AQA Invitational Meeting

Framework for Performance Measurement

Kevin Weiss, American College of Physicians

Kevin Weiss offered a framework for discussing clinically relevant, standardized performance measurements for ambulatory care. Urging quick action, he said that the lack of agreement on performance standards and data conventions has resulted in confusion about the validity of measures, imposed a burden on providers, and created confusion among patients over which provider or institution to select. Weiss endorsed the work of the Strategic Framework Board in setting up an evidence-based conceptual framework for quality measurement and reporting.

Weiss noted two pathways for quality improvement—intrinsic motivation (i.e., must be ready to engage in change) and extrinsic motivation (i.e., pay for performance)—and said that everyone must be mindful of the tension between the two. For example, there is concern that physicians can be driven to do too much if the incentives are set too high.

Weiss noted that the Institute of Medicine's priority levels are evidence-based, which he said should help to narrow the number of potential measures. If there are 20 priority areas, he said, there will probably be 3 or 4 good evidence-based ones and 10 to 20 performance measurements.

Weiss said that performance measures must be:

• Evidence-based.
• Scientifically sound (including a clear linkage between clinical process and outcome; and valid, reproducible, responsive, and risk adjustable as warranted).
• Clinically relevant (to providers, purchasers, and consumers).
• Transparent (i.e., open source and fully disclosed).
• Feasible (i.e., of reasonable cost and precise specification, and they must not add an undue burden on providers).

Feasibility, said Weiss, is probably the most difficult of these elements. How do we facilitate measurement so providers aren't burdened?

Next, Weiss looked briefly at existing processes, and noted that the National Quality Forum provides a national forum for developing consensus-based performance measures. The American College of Physicians, meanwhile, has a policy that supports well-designed demonstration projects on the use of performance measures for accountability.

It is time to move forward with accountability, said Weiss, adding that it was important to be careful about how accountability is rewarded. Weiss also noted that performance measures should have the same attributes for quality improvement and accountability.

The ACP's recommendation, he said, is "that pay for performance should become a top national priority, and that Medicare payments should lead to this effort."

Weiss also noted a couple of elements that he said have been largely missing from the discussion:

• The need for continuous review (to find the benefits of the improvement and to be mindful of the unintended consequences of bad measures).
• Recognition of the vulnerabilities of performance measures (including the need to ensure that each measure promotes not only quality, but also equity and access).

Finally, Weiss offered a road map to change. He said this road map involves raising public awareness, redesigning measures and reports, delivering information in a timely fashion, public reporting, leadership, and a system that rewards quality. Most important, he said, is to recognize that the Institute of Medicine has already provided the goals, and that the work ahead involves enhancing the evidence-based process, improving accountability, and making sure there is a surveillance cycle in place.

Discussion

The discussion opened with a remark about health disparities among uninsured Americans and the need to balance access along with quality and equity. A second person suggested that there will be even more uninsured in the next 2-3 years as insurance costs rise, thus increasing the urgency to act now.

One participant suggested that there already existed a broad consensus on performance measures, and that what was really needed to move forward was to put in place the right measures (including system metrics), the right levers (the stakeholders in the room), the right incentives (aligning professionalism with payment), and the right supports at the office level.

The discussion quickly moved into the area of accountability. Without accountability, performance measurements will have no meaning, said one participant, who added that any changes must also reduce the administrative burden on physicians. This concept of reducing the administrative burden on physicians was echoed by several other participants during the course of the discussion. Another, however, suggested that quality improvement need not be burdensome and challenged the provider community to improve the coding in order to pay for the interventions that provide the highest quality of care. Similarly, another participant said that the burden on physicians will disappear once measurement is incorporated into a practice and becomes part of the work flow. He added a challenge to the health plans: Is there a way to collect and aggregate data in local communities and give physicians one report that is credible? Someone else noted that the health plans have a great deal of data that can provide a window into some aspects of performance.

The way to get new codes approved is through the American Medical Association process, said one participant. He cautioned, however, that the problem with building on the existing system is that it is a per-procedure/per-visit model. We need to be looking more broadly at how we change the payment system to reward performance and quality care over time (for management of chronic disease), he said. He added that little progress would be made, however, unless the Medicare system is revamped at the same time.

Still regarding accountability, one participant asked how to engage physicians on quality improvement and accountability. These measures don't require sophisticated information technology, but they do require implementation. Another person suggested that there needed to be joint accountability (and better data coordination) at the physician level to address gaps in care, since patients often see more than one physician for their care.

How quickly can a better health information technology system be put in place? One participant noted that it will take a while for smaller private practices to acquire, manage, and report data on quality of care—since physicians would be doing this in addition to their current duties. The participant suggested progressively incentivizing practices to do this. Another stressed that health information technology will enable physicians to be more efficient—especially at the small-practice level. Another suggested that the answer to the "when" question will come after there is an answer on how to finance new information technology systems.

One participant said that there was work that could be done before an electronic medical records system was put in place. The participant noted that most physician systems are currently set up for billing, not for managing patient data. He suggested that it would be very helpful if the insurance plans, which do receive patient data, could act as a registry around a few, agreed-upon performance-based measures.

From a purchaser perspective, said one participant, it would be helpful to have a set of recommendations that were tiered based on the level of evolution of a physician's practice. For example, level one would be a practice that is unwilling to collect data for medical records; level two, those who are willing to collect information; level three, those who have implemented at least a minimal health records system; and, level four, measures that are geared to physicians who have put in place an advanced, interoperable electronic medical records system.

One participant suggested that the various Medicare pilot projects under way would provide a lot of useful data on quality improvement and information technology over the next three years. He suggested that private payers also set up similar pilots. The participant noted a sea change in the way that physicians were managing small offices, and stressed the need to learn from their efforts before rolling out new systems for improving quality.

One big question that came up repeatedly was: How are we going to put in place new health information technology and pay for changes to the system? One participant said there was no question that there were considerable costs to developing and implementing evidence-based performance measures. Another participant suggested that the costs must be shared by private and public payers.

Regarding payment models, one participant suggested that quality improvement needed to be thought about in two different pathways: one for those "at the top of the class" and another for those at the bottom. If we really want to improve quality for those at the bottom of the class, said the participant, then this implies a different incentives model. Another participant suggested that some of the existing pilot projects would offer a logical, reasoned, and measured way to deal with some of these issues.

One participant noted that there is no unanimity in what constitutes a pay-for-performance program. She noted three basic versions (including the CMS model) and said she felt that none is sustainable as a business model over the long haul. The challenge for those assembled here today, she said, is not to fall into a trap and think the versions on the books today are the only options.

There was also discussion about how to provide individual physicians with the flexibility to take the steps they want to address quality improvement. The current system is not allowing them to do this, said one participant. Another person, meanwhile, noted that a section of the provider community has already demonstrated success in managing chronic conditions and disease management, and is already adopting electronic health records and other automated technologies. Perhaps we should begin where successes have already taken place and try to build on these, he said.

One participant suggested that there are already agreed-upon measures at the physician level for prevention (of diabetes, hypertension, etc.) and that his organization was ready to move forward with demonstration projects to validate and show that these measures can be implemented in small practices.

Another participant expressed concern that the current system is not helping to create the uniformity that every provider says it needs. If we can't get uniformity on what we're measuring, it is hard to figure out how we can move forward, she said.

There was also concern raised about proprietary care programs. One participant noted that physicians are being offered incentives to compete in proprietary care programs rather than incentives for improving quality. We need to communicate that proprietary care programs aren't the right way to go, he said, but how do we do this given existing market forces?

Regarding a timetable for action, one participant said that it was critical to think about near-, medium-, and long-term horizons. It is extremely important that we have these different time horizons, he said.

In wrapping up the discussion, AHRQ's Carolyn Clancy asked each participant to write down two suggestions that reflect short-term priorities.

---

Previous Section         Contents         Next Section

---