Effects of Health Care Payment Models on Physician Practice in the United States

Abstract

The project reported here, sponsored by the American Medical Association (AMA), aimed to describe the effects that alternative health care payment models (i.e., models other than fee-for-service payment) have on physicians and physician practices in the United States. These payment models included capitation, episode-based and bundled payment, shared savings, pay for performance, and retainer-based practice. Accountable care organizations and medical homes, which are two recently expanding practice and organizational models that frequently participate in one or more of these alternative payment models, were also included. Project findings are intended to help guide efforts by the AMA and other stakeholders to make improvements to current and future alternative payment programs and help physician practices succeed in these new payment models—i.e., to help practices simultaneously improve patient care, preserve or enhance physician professional satisfaction, satisfy multiple external stakeholders, and maintain economic viability as businesses. The article provides both findings and recommendations.

Purpose

The project reported here, sponsored by the American Medical Association (AMA), aimed to describe the effects that alternative health care payment models (i.e., models other than fee-for-service payment) have on physicians and physician practices in the United States. These payment models included capitation, episode-based and bundled payment, shared savings, pay for performance (PFP), and retainer-based practice. Accountable care organizations and medical homes, which are two recently expanding practice and organizational models that are based on one or more of these alternative payment models, were also included. Project findings are intended to help guide efforts by the AMA and other stakeholders to make improvements to current and future alternative payment programs and help physician practices succeed in these new payment models—i.e., to help practices simultaneously improve patient care, preserve or enhance physician professional satisfaction, satisfy multiple external stakeholders, and maintain economic viability as businesses.

Methods

To describe the effects that payment programs have on physician practices, this project employed qualitative methods, incorporating multiple case studies, with each of 34 physician practices constituting a “case.” Because the project sought to incorporate contextual information on market-level characteristics that might affect how practices respond to alternative payment models (e.g., the mix of competitors, health plans, and payment programs operating in the geographic area served by each practice), these 34 cases were nested within six geographically defined health care markets in the United States. Thus, within each market, we gathered data from physician practices and other market participants and observers: leaders of health plans and hospitals operating in the market, state or county medical societies, and state Medical Group Management Association (MGMA) chapters.

Main Findings

Effects of Payment Models at the Organizational Level

Changes in Organizational Structure

Multiple practice leaders and market interviewees reported that their own practices or others in their markets were changing their organizational models—predominantly by affiliating or merging with other physician practices or aligning with or becoming owned by hospitals—in response to new payment models. From the practice leader perspective, the most prominent payment model–related reasons for these mergers were to enhance practices' ability to make the capital investments required to succeed in certain alternative payment models (especially investments in computers and data infrastructure), to negotiate contracts with health plans (including which performance measures and targets would be included), and to gain a sense of “safety in numbers.” Leaders and physicians in multiple practices described uncertainty about how they would fare in alternative payment programs (and how such programs might evolve over time). For some of these practices, joining with a larger organization was seen as providing a general sense of security, no matter what payment programs might be introduced.

The reported effects of alternative payment models on practice stability, including overall financial impact, ranged from neutral to positive. No practice in our sample indicated experiencing major financial hardship as a result of new payment models. The retainer-based practices in our sample were small, and their physician-owners described their conversions to retainer-based payment as enabling an escape from market pressures that might otherwise have led to merger with other practices or early retirement.

Changes in Practice Operations

Respondents to our study perceived that alternative payment models have encouraged the development of team approaches to care management, featuring prominent roles for allied health professionals. In primary care practices in particular, physicians and practice leaders described appreciating how medical home programs and shared savings models (based on virtual global capitation) had allowed them to fund care manager positions. These dedicated care managers, who were allied health professionals in all cases in our sample, could concentrate on patient management between office visits, alleviating some of the pressure that physicians would feel if these activities were added to already-packed visits.

Alternative payment models that incentivize containment of total costs of care also increased the importance of offering expanded options for patients to access care from physician practices. Two examples of such expanded access were communication options for after-hours care (via web portal or telephone) and provision of in-person care in the community, outside the office.

Because global capitation and related shared savings models focus predominantly on primary care services for patient attribution and performance measurement, market observers and physician practices reported that these alternative payment models were changing relationships between primary care and subspecialist physicians. In some cases, these changes were collaborative, with multispecialty teams working to prevent progression of disease, without necessarily changing referral patterns. In other cases, changes in referral relationships were prominent, especially when alternative payment models led practices to reduce “leakage” to subspecialists in other organizations.

Increased Importance of Data and Data Analysis

In response to alternative payment models, physician practices reported making significant investments in their data management capabilities, ranging from adopting or upgrading electronic health records (EHRs) to committing physician and staff time and effort to data entry, management, and analysis. Several practices lacked in-house data management expertise and therefore invested in new types of staff dedicated to data management. Such investments were almost uniformly reported as being critical to practices' ability to succeed in these payment models.

In practices with more–highly developed data management capabilities, several leaders and physicians reported lacking the timely, accurate data they needed to respond to alternative payment models effectively. When present, these data deficiencies were a source of considerable frustration. The greatest concerns about data centered on the potential mismatch between data internal to a practice, from the EHR, reflecting “what actually happened” to the patient, and reports based on claims data, reflecting what was documented for billing purposes. Sometimes data in an EHR were not entered in a manner that facilitated capture during the coding and billing process (e.g., in free text rather than defined fields). Overall, practices seemed to trust their own internal EHR data more than they trusted external data, which they felt were at least one step removed from the “source of truth.”

Finally, respondents noted that accurate price data for health care services and commodities (e.g., specialty drugs) could be difficult to obtain. When data on prices were unavailable, this limited practices' abilities to contain the costs of care—as encouraged in alternative payment models, such as capitation, shared savings, and episode-based and bundled payments.

Interactions Among Payment Programs and Between Payment Programs and Government Regulations

The multiplicity of PFP and other incentive programs has created a heavy administrative burden for some physician practices. Merely keeping track of payment program details, which vary from payer to payer, required management effort that might exceed the capacity of some practices. In response, larger physician practices and hospital systems have stepped into the role of boiling those incentives down into something that is more manageable, and palatable, for their physicians.

Performance incentives offered by multiple payers can reinforce each other, and incentives from one payer, in some cases, led to practice-wide changes affecting all of the practice's patients. But, a serious tension could also arise when practices participated in a mix of both fee-for-service (FFS) and risk-based contracts. In those situations, some practices reported facing fundamentally conflicting incentives—to increase volume under the FFS contract while reducing costs under the risk-based contract. This conflict was especially acute for hospital-owned physician practices, in which reductions in hospital utilization—which are strongly incentivized under risk-based contracts—could undermine the financial well-being of the parent organization.

In addition, multiple practices described spillover effects from the EHR installations and upgrades encouraged by meaningful-use incentives. In some cases, EHRs had positive effects, facilitating the achievement of performance targets in PFP and shared savings programs. In other cases, especially when customized EHR modifications tailored to an alternative payment model could not be transferred to a new EHR, some interviewees described significant setbacks in their ability to meet the goals of these alternative payment models.

Effects of Payment Models at the Individual Physician Level

Physician Incentives and Compensation

In general, we found that the financial incentives applied to physician practices via alternative payment models were not simply “passed through” to individual physicians. Even practices of relatively modest size reported shielding their physicians from direct exposure to the financial incentives created by payers—except in the case of traditional FFS payment. In fact, the greatest marginal financial incentive facing nearly all physicians in the study, even those in practices with substantial exposure to payment models intended to contain the costs of care (capitation, shared savings, and episode-based payment), was to increase “productivity” as measured by revenues or relative value units (RVUs), a unit of measurement for physician services originally created for Medicare FFS payment.

However, physician practices did not ignore the quality performance or cost-containment incentives they received from payers or seek to insulate individual physicians completely from making changes in response to practice-level financial incentives. Rather, practice leaders described transforming certain practice-level financial incentives (especially those concerning cost containment) into internal nonfinancial incentives for individual physicians, choosing instead to appeal to physicians' sense of professionalism, competitiveness, and desire to improve patient care. Common nonfinancial incentives included performance feedback and selectively retaining or terminating their physicians based on quality or efficiency performance.

In several practices, leaders acknowledged the presence of inconsistencies between financial and nonfinancial incentives (e.g., applying RVU-based financial incentives simultaneously with admonitions to contain costs). Reported barriers to achieving better alignment included a lack of alternatives to RVUs for measuring physician “productivity,” a desire to avoid dramatic reallocation of income between physicians within the practice, and a need to balance the economic efficiency of physician compensation formulas with practical considerations (such as the operational costs of administering more-complex physician compensation formulas and the trade-off between the complexity and understandability of compensation incentives to physicians).

Generally speaking, alternative payment models had negligible effects on the aggregate income of individual physicians within our sample.

Some physicians reported wanting to have their incomes more closely linked to quality and efficiency of care. These physicians expressed an underlying desire to have better alignment between what they thought they should do for patients and what they were paid to do.

Physician Work and Professional Satisfaction

Within our sample, alternative payment models had not substantially changed how physicians delivered face-to-face patient care. However, the overall quantity and intensity of physician work had increased because of growing patient volume expectations and ongoing pressure for physicians to practice at the “top of license” (e.g., by delegating less intense patient encounters to allied health professionals), which was described as a potential contributor to burnout because lower-intensity patients could be an important source of respite for busy physicians.

Additional nonclinical work, particularly documentation requirements, created significant discontent. Physicians recognized the value of documentation tasks that were directly related to improvements in patient care, such as identifying patients with diabetes to facilitate better management of all patients with this condition, but they disliked the extra burden generated when documentation requirements were perceived as irrelevant to patient care.

Most physicians in practice leadership positions were optimistic and enthusiastic about alternative payment models, while most physicians not in leadership roles expressed at least some level of apprehension, particularly with regard to the documentation requirements of new payment models. Overall, even these physicians seemed to believe that major changes in payment methods would continue and acknowledged that some changes were useful. Nevertheless, their attitude was frequently one of resignation, rather than enthusiasm, because their day-to-day work life was more difficult and included burdens they did not believe would improve patient care.

Features of Payment Model Implementation

Factors Limiting the Effectiveness of New Payment Models as Implemented

Physicians and practice leaders described encountering three general types of operational problems in new payment programs that limited their effectiveness and sapped physicians' enthusiasm for them. By taking steps to avoid or prepare for these stumbling points, designers and implementers of future payment programs might be able to enhance their likelihood of achieving program goals.

First, physicians and practice leaders participating in a variety of alternative payment models described encountering errors in data integrity and timeliness, performance measure specification, and patient attribution (the process by which patients are assigned to a specific physician or practice). These payment models shared characteristics that might have made errors more likely: They were administratively more complex than FFS payment; some required payers to develop new measurement systems; and some were deployed for the first time quite quickly, without a “dress rehearsal” in which errors could be corrected before payments were on the line. Future participants in such models might consider such dress rehearsals or at least asking payers to design systems to quickly detect and correct implementation errors, which might be inevitable even in the best of cases.

Second, physicians had a variety of concerns about the implementation of performance and risk-adjustment measures underlying PFP, shared savings, and capitation programs. Broadly speaking, these concerns stemmed from a sense that the multiplicity of measures within and across programs could distract physician practices from making the changes to patient care that were actually the ultimate goal of many payment programs.

Third, the influence of uncontrollable, game-changing events in shared savings and capitation programs (e.g., the introduction of very high-cost specialty drugs) sapped physician practices' enthusiasm for these payment models. Finally, some physicians reported that they could not understand exactly what behaviors were being encouraged or discouraged by certain performance-based payment programs—even after seeking clarification from payers. Although these physicians reported that the performance bonuses they received were welcome, an incentive not understood by its target might not function as intended. Increasing the understandability of such incentive programs could enhance their effectiveness.

Conclusions

Nearly all physicians, physician practice leaders, and market observers who participated in this project described multiple simultaneous changes in payment models and regulations (such as meaningful use—a federal program to encourage physician practices to adopt and use EHRs). Most interviewees therefore described how interactions between these simultaneous changes, rather than the introduction of a given specific alternative payment model, affected physicians and physician practices. Prominent among these interactions were the tensions caused by lack of alignment among the plethora of performance measures and payment incentives deployed by different private and public payers.

Although our study did not seek explicitly to investigate the role of EHRs, the rapid uptake and upgrading of EHRs as a consequence of meaningful-use regulations was repeatedly described as the single greatest change in most practices, with both positive and negative spillover effects on nearly all practice efforts to respond to alternative payment models.

Physician practices played important roles as intermediaries and buffers between these changes in the health care marketplace and individual physicians within the practices. In some instances, practices magnified the impact that alternative payment models have on physicians' approaches to patient care—for example, by making substantial investments in new care pathways to enable successful performance in episode-based payment programs, even when such programs accounted for a negligible percentage of practice revenues. In other instances, practices shielded their physicians completely from specific aspects of alternative payment models—for example, when practice leaders made conscious decisions to ignore certain PFP measures to give their physicians a manageable array of targets for improvement. Physician practices also described translating external financial incentives from health plans into nonfinancial interventions (e.g., performance feedback and coaching) for individual physicians within the practice; this translation was nearly universal for financial incentives to contain the costs of care.

Practice leaders expressed considerable uncertainty about best strategies for responding to the combinations of alternative payment models that they faced, and doubts about the future compounded these uncertainties. Guided by practical limits on available capital and how much change their physicians could absorb quickly (especially when “change” amounted to “more work for each physician”), practice leaders tended to proceed cautiously, prioritizing areas in which multiple payment incentives overlapped with each other and with practices' internal priorities. For some smaller, independent practices, merging with larger practices or hospitals was an attractive option for accessing the capital necessary to succeed in alternative payment models (and for complying with new regulations, such as meaningful use) and enhancing their ability to control what alternative payment models they faced and how these would affect their physicians.

Implications

Informed by their experiences with alternative payment models, physicians, practice leaders, and other market observers described ways to enhance physician practices' abilities to respond successfully:

• Physician practices need support and guidance to optimize the quantity and content of physician work under alternative payment models. Alternative payment programs could create opportunities to reallocate physician work toward more satisfying content, which also could produce better, more efficient patient care. However, such payment programs also carry the risk of simply adding more work for already-overburdened physicians, risking burnout especially when such work is perceived as unrelated to good patient care. Developing physician leadership could help guide practices and health systems in their efforts to succeed in alternative payment models while preserving or enhancing physicians' professional satisfaction.

• Addressing physicians' concerns about the operational details of alternative payment models could improve their effectiveness. Although physician practices reported these problems in a minority of cases, failure to execute payment programs as intended, use of clinical performance measures with unclear validity, and deployment of financial incentives that physicians do not understand can all undermine the effectiveness of alternative payment models. Health plans can anticipate and correct these problems by conducting dry runs of alternative payment programs before “going live” and clearly communicating their intent to physicians (i.e., communicating what, if anything, physicians should do differently in response to the program). If a program's specific intent cannot be communicated clearly, this could be a sign that the program should be redesigned.

• To succeed in alternative payment models, physician practices need data and resources for data management and analysis. Practices must make substantial data infrastructure investments to manage patient care effectively and monitor the performance measures that underlie many alternative payment programs. Although the financial resources necessary to make these investments can come from practices merging with each other and with hospitals, health plans also should consider investing in physician practices' data management capabilities. Such investments could enhance the effectiveness of alternative payment models. In addition, greater data sharing with physician practices (particularly sharing the prices of all health care services, including drugs) would help practices make the best possible use of their data management infrastructure.

• Harmonizing key components of alternative payment models, especially performance measures, would help physician practices respond constructively. Within the bounds of antitrust law, steps by health plans to align their payment models with each other will free up the substantial physician practice resources currently spent on wrangling hundreds of performance measures and trying to create a coherent response to the problem of “50 people shouting their priorities at you.” If this cacophony can be ameliorated—and especially if government regulations can be harmonized with alternative payment models—practice leaders can better devote their attention to the difficult work of making meaningful and beneficial changes to their processes for patient care.