for patients have emerged.
Health maintenance organizations have always required plan members to identify their primary care physician at the time of enrollment, given that the underlying payment approach relies on capitation, which requires knowing the physician and/or practice to pay for each enrolled member. The dwindling popularity of HMO coverage—representing only 12% of the employer-sponsored insurance market in 2022, down from a peak of 31% in 199612—resulted in the practice of identifying a preferred primary care physician as part of one’s insurance enrollment falling by the wayside for most patients until recently. |
“Attribution” or “alignment” are generally interchangeable terms frequently used by payers and others to describe the process of establishing accountability for payment and performance under a VBP model. For ease of discussion, “attribution” describes different approaches to attribution and alignment.
In a majority of VBP contracts, physician-patient relationships are identified using “claims-based attribution,” which are specific methods that rely on historic claims data to match patients to the primary care physician and/or non-physician clinician who provided the majority of their primary care services over a recent period of time—usually at least one year and not more than three years. Other limiting parameters are frequently applied in claims-based methods, such as limiting the place of care to an ambulatory clinic and/or the specialty of the treating physician and/or non-physician clinician to those generally considered to be primary care (family medicine, general internal medicine, pediatrics, and geriatrics are the most common). Payers and others using this approach may prioritize the operational expediency of using claims data as a proxy for patient preference over asking members directly. A less commonly used but preferred method of attribution in VBP models is asking individuals when they enroll who is their primary care physician, called “voluntary attribution.” Using this approach prioritizes patient preference.
Despite the drawbacks, most VBP models currently utilize claims-based attribution methodologies for matching patients with physicians and practices. Under these methodologies, payers use claims data to identify where patients are likely to receive care in the future based on their past utilization patterns. While relying on claims data to assign accountability is central to this approach, there is significant variation across VBP models in the timing, services and logic used to attribute patients.
One example of methodology variation is in the duration of the look-back period (usually one to three years) that is applied to claims histories to identify a patient’s primary source of care. The longer the look-back period, the more likely a plan is to capture the patients who do not visit their primary care physician annually. On the other hand, longer look-back periods can increase the risk of holding physicians accountable for patients who are no longer under their care.
There are also variations in the services and codes used to assign patients. While claims-based methodologies generally assign patients to physicians based on the plurality of prior primary care services, one methodology might rely strictly on wellness visits, another might include all evaluation and management (E/M) services and another might utilize E/M codes and prescription data. Similarly, there are variations in the types of clinicians recognized and/or prioritized in claims-based attribution methodologies. Understanding the critical role of primary care relationships for population health management, claims-based attribution methodologies tend to focus on primary care services administered by primary care physicians. Some methodologies also recognize primary care services provided by non-physician clinicians (NPCs)and/or subspecialists.
Timing is another area of variation, with some VBP models determining attribution retrospectively (i.e., at the end of the performance period based on actual utilization patterns), while others analyze prior claims to attribute patients on a prospective basis (i.e., at the start of the performance period based on anticipated utilization). Retrospective attribution prioritizes accuracy (e.g., when looking back, payers can track which patients actually received care under each physician or practice, so attribution errors are unlikely). Prospective attribution, on the other hand, prioritizes predictability, aiming to provide accountable physicians with a clearer sense of their attributed patients from the beginning.
Under VBP models, the ideal method for determining performance often uses a hybrid approach—prospective attribution with retrospective reconciliation. Under this method, physicians are notified of the patients for whom they will be taking accountability prior to the performance period, but the final attribution list is adjusted at the end of the performance period to remove patients who should have been attributed elsewhere.
While prospective attribution with retrospective reconciliation is the ideal method for evaluating cost and quality performance under VBP, determining other things—such as prospective payments—may require different attribution methods and/or timing. For example, attribution for non-FFS payment purposes (i.e., care coordination or management fees) should be effective immediately when new patients are attributed, whereas attribution for year-end performance evaluations should be limited to only those patients who have been attributed to the physician for the whole performance year.
Although claims-based attribution is currently the most common method for aligning patients with physicians for the purposes of assigning accountability under VBP, the prevalence of this approach is largely due to its operational feasibility for the payers administering value-based programs rather than its superiority over other alignment methods. Recognizing the importance of patient choice and the powerful impact of a strong relationship between patients and their primary care physicians, payers are increasingly encouraging, incentivizing or requiring patients to identify their primary care physician as part of the insurance enrollment process. Many labels describe this approach to establish accountability—voluntary alignment, voluntary attribution or self-attestation. Regardless of what it is called, asking patients to express their preference is always a good idea.
As with claims-based attribution methods, member/patient choice approaches also vary. Voluntary methods have different mechanisms used for patient selection (e.g., online portals, over the phone or physical paper forms), the timing of when patients’ preferences are collected and reflected in model performance (e.g., at the start of the performance year, quarterly updates, etc.) and the types of clinicians/entities patients can identify as their source of care.
Since it may be unrealistic to expect every patient to identify their preferred physician, voluntary alignment is often coupled with other methods of assigning patients, like claims-based attribution. In these VBP models, payers design attribution hierarchies and rules for determining when voluntary and claims-based methods are used. VBP models designed by the CMS Innovation Center—which often influence the approaches used by other payers—are increasingly emphasizing voluntary alignment by making the process easier and less confusing for patients, updating attribution lists periodically throughout the performance year based on voluntary alignment data and prioritizing voluntary alignment over other methods. Although the share of patients voluntarily aligning remains low in most VBP models, the hope is that these recent changes will increase the use of this method.
Lastly, in addition to utilizing different methods for aligning patients—whether claims-based, voluntary or other methods used by payers to assign accountability for patient outcomes—patient attribution can vary in the level at which accountability is assigned. Some attribution methodologies assign patients to individual physicians or non-physician clinicians (i.e., physician-level accountability), while others assign accountability at the practice or entity level (i.e., practice-level accountability). While physician-level attribution is ideal, it can present practical challenges—particularly as organizations increasingly adopt team-based models of care and because current payment systems are built to address practice-level attribution.
What Is Not Working
Over-reliance on retrospective claims-based attribution. To date, most VBP models largely rely on claims-based attribution methods for assigning accountability despite the many issues associated with this approach. For example, claims-based attribution methodologies have a problematic lag when reflecting accurate patient preferences because they rely on historical claims data. However, without timely reporting of patient preferences, physicians can find themselves accountable for patients of whom they are unaware. Likewise, relying on past utilization for attribution can cause practices to miss opportunities to recognize and intentionally engage patients not currently receiving care—especially within the primary care setting.
Furthermore, claims-based attribution methodologies may not always recognize care teams' important role in primary care. High-value primary care requires a team-based approach, especially as primary care practices help address patients’ expanded clinical and social needs through interventions like chronic care management programs, behavioral health integration and interventions to address health-related social needs. When multiple team members care for the same patient, attribution can quickly become distorted. While matching patients to individual physicians is important, attribution methodologies should seek to recognize how other care team members contribute to care delivery and patient outcomes. Methodologies should also account for the unique structure of the primary care organization and whether it is a large group practice, a multi-level health system or a small group/independent practice.
Although the many approaches used to establish accountability in VBP are improving (e.g., increasing prospective identification, notifying attributed patients with a retrospective review and prioritizing patient choice via voluntary mechanisms), we have yet to solve many issues, including determining accountability when VBP models overlap, accounting for the role of expanded care teams and ensuring alignment methods balance accuracy and operational feasibility while prioritizing existing physician-patient relationships above all else.
Certain populations, such as Medicaid enrollees, are especially difficult to contact and suffer from high churn, resulting in even higher chances of data lag and attribution ambiguity.
|
Lack of timely information for care teams regarding patients for whom they are accountable. Even when patients are offered the advantage of voluntary alignment, physicians do not always receive timely and actionable information regarding the patients attributed to them by different insurers. With more than 60% of family physicians reporting contracts with seven or more payers,13 collecting, aggregating and disseminating attribution data for all payers in an actionable format for physicians takes time—especially when collecting patient-reported information.
Patients are largely unaware of processes that do not involve them. Often, patients are unaware that they are being “attributed” or “assigned” to a specific physician, care team or accountable entity for the purpose of payment and/or quality measurement. Without this understanding, care management and establishing a robust plan for holistic care become more difficult. On the other hand, patients who misunderstand accountability in this context may mistakenly believe that voluntarily identifying a primary care physician could limit their choice of, and access to, health care services.
Physician employment and consolidation. With increasing consolidation and employment of physicians by large groups, accountability for specific patients does not always result in continuity of care in every case. Increased structural changes and burdens associated with this consolidation can lead to disruptions in a continuous primary care relationship that may be inaccurately reflected in claims-based attribution methods. For example, larger physician groups (and organizations that employ family physicians) lead some patients to develop more commitment to a practice than a particular physician. Likewise, the changing structure of physician employment can lead to a discontinuity of care as physicians leave or change their practice site.
What Is Working
Incremental improvements to methodologies used to establish accountability in Medicare alternative payment models (APMs). Although patient attribution in its current form is far from perfect, APMs—particularly those designed and administered by CMS—have made incremental and noteworthy improvements over time. One improvement has been the greater use of prospective attribution with retrospective reconciliation. While retrospective attribution used to be the default design across Medicare APMs, most new models align patients on a prospective basis, offering participants greater predictability while also reconciling at the end of the performance year to confirm accuracy.
Likewise, CMS has made a special effort to emphasize the importance of voluntary attribution in its newest APMs—adjusting attribution logic to prioritize patient choice and learning from beneficiary feedback to past program iterations (such as the Next Generation Accountable Care Organization (ACO) Model) to improve the way voluntary alignment is messaged to patients.
Another improvement is the addition of more frequent updates to attribution lists during the performance year, enabling physicians to more accurately understand and care for their attributed patients. For example, under the ACO Realizing Equity, Access, and Community Health Model or ACO REACH Model, attribution lists can be updated quarterly to reflect new voluntarily aligned beneficiaries.
Lastly, another example of incremental improvement is exemplified by changes to the Medicare Shared Savings Program to include primary care services administered by NPCs (in addition to physicians) in attribution logic, with algorithms increasingly recognizing other types of primary care practitioners in their priority logic and better reflecting actual accountability.14
Growing recognition of the importance of primary care as foundational to individual and population health. In addition to incremental improvements to APM designs, the growing recognition of primary care’s importance to the health of individuals and the broader health system drives many payers and purchasers to revisit their roles and processes to better identify members’ preferred sources of primary care. While not specific to assigning accountability under VBP models, these efforts are complementary and help to further establish primary care relationships, leading to improved care and outcomes.
Purchaser organizations (i.e., large employers and union trusts, such as Covered California and SEIU 775) are beginning to see the benefits of incentivizing or requiring identification of a preferred source of primary care at the point of enrollment for preferred provider organization or PPO plans—rather than only health maintenance organization or HMO plans, which have historically required a preferred source of care.15,16
Call to Action
The AAFP calls on payers, policymakers and primary care physician employers to prioritize prospective methods of establishing accountability where physicians and their care teams are informed in advance of the patients they are to be accountable for and provided with accurate and timely information regarding those patients. Establishing prospective accountability should always be accompanied by a clear and transparent approach that retrospectively adjusts for actual patient utilization.
The AAFP calls on stakeholders to prioritize patient selection whenever identifying an individual’s ongoing source of primary care, including adding a patient-verification step when historical claims data are used for attribution in VBP.
The AAFP calls on stakeholders to attribute patients to primary care physicians or non-physician clinicians participating in physician-led care teams.
Lastly, the AAFP calls on payers and policymakers to align patient-attribution guidelines across plans and patient populations, considering instances when patient-attribution methodologies may differ for payment than for quality or financial performance evaluation.
(April 2024) (September 2024 COD)