Effects of Health Care Payment Models on Physician Practice in the United States: Follow-Up Study

Short abstract

This study, sponsored by the American Medical Association (AMA), describes how alternative payment models (APMs) affect physicians, physicians' practices, and hospital systems in the United States.

Abstract

This study, sponsored by the American Medical Association (AMA), describes how alternative payment models (APMs) affect physicians, physicians' practices, and hospital systems in the United States and also provides updated data to the original 2014 study. Payment models discussed are core payment (fee for service, capitation, episode-based and bundled), supplementary payment (shared savings, pay for performance, retainer-based), and combined payment (medical homes and accountable care organizations). The effects of changes since 2014 in the Affordable Care Act (ACA) and of new alternative payment models (APMs), such as the Medicare Access and CHIP Reauthorization Act of 2015 (MACRA) Quality Payment Program (QPP), are also examined. This project uses the same qualitative multiple–case study method as the 2014 study, relying primarily on semistructured interviews with physician practice leaders, physicians, and other observers. Findings describe the challenges posed by APMs, strategies adopted to deal with APMs, the effects of rapidly changing and increasingly complex payment models, and how risk aversion influences physician practices' decisions to engage in new payment models. Project findings are intended to help guide efforts by the AMA and other stakeholders to improve current and future APMs and help physician practices succeed in them.

Purpose

In 2014, RAND and the American Medical Association (AMA) conducted a qualitative research project entitled “Effects of Health Care Payment Models on Physician Practice in the United States” that described how alternative payment models such as capitation, episode-based and bundled payment, shared savings, and pay for performance affected multiple aspects of physician practice. This study updates our 2014 study with data from interviewees in the same six markets we examined in 2014. Our interviews focused on changes since 2014, including but not limited to the effects of new alternative payment models (APMs) such as the Medicare Access and CHIP Reauthorization Act of 2015 (MACRA) Quality Payment Program (QPP). Project findings are intended to help guide efforts by the AMA and other stakeholders to improve current and future alternative payment models and help physician practices succeed in them.

Methods

To describe the evolution of payment models and their longitudinal effects on physician practices since 2014, this project used the same qualitative multiple–case study method as our 2014 study, relying primarily on semistructured interviews with physician practice leaders, frontline physicians, and other observers. Whenever possible, we sought to interview the same individuals and organizational representatives that we interviewed in 2014. We then supplemented the sample with new interviewees, aiming to collect data on payment models and practices that have emerged since 2014. As in our 2014 study, we included a relatively large number of cases (31 physician practices in six markets) because we wanted to capture a diversity of practice sizes, specialties, and ownership models. To collect updated data on market context, we also interviewed market observers from the same categories we used in 2014: leaders of health plans and hospitals operating in the market, state or county medical societies, and state Medical Group Management Association chapters. The project did not seek to describe effects of alternative payment models on patient outcomes or the costs of care.

Main Findings

Persistent Findings: Challenges Associated with Alternative Payment Models

Our 2014 study identified several key challenges facing physician practices that participated in APMs. At the practice level, alternative payment models increased the importance of data and data analysis (as well as data deficiencies and inaccuracies). At times, alternative payment models also conflicted with each other and with government regulations, complicating practices' ability to respond in a constructive manner. For individual physicians, core clinical activities were unchanged, but participation in alternative payment models had increased the volume of nonclinical activities, particularly documentation needs. Finally, our 2014 study identified problems with data integrity and timeliness, errors in payment model execution (including inaccurate measure specification and patient attribution), incomprehensible incentives, and concerns about measure validity, all of which limited the effectiveness of alternative payment models.

All the challenges described by respondents in 2014 persisted in the current study. In particular, issues related to data continued to constrain practices' ability to understand and improve their performance. Operational errors in payment models also continued to be a source of frustration for physician practices, at times with financial consequences. In some cases, these negative experiences reduced practices' future willingness to participate in alternative payment models, even when offered by different payers. Because physician practices typically participated in multiple payment models from a variety of payers, challenges related to interactions between payment models also persisted.

Persistent Findings: Physician Practice Strategies Regarding Alternative Payment Models

As in 2014, practice leaders and other stakeholders across the study markets implemented a variety of strategies in response to APMs. These included new capabilities and models of care, investments in data and analytics, and internal financial and nonfinancial incentives for individual physicians. Some of these strategies were a continuation of previous efforts (e.g., expanding the roles of care management and care coordination staff), while others involved adding new capabilities (e.g., enhancing a primary care practice's ability to provide behavioral health services), often enabled by new information technology. Practices also augmented their capabilities to collect and manage data from internal and external sources by investing in staff and information technology.

Despite engagement with new APMs, most practices reported that internal financial incentives for individual physicians had not substantially changed since 2014. Modest bonuses for quality performance remained common, and with the exception of small, independent practices (for which physician-owner incentives were inseparable from practice-level incentives), individual physician financial incentives based on costs of care were almost nonexistent. As in 2014, practice leaders deployed a range of nonfinancial strategies to influence physician decisionmaking, such as internal performance reports, that appealed to physicians' competitiveness and self-esteem.

New Findings: Accelerating Pace of Change in Payment Models

Across study markets, multiple practice leaders and market-context interviewees perceived an accelerating pace of change in payment models since 2014. This acceleration, partially driven by the MACRA QPP, capped an already fast-paced prior decade of successive payment and regulatory changes including the Physician Quality Reporting System (PQRS), the Health Information Technology for Economic and Clinical Health (HITECH) Act, the Patient Protection and Affordable Care Act (ACA), patient-centered medical homes (PCMHs), accountable care organization (ACOs), and other pay for performance (PFP), shared savings, and episode-based payment programs. The pace of change was challenging for most practices and especially so for small and independent practices. Larger organizations described partially shielding their physicians from payment changes by focusing on long-standing internal goals and by modifying their internal incentives more slowly than the external incentives received from payers.

Market observers noted that practice management consultants were sometimes unable to keep up with the pace of change, making it harder for small primary care practices to find trustworthy advice. Some practice leaders called for a “time out” from further payment changes, so that they could better understand how to respond to their current financial incentives.

Sudden or unexpected discontinuations of APMs were particularly challenging for physician practices and other market participants. Respondents in the current study described these APM reversals as resulting from transitions in state and federal political leadership rather than the performance of the models themselves. In some cases, reversals undermined practices' ability to receive a return on their investments in performance improvement. In other cases, such as unanticipated shifts from APMs back toward FFS (fee-for-service) payment, physicians described feeling frustrated by the return of volume-based incentives.

New Findings: Increasing Complexity of Payment Models

Interviewees from a wide range of practices, markets, and roles said that alternative payment models have become increasingly complex since 2014, citing an expanding number of performance measures, uncertainty concerning performance thresholds for penalties and rewards, and interactions between different payment models as sources of this complexity. The MACRA QPP was a key contributor to this complexity because it introduced new decision points for practices (e.g., the choice between the Merit-Based Incentive Payment System [MIPS] or an Advanced Alternative Payment Model [AAPM], or the choice of performance measures from a large menu). Practices of all sizes and specialties reported that understanding complex new payment models often entailed a significant resource investment, either to hire consultants or to build internal capabilities to analyze APMs. In our sample, larger practices and those affiliated with large health systems made these investments, while leaders of smaller, independent practices were more likely to express confusion and disengagement.

For practices that did invest in understanding APMs, the increased complexity of payment models presented new opportunities for financial success. Some of these practices found ways to receive more credit for their preexisting quality—without materially changing patient care—by enhancing their documentation and data abstraction practices, thoroughly coding risk adjustment diagnoses, actively managing patient attribution, or purposefully selecting their performance measures to maximize the likelihood of rewards.

New Findings: More Prominent Risk Aversion Among Physician Practices

Despite general enthusiasm for APMs that involved bonuses, interviewees in practices that spanned multiple specialties and markets reported a high degree of financial risk aversion, which influenced their decisions to engage in new payment models. Risk aversion was especially prominent among practices that had experienced losses in APMs or that were inexperienced in managing risk. These practices sought either to avoid downside risk or to off-load it to partners (e.g., hospitals and device manufacturers) whenever possible.

For smaller practices, taking on debt to finance infrastructure investments that were necessary to succeed in APMs represented another form of financial risk. These practices were attracted to APMs that offered subsidies for up-front infrastructure investments and to partners that provided such infrastructure at nominal cost, in exchange for a share of any bonuses received. Some larger practices—especially those with prior experience taking losses in APMs—renegotiated their contracts to shift more risk back to payers.

Implications

Simpler APMs Might Help Practices Focus on Improving Patient Care

The complexity of new APMs has confused some physician practice leaders, disengaged others, and induced a small set of practices to make substantial investments in understanding these APMs in detail. When practices do not understand APMs, they are unsure of whether to invest in care improvement, or how to do so in ways that will be financially rewarded or reimbursed. However, when practices have invested in understanding these APMs, they can find ways to earn bonuses and avoid penalties without necessarily changing patient care (e.g., through strategies that affect patient attribution), especially when practice leaders believe their quality is already high. Our findings suggest that the greater the complexity of APMs, the greater the potential financial return on practices' investments in understanding them. Simplifying APMs might help tip the balance back toward improving patient care as the preferred strategy for earning financial rewards.

Practices Would Benefit from a Stable, Predictable, Moderately Paced Pathway for APMs

The accelerating pace of change in APMs has exhausted not only physician practices but also the consultants advising them. Worse, unanticipated reversals in payment policies have prevented practices from recovering the costs and reaping the rewards of their substantial APM-driven investments in care improvement. A slower, more predictable pace of change in payment models seems likely to benefit practices, payers, and other stakeholders. Practices might consider negotiating longer-term contracts with payers, with built-in penalties for early unilateral termination of the model (or substantial deviation from the prespecified course of change). To insulate APMs from the vagaries of electoral politics, efforts that involve private-sector payers might prove more stable over time.

Practices Need New Capabilities and Timely Data to Succeed in APMs

As in 2014, many physician practices in our current study—especially small, independent practices—reported that they lacked the internal skills and experience necessary to perform well in APMs. Data management and analysis were seen as critical skills; yet, without timely, complete, and accurate performance data, even practices with well-developed data infrastructures were unable to assess their improvement efforts, make course corrections, or even know their positions relative to performance targets. Helping practices invest in these skill sets and supplying them with timely, understandable performance data will likely be a critical contributor to the ultimate success of many APMs. Conversely, payment models that are poorly executed (e.g., with serious, persistent operational errors) and unsupportive of physician practices could undermine future engagement in APMs and fail to improve patient care.

Reducing Practices' Access to Upside-Only APMs Risks Disengaging Them

In our current study, physician practices expressed increasing aversion to downside risk in APMs (regardless of APM complexity). With the exception of larger practices that had extensive experience in managed care, practices generally sought upside-only APMs. When faced with APMs that featured downside risk, practices took steps to off-load this risk to partners. Given this risk aversion, and given the likelihood that practices will find ways to minimize downside risk in any event, payers should carefully weigh the anticipated advantages and disadvantages of mandating APMs with downside risk. In some cases, continuing to offer upside-only APMs or finding other ways to help practices manage downside risk (e.g., subsidizing up-front investments in new practice capabilities) might improve APM uptake—especially among practices with limited experience in risk contracts.

Designing APMs to Encourage Clinical Changes That Individual Physicians See as Valuable Might Improve Their Effectiveness

As in 2014, physicians were broadly supportive of APMs that enabled their practices to make noticeable improvements in patient care. In such cases, physicians reported intrinsic satisfaction with clinical improvements, sometimes even when these improvements did not result in financial bonuses. However, when APMs' principal impact on physicians was to create new documentation and reporting burdens—or there was no perceptible improvement in patient care—physicians generally reported disengagement and skepticism that anything had improved, even when they received bonuses. Co-designing APMs with practicing physicians and other leaders of their practices might help improve physician engagement and the chances that APMs will produce real improvements in patient care.