In their efforts to improve health care system efficiency, payers have become increasingly active in promoting strategies to improve the value of care. While initially focusing on quality, in order to address their continuing increases in costs of care, payers have recently added cost or efficiency measures to their programs. This paper reviews the four most popular strategies to address cost and efficiency: pay-for-performance, public reporting, consumer-directed health plans, and tiered or limited networks.
We begin by considering the strengths and weaknesses of the current methods of collecting and interpreting cost efficiency data, reporting results to interested parties, and providing incentives to shape provider behavior and patient outcomes. We then examine the four specific strategies in detail.
General Measurement Issues
Accuracy
The current trend in efficiency measurement is to calculate an index of physicians' costs relative to those of their peers. Efficiency indexes are built on the assumption that underlying claims data are correct, but there are a number of problems with that assumption. For example, on their visit billing forms, the doctors may use disease descriptions but not diagnostic codes. If the doctor writes, for example, "depression," a clerk must later decide whether this means minor depression, major depression, psychotic depression, reactive depression, or adjustment disorder with depressed mood, all of which are coded differently. In addition, because of the complex insurerhospital contracts, which include payment formulas tied to volume and sites of service, the cost of any single hospitalization or procedure may be difficult or impossible to calculate. The result is estimated costs that may not accurately incorporate facility costs into the overall cost calculation. In episodes of care that involve hospitalization, those hospitalization costs are often a major determinant of total episode costs. The accuracy of each health plans management of data is also crucial to reporting validity and should be transparent to the subjects of the reporting (i.e., providers) and those using the data to make decisions.
Attribution
There are many ways to assign a patient's costs to a practitioner, practice group, or integrated system, whether for population- or episode-based measures. In fact, different attribution rules can result in significantly different conclusions. Recent data suggest that some beneficiaries change assigned physicians from one year to the next, making assignment by individual physician inaccurate. Determining how patients' costs are assigned may influence how some clinical decisions will be made and the degree to which physicians are likely to effectively collaborate. For example, if assignment of costs is based on the physician with the highest cost in the episode of care, the collaborating physicians will have incentive to have others order expensive tests. If, instead, all the physicians who contribute to more than 30 percent of the overall costs are assigned the cost of the episode, all the involved physicians will have incentive to talk with each other about how to minimize unnecessary care and costs.
Assignment of Cost
Practitioners reasonably argue that they can only be held accountable for the costs that they themselves engender. However, scoring on these direct or responsible costs creates an incentive to make other practitioners order the more expensive interventions, such as medications, imaging, consultations, or hospitalizations. This approach can cause delays in appropriate care because of disagreements over who should prescribe chronic medication refills or who should order an indicated MRI.
Incentive Program Goal
Stakeholders debate whether an efficiency measure should pay for achieving a target or for significant improvement. Those arguing for reaching targets say that to deliver true quality care, a particular level of care must be assured. Therefore, they conclude, one should be rewarded for reaching that target. On the other hand, if one employs an all-or-none model, those who were top performers before creating the incentive are rewarded, discouraging the lowest performers from investing in an improvement program with a target they are unlikely to reach, and encouraging providers to preferentially select patients already close to the target.
Actionability
For efficiency-reporting programs to be successful they must predictably pinpoint practitioners' overuse and misuse behaviors, those that add cost but do not improve desired outcomes. Simply providing a global rolled-up measure, such as a cost-efficiency index that compares them to a specialty average, is itself inefficient. For the busy practitioner, trying to find what to do differentlywithout the proper toolsresults in wasted time, frustration, and, eventually, a lack of trust in the sponsoring organization. Meanwhile, in trying to become more efficient and improve their scores, they may reduce their treatment of necessary as well as unnecessary services. Finally, if the desired action is to move patients from less efficient to more efficient practitioners, program administrators should be certain there is excess capacity in the more-efficient group to accommodate the shift. If there is not capacity in the "more-efficient" group, patients will have to pay a higher copay for care that is ranked as less-effective care. That situation can only lead to frustration and raise patients' concerns about their physicians' quality.
Pay-for-Performance
Pay-for-performance is designed to provide a direct link between payment and the performance of carefully selected services. The measures selected are intended to evaluate the practitioners' performance on certain combinations of cost/efficiency, quality, and a patient's experience of care. Although there is growing acceptance of pay-for-performance programs, its influence has been hampered by small incentives, lack of adequate post-intervention follow-up, and methodological problems with collecting and reporting the data.
We believe that insurers will come to understand that rather than unsuccessfully attempting to identify and reward the best practitioners, the most productive use of pay-for-performance programs is promoting behavioral change for a small set of carefully selected measures for which there is ample evidence of need for improvement. Organizing the professional community around a specific set of behaviors that are known to have a direct impact on outcomes creates incentive to improve. Although there is concern about incentives diverting attention from nonselected activities, if the measures chosen are of significant benefit, the time focused on them would be well spent.
Public Reporting
Public reports are designed to drive system change in three ways: using comparison data to motivate providers to improve performance, stimulating payers to reward quality and efficiency, and providing information to patients so that they can choose their care more wisely.
But physicians tend to see the public reporting of their individual quality and cost data as misleading and even threatening. Their concerns about accuracy, sufficient sample size, and the inherent judgment imposed by forced rankingsseemingly to motivate physicians to change through fear of humiliation or shameput the intent of public reporting programs into question.
Data should of course be available to consumers to allow them to decide from whom to seek care and what outcomes they may expect. The challenge, however, is to present the data in a way that minimizes the judging of physicians and focuses on encouraging quality improvement. Prematurely providing misleading or insufficient information to consumers risks alienating practitioners while offering no real benefit to patients.
Consumer-Directed Health Plans (CDHP)
All variations of CDHP try to offer consumers broad choices of providers and services, coupled with greater information aboutand liability fortheir prices. In theory, if there is access to information on cost, care options, quality, and risks, patients will act as informed consumers in a competitive market, helping to manage their costs by taking more responsibility for their health care decisions.
But CDHP depends on patients proper interpretation of reported data, an area in which even physicians have difficulty. Even more important, there is good evidence that patient cost-sharing can actually decrease the quality of care. As a patient's cost increases, fewer services, appropriate along with inappropriate, tend to be used.
Meanwhile, physicians are deeply concerned that they will spend valuable time looking up comparative fees, become more involved in negotiating rates, and have higher office expenses as bills are increasingly collected from patients' health savings accounts or out-of-pocket. To many practitioners, this model increases rather than decreases system inefficiencies. This is especially true now, when the data available to patients is simply inadequate to reliably improve decision-making.
Tiered and Limited Networks
Tiered networks utilize two mechanisms for controlling cost. The first is to encourage patients to switch from lower-value to higher-value physicians by offering lower copays for higher-value physicians or higher copays for lower-value physicians. The second mechanism is to encourage physicians to become more cost efficient in order to avoid the negative outcomes of being advertised as a lower-tier physician, with an attendant loss of patients.
Unfortunately, the criteria for selection in tiers can vary significantly. Selection is confounded, moreover, by issues related to geographic needs and sociodemographic variables, making the validity of these forced rankings suspect.
One obvious limitation of these approaches is access to providers. The concept of restricting access to a subset of providers necessarily implies that there is a pool of excess high-value providers from which to choose in that geographical area. In many communities, this is simply not the case, especially in rural and some inner-city areas.
Conclusions
The lack of demonstrated efficacy, the problems with trying to identify the precise qualities of a physician that lead to predictable increases in health care value, and the inherently judgmental nature of these four strategies suggest that public reporting, tiered and limited networks, and pay-for-performance programs that try to judge the overall quality of a physician are unlikely to fulfill their expectations. Results to date lead us to conclude that, in the health care industry, basing change on the foundation of winlose relationships does not work. As we look to create a more effective and efficient health care system, it is critical that we not mistake programs to force behavioral change with successful quality-improvement processes.
In such a process, the physician's behavior is seen as contributing to a system of care. As that system is evaluated for how it delivers services, behaviors that help or hinder the attainment of desired outcomes may be identified. By focusing on specific units of behaviors rather than critiquing the overall performance of one individual physician, the focus remains on the need to improve rather than on placing blame.
We believe that the current focus on judgment programs has accomplished little and will continue to meet with limited success. These programs ignore some of the more important lessons that the business community has learned about true quality improvement. Successful improvement is rooted in partnership based on a commitment to achieve mutually agreed upon goals and a sense of shared benefit. Transparency must be delivered not only in the data reported to patients and purchasers but also in the processes that health plans use to make decisions about physician performance.
Successful transformation of our present system to one based on value requires the collaboration of multiple stakeholders. We focus on collaboration because effective partnerships are needed to effectively coordinate behaviors, provide a continuum of services, and ensure that evidence and patient-centeredness underlie medical decision-making. The current medical system's silos must be broken down and replaced with integrated programs that encourage and reward effective and efficient system solutions.