By Janet Silversmith on behalf of the MMA Work Group to Advance Health Care Reform
■ Among the leading strategies to reform health care is the development and implementation of new payment models. The goal is to change the way physicians, hospitals, and other care providers are paid in order to emphasize higher quality at lower costs—in other words, to improve value. In an effort to build on its health care reform activities that began in 2005, the Minnesota Medical Association convened a work group in 2010 to develop recommendations on how payment reform can best be advanced. Among the work group’s output was a comparative review of five payment models with respect to how they can support a value-driven health care system. This article summarizes the pros and cons of the five models—fee for service, pay for coordination, pay for performance, episode or bundled payment, and comprehensive care or total cost of care payment. It also offers the work group’s recommendations for how these models might be applied in a reformed health care system.
With the 2005 publication of the Physicians’ Plan for a Healthy Minnesota, the Minnesota Medical Association’s (MMA) recommendations for health care reform, the MMA articulated a goal of changing from the current fee-for-service payment structure, which provides incentives for volume and visit-based care, to one that rewards value and supports innovation in care delivery. One challenge facing physicians, hospitals, payers, and policy makers is managing the transition to new payment models. Because numerous models and countless combinations of them may be employed as alternatives to fee for service, selecting the most appropriate options can be confusing. The MMA’s Work Group to Advance Health Care Reform, which was convened in mid-2010 by the MMA Board of Trustees to review a variety of health care reform topics, assessed five payment models with respect to how well each one supports a value-driven health care delivery system. This article summarizes the work group’s findings in regard to the characteristics of each model.* The hope is to increase understanding of the strengths and weaknesses of each model in order to promote more balanced debate about health care policy.
■ Payment Models
The five payment models that the MMA work group reviewed are fee for service, pay for coordination, pay for performance, episode or bundled payment, and comprehensive care or total cost of care payment. Although variations and combinations of these models exist (and they may be known by different names), the work group concluded that they represented the most common ones currently in use or under consideration.
Several different perspectives can be used to evaluate payment models. For example, the relative financial risk to physicians and other providers may be considered as well as the potential for overtreatment or undertreatment of patients, as illustrated in the figure on p. 46. The framework for analysis used by the work group was how well the various models support the following 11 attributes of a value-driven health system:
- Care is patient-centered (ie, it takes into account the patient’s cultural traditions, personal preferences and values, family situation, and lifestyle; the patient is an integral part of the care team who collaborates with providers in making clinical decisions);
- Care is safe and effective;
- Care is timely and accessible (ie, the system is structured in a way that reduces waiting time for both patients and caregivers, and that care and the patients’ health information are accessible);
- Care delivery is efficient (waste is reduced or eliminated);
- Care is coordinated among providers and across facilities;
- Continuity of care and care relationships are supported and facilitated;
- Providers of care collaborate to deliver high-quality, high-value care;
- Care is optimized by the effective and efficient exchange/communication of patients’ clinical information;
- Physicians and other caregivers engage patients in ways that can maximize health;
- Accountability for each aspect of a patient’s care and for a patient’s total clinical care is clear; and
- Continuous innovation and learning occur.1
■ Fee for Service
Fee-for-service payment is reimbursement for specific, individual services provided to a patient. Fee for service is fairly easy to understand as a payment method, as each specific service (or procedure or intervention or piece of equipment) provided is billed and paid for. In its most common form, fee-for-service payment in health care differs from payment for goods or services in other sectors of the economy in the way it is priced. In most consumer markets, the list price is determined by what the consumer is willing to pay for an item or service. In health care, the amount paid for services is usually negotiated between insurers and other payers and providers. In the case of government payers, it is based on defined or administered rates often determined by a formula or funding levels. In addition, fee-for-service payments are somewhat constrained by coding guidelines and rules (ie, CPT and ICD-9) that define what can be billed and paid for.
When analyzed with respect to the 11 delivery system attributes, fee-for- service payment has several benefits. Among them is its emphasis on productivity. Fee for service encourages the delivery of care and maximizing patient visits. As a payment mechanism, it is relatively flexible in that it can be used regardless of the size or organizational structure of a physician’s practice, the type of care provided (eg, clinic visit, surgery, therapy session), the place of service (eg, physician’s office, nursing home, hospital, surgery center), or the geographical location of care. Fee for service does support accountability for patient care, but it is often limited to the scope of the service a particular physician provides at any point in time.
There are, of course, negative features associated with fee-for-service payment. For one, it offers little or no incentive to deliver efficient care or prevent unnecessary care. In its current form, it is generally limited to face-to-face visits and thereby thwarts activities such as care coordination and management of conditions by phone and/or email.
Although fee for service is easy to understand conceptually, it can be difficult to understand in practice. Patients may struggle to decipher the coding and nomenclature involved in billing, manage the numerous bills and explanations of benefits they might receive, and understand its application in inpatient settings, especially for lab, radiology, and anesthesia services. Because payment is limited to one provider for one interaction, fee for service does little to encourage management of care across settings and among multiple providers.
The work group identified the following types of care as being best suited for fee-for-service payment: emergency and trauma care; elective procedures that are not covered by insurance; and complex diagnostic services and treatments that are difficult to categorize in a bundle or episode of care.
■ Pay for Coordination
This model involves payment for specified care coordination services, usually to certain types of providers. The most typical example of this is the medical or health care home model whereby the medical home receives a monthly payment in exchange for the delivery of care coordination services that are not otherwise provided and reimbursed.
This model has garnered tremendous support among primary care providers. Minnesota’s 2008 health care reform act included provisions to promote health care home development and established requirements for health care home certification. Payments to health care homes are based on a patient’s chronic health and care coordination needs. It is too early to know whether Minnesota’s health care home model is successful because payments to health care homes have only recently begun and the number of certified health care homes is small.
A number of benefits are associated with the concept of paying for care coordination, which often is payment for support services or work that would not be paid for under a fee-for-service model and, therefore, would not be provided. Those benefits include the potential to improve and enhance the physician-patient relationship and communication between patients and providers; to increase the level of patient and family involvement in care decisions; and to improve flexibility in how, where, and by whom some care can be provided. The model is intended to reduce the delivery of unnecessary care (eg, duplicative tests and procedures, futile care) and some inefficient care (eg, emergency room visits for conditions that would be better handled by urgent care or in a physician’s office), thereby enhancing efficiency. Recipients of pay-for- coordination payments also may be able to support care between visits in more cost-efficient ways such as through phone calls, email, or group appointments.
The limitations of the model include the fact that many patients, and possibly some payers and purchasers, may assume or expect care coordination to be provided without additional or separate payment. Explaining the rationale for the coordination payment, a portion of which may or may not come out of the patient’s pocket, may be difficult and could undermine the patient-physician relationship. There are also questions as to the specific scope of care-coordination services and the expectations on the part of patients and providers regarding what should be offered in exchange for the care-coordination fee. Because of the time-intensive nature of services associated with this model, it is possible that, if used exclusively, it would limit time available for visits by other patients.
Among the types of care best suited for pay for coordination, as identified by the work group, are primary care management and care coordination for patients with chronic conditions, and care coordination for healthy patients who are at risk for chronic illness.
■ Pay for Performance
Pay for performance can be defined as a payment or financial incentive (eg, a bonus) associated with achieving defined and measurable goals related to care processes and outcomes, patient experience, resource use, and other factors. The idea of pay for performance has generated significant debate and has been used by most Minnesota payers—both public and private—for several years. The MMA developed principles to guide pay-for-performance implementation in 2007 and has worked hard to assure uniformity in measurement standards.
The evidence regarding the effectiveness of pay for performance in improving quality and reducing costs is mixed.2 When evaluated against the work group’s delivery system attributes, pay for performance offers the potential to improve the quality of care delivered (particularly for care that is measured), enhance the efficiency of care (if measured), encourage collaboration and promote accountability among providers, and encourage improvement by emphasizing outcomes of care.
The limitations center around the operational challenges associated with measurement. As it is currently implemented, many pay-for-performance programs use only single condition-focused measures that do not reflect the complexity of caring for patients with multiple conditions. Although pay-for-performance programs may drive improvements in care that can be measured, such care may be inconsistent with patient preferences. Programs with rigid measures and standards could create incentives for physicians to avoid high-risk patients and fire noncompliant ones. In addition, the administrative work associated with data collection and reporting may take time that otherwise could be devoted to direct patient care.
The work group determined that among the types of care best suited for pay for performance are services for which metrics already exist including management of some chronic conditions (eg, diabetes, asthma, heart failure) and certain surgeries.
■ Episode or Bundled Payments
Episode or bundled payments are single payments for a group of services related to a treatment or condition that may involve multiple providers in multiple settings. This model has been tested in a number of settings. Geisinger Health System in Pennsylvania, for example, developed its ProvenCare model as a bundled payment model for coronary artery bypass graft (CABG) surgery; other organizations have been experimenting with the Prometheus model in which evidence-based case rates are used to determine the total resources required to deliver clinically appropriate care for acute and chronic illnesses. The largest evaluation of a bundled payment model was Medicare’s CAGB surgery demonstration, which ran from 1991 to 1996. Four U.S. hospitals participated in the program, and each was paid a single fee for inpatient and physician services during hospitalization, readmissions within 72 hours, and related physician services during the 90-day global period, but not other pre- and post-discharge physician services.3
Minnesota’s 2008 health care reform act included a variant of the bundled payment approach in the form of baskets of care. Baskets of care were developed for eight conditions and services: diabetes, prediabetes, preventive services for children and adults, childhood asthma, low-back pain, obstetric care, and total knee replacement. But to the best of our knowledge, no providers in the state are offering the baskets and no health plans are paying for them. Minnesota’s baskets of care experiment was likely limited by the fact that the baskets were designed as products to be purchased directly by consumers rather than as an alternative payment mechanism.
The plusses of the episode or bundled payment model include its potential to improve coordination among multiple caregivers; its ability to support flexibility in how and where some care is delivered; its incentive to efficiently manage an episode (reduce treatment/manage costs); its simplicity from a billing perspective (one bill instead of many); and, its clear accountability for care for a defined episode.
The challenges associated with it include the difficulty of defining the boundaries of an episode (what care falls within and outside of the episode); its potential to increase barriers to patients’ choice of provider and/or geographic preferences for care if adoption is not widespread; lack of incentive to reduce unnecessary episodes; and the potential to avoid high-risk patients or cases that may exceed the average episode payment.
The work group identified the following types of care as being best-suited for episode or bundled payments: obstetric/maternity care, transplants, coronary bypass surgery, joint replacement surgery, other general surgeries, angioplasty, pacemaker/ICD implantation, and other ambulatory diagnostic or therapeutic procedures.
■ Comprehensive Care/Total Cost of Care Payment
The comprehensive care or total cost of care payment model involves providing a single risk-adjusted payment for the full range of health care services needed by a specified group of people for a fixed period of time.
Total cost of care payment is very similar to capitation, but the main differences are the use of more sophisticated risk-adjustment methodologies, limits on risk exposure, and incorporation of quality measurement.4
In Minnesota and elsewhere in the United States, adoption of the total cost of care model has been fairly limited. Some Minnesota commercial payers have expressed an interest in moving toward it as quickly as possible, and some have begun to modify their contracts with larger provider systems in a way that does that (eg, the performance-based payment arrangement between Fairview and Medica; the Northwest Metro Alliance shared-savings collaboration involving Mercy Hospital, Allina, HealthPartners Medical Group, and HealthPartners Health Plan; and Blue Cross and Blue Shield of Minnesota’s expanded incentive payment contracts with Allina, Essentia Health, Fairview, and HealthEast).4-6
The benefits associated with this model are improved flexibility for providers in terms of care delivery; greater potential for innovation in delivery design; incentive to deliver care efficiently; improved incentive for providers who serve a particular population to collaborate with each other; and improved emphasis on maximizing health.
The limitations of the model include the relative sophistication of data and information systems and analysis required of providers; the likely limited application of the model to larger and more integrated practices; the model’s potential to overemphasize population health at the expense of the health of individual patients; the incentive it creates to avoid high-risk or noncompliant patients; the possible decrease in patient choice of provider and/or geographic preferences for care if adoption of the model is not widespread; and the potential for care to be withheld (or perceived to be withheld).
Interest in payment reform is likely to intensify as new models of care delivery are tested and refined. Additional demonstrations and evaluations of the various models are needed to fully understand their relative advantages, disadvantages, and operational feasibility. There is, however, no silver bullet among the options. No single payment model is appropriate for all types of care or applicable in all settings, practice types, and geographic locations. As physicians, policy makers, and others search for improvements in how care is paid for, the work group hopes that their analysis will help shine a light on the best paths to pursue. MM
Janet Silversmith is the Minnesota Medical Association’s health policy director.
Acknowledgement: We would like to thank the following individuals who participated in the MMA’s Work Group to Advance Health Care Reform: Peter J. Dehnel, M.D., chair; Mary E. Braddock, M.D.; Kathleen D. Brooks, M.D., M.B.A., M.P.A.; Roberto G. Gamez, M.D. (Resident Fellow Section); Julie S. Gerndt, M.D.; Ronnell A. Hansen, M.D.; Donald M. Jacobs, M.D., FACS; Stephen J. Kolar, M.D.; Robert A. Koshnick Jr., M.D.; Thomas E. Kottke, M.D.; Mark Liebow, M.D.; Audrey M. Park-Skinner, M.D.; Randy Rice, M.D.; Rahul Suresh (Medical Student Section); Douglas L. Wood, M.D.; Patricia J. Lindholm, M.D.; David C. Thorson, M.D.; Benjamin H. Whitten, M.D.