Time to Do the Right Thing: End Fee-for-Service for Primary Care

You can always count on Americans to do the right thing - after they’ve tried everything else.

–Winston Churchill

Current fee-for-service (FFS) payment rates for physician visits trace to the origins of Blue Cross Blue Shield insurance in the 1930s. At that time, rates were set that paid generously for hospitalizations and for procedures, such as surgery.^1,2 Payments for so-called “cognitive services” were lower per minute of physician time. This disparity has been perpetuated since the 1980s in the calculation of rates set by the Centers for Medicare & Medicaid Services (CMS), based on “Relative Value Units,” for payment of the Evaluation and Management codes most often billed by primary care physicians.³ Despite recognition by the Medicare Payment Advisory Commission (MedPAC)⁴ and others⁵ of the adverse effects of inadequate payment for primary care, only limited progress has been made toward correction of the disparity. This may be due, at least in part, to treatment of total payment for physicians as a zero-sum game in which decision making is dominated by non–primary care physicians through mechanisms such as the Relative Value Scale Update Committee (RUC).^6,7

Although in the United States we are most familiar with this fee-for-service payment arrangement, it is a creation of man, not nature. And even in this country, other models have existed since at least the mid-20th century. Most prominent of these others is the capitated payment model pioneered by Kaiser Permanente beginning in 1945⁸ and expanded in the Health Maintenance Organizations (HMOs) of the 1980s and 1990s.⁹ Following collapse of many of the HMOs, payers have experimented with multiple smaller changes in payment models, mostly incremental adjustments to existing FFS, perpetuating structural disadvantages for primary care.

In this issue of the Annals, Basu, et al report on their study in which they calculated potential effects on primary care practice costs and revenue resulting from 3 modifications of FFS payment: increased FFS, traditional FFS plus per-member per-month (PMPM), and traditional FFS plus PMPM plus pay-for-performance (P4P).¹⁰ The authors drew on extensive published literature about how primary care practices can staff and organize to deliver patient-centered medical home (PCMH) services such as enhanced access, care management, and alternative visits, and the impact of these changes on revenue and expense within the practices. The authors created simulated models of these financial effects based on detailed profiles of patient demographics, insurance coverage, and disease burden. They conducted tens of thousands of iterations to determine results assuming random variations within expected ranges, and sensitivity of their results to reasonable variations in assumptions, such as about payment amounts.

This was an extraordinary project, impossible to conduct as an experiment in real life due to the huge number of variables considered and number of iterations possible in the simulation environment. The findings are granular, robust, and convincing: none of the payment strategies considered provides sufficient support for primary care practice transformation beyond the minimum needed to meet PCMH criteria. And meeting minimum criteria is not the same thing as actually transforming care to improve quality and reduce cost.¹¹

The authors report their results in exhaustive detail, including extensive online appendixes. If each data point in their results is 1 grain of sand, together they form an arid desert, one which the shallow wells of modified FFS cannot turn into blooming oases. As the authors dryly conclude, “Achieving goals of the PCMH will likely require more radical payment reforms to specifically target funding towards the delivery of desired services.”¹⁰ Bluntly put: tweaking FFS payment will not save health care.

Although this is a seminal and powerful paper, the simulations do not address all possible futures. Rather, they only extrapolate from the present, and PCMH service models are sure to evolve over time. That said, the authors have developed a methodology that can be applied to future payment and practice models and incorporate new data as it becomes available. In addition to, for example, extended interdisciplinary teams in primary care, these might include effects of new value-based payment models such as the CMS Merit Based Incentive Payment System (MIPS), Alternative Payment Models (APMs),¹² Next Generation ACOs, Shared Savings,¹³ and Direct Primary Care (DPC).¹⁴

Future applications should go beyond assessing payment effects on revenues and expense within the PCMH alone to address effects on total cost of care. This is because the financial benefits, if any, of advanced primary care such as PCMHs, accrue not within the practice itself but as a result of “downstream” savings such as from reduced emergency department visits and hospital admissions.^15,16 Achieving the benefits of overall care redesign will require new payment models that reward primary care practices for results across the full continuum of care, not simply pay for services provided within the primary care practice itself.

This report provides ample evidence for a simple truth: it is time to end FFS payment for primary care.¹⁷ The paper puts a very large nail in the coffin for failed models of payment based on FFS. We have tried everything else. It is time for America to do the right thing and implement comprehensive payment for primary care that rewards physicians and their colleagues, not just for providing “desired services,” but for achieving the Quadruple Aim of better care, lower cost, better health, and clinician satisfaction.¹⁸ This will enable a system that is incredibly attractive and satisfying to teams of skilled professionals working together. It can become one that is financially sustainable, not just for primary care practices but for American health care, for our communities, and for our patients.

• Received for publication August 2, 2016.
• Revision received August 2, 2016.
• Accepted for publication August 2, 2016.

• © 2016 Annals of Family Medicine, Inc.