Improving Health Care Quality: The Path Forward

Mark B. McClellan

Chairman Baucus, Ranking Member Hatch, and members of the Committee, thank you for this opportunity to discuss the current landscape and next steps for improving the quality of health care. I appreciate your leadership in focusing the nation’s attention on improving quality, as this is fundamental to achieving better health outcomes while avoiding unnecessary costs. Simply expanding health insurance coverage to promote access, or trying to lower costs by cutting prices or covered services, will not achieve the best health and the lowest health care costs for Americans. Improving how care is delivered is essential. Health care providers and patients have many good ideas for how to improve quality and lower costs, but often these approaches are not supported well or at all by fee-for-service payments, traditional insurance benefit designs, or current health care regulations.

Much of my work, and the work of my collaborators and colleagues, remains focused on health care policy reforms, reforms in the private sector, and public-private collaborations to support providers and patients in their efforts to get to better care. I am a Senior Fellow at the Brookings Institution, where I direct a range of projects related to improving innovation and value in health care. I co-chair of the Quality Alliance Steering Committee, a multi-stakeholder group of employers, insurers, providers, and consumers that focuses on overcoming the practical challenges in implementing quality measures and using them to improve care. I chair the Roundtable on Value and Science-Driven Health Care of the Institute of Medicine (IOM), which focuses on improving clinical evidence and its use to achieve better care. I chair the Clinician Workgroup of the National Quality Forum (NQF)’s Measure Application Partnership (MAP), which prioritizes and recommends performance measures for implementation in Medicare and other Federal programs. Previously, as Administrator for the Centers for Medicare & Medicaid Services, I oversaw the implementation of a range of quality-related payment reforms, including provider reporting on quality and patient experience, and payment reforms related to “shared savings” and accountable care.

Opportunities for Improving Health Care Quality

We’ve made a lot of progress to support better quality care in recent years. Building on bipartisan legislation and support from the Congress, Medicare has established quality reporting systems for providers. There is more activity than ever around the development of quality measures, thanks to private organizations like the American Medical Association’s Physician Consortium for Performance Improvement (PCPI) and the National Committee for Quality Assurance, as well as public support and initiatives in the Centers for Medicare and Medicaid Services (CMS) and the Agency for Healthcare Research and Quality (AHRQ). The National Quality Forum (NQF) has taken important steps including assisting with the prioritization of measures for development and implementation, and especially in “endorsing” quality measures to promote the consistent use of meaningful, well-understood measures.

Today, there are numerous and diverse quality improvement initiatives underway at all levels of the health care system – federal, state, regional, local, and within health care organizations – that are putting quality measures to use. Quality improvement initiatives within and across health care organizations are core to these efforts. They require measurement in order to identify opportunities for improvement, often through “registries” that enable providers to assess and track how their patients are doing in terms of key aspects of care and potential complications in order to identify areas for improvement. Quality measures are also being used for payment reforms, which can enable health care providers to get more resources to take steps like setting up registries and implementing other changes in care delivery to improve care and avoid unnecessary costs. As an illustration, fee-for-service payments in Medicare, Medicaid, and the private sector have historically provided little financial support for many activities that can improve patient care and potentially reduce costs. Examples include taking time and implementing systems to coordinate care to avoid duplicative or inappropriate services; answering patient calls or emails to avoid the cost and delay of an office visit; and spending more time with a complex patient (or implementing a care team with a nurse practitioner, pharmacist, and other non-physician clinicians) to improve medication adherence, lifestyle changes, or other care management steps that can enable patients to prevent their diseases or health risks rom progressing. Private payers, employers, Medicare, and Medicaid are all undertaking a range of payment reforms to provide better support for such activities, generally in conjunction with using quality measures. Finally, quality measures are used increasingly in public reporting, thanks to national efforts like those supported by Medicare as well as impressive regional efforts, such as Puget Sound Health Alliance, Minnesota Community Measurement, Wisconsin Collaborative for Healthcare Quality, Pittsburgh Regional Health Initiative, and many others.

But you only have to look at the evidence on what these quality measures show to know we have a long way to go. Last September, in its report on “Best Care at Lower Costs: The Path to Continuously Learning Health Care in America,” the IOM noted that patients get effective care only about half the time, that gaps in coordination remain widespread, that serious preventable medical errors are common, and that perhaps more than 30 percent of health care costs could be avoided as a result of improving quality and efficiency. These are not new findings; studies have been using progressively better quality measures to document gaps in quality and broad variations in costs that are not related to quality for decades.

We also still have a long way to go in quality measurement. Many important quality measures available today have not been widely or consistently implemented. We lack robust quality measures for many important aspects of health care. We don’t have reliable, widely available quality measures for most of the things that really matter to patients, like the experience of care for patients like them, or measures related to their outcomes like how well they can function, work, and undertake their activities of daily life. And as you heard at last week’s very important hearing, we don’t have reliable and consistent information on the price and costs of care. Again, these are not new problems.

Some of the challenges facing quality measurement include: lack of alignment of key measures between public and private sector quality improvement efforts; issues regarding data transfer such as merging data across different information technology systems; ensuring secure protection of sensitive patient data and proprietary information; and developing, endorsing, and implementing measures of value – that is, sets of measures that include both quality and cost information. But the most important obstacle to greater use and impact of quality measures is that, today, quality still doesn’t matter that much in health care financing systems, including Medicare and Medicaid.

Policy Reforms to Support Better Quality and Lower Costs

Recently, along with a group of health care leaders and experts, I authored a report on “Person-Centered Health Care Reform: A Framework for Improving Care and Slowing Health Care Cost Growth,” which described how to address the persistent problem of health care quality in all parts of our health care system. The report was the third in a Brookings series on “Bending the Curve” of rising health care costs. It included a wide range of health care experts as well as public policy leaders including Dan Crippen, Glenn Hubbard, Peter Orszag, Mike Leavitt, Donna Shalala, and Tom Daschle. What all of us concluded together was that the best way – really, the only way – to assure that we could achieve health care that was financially sustainable was to reform our health care financing and regulatory policies to do much more to support better-quality care and lower costs at the person level.

Our report proposes a framework for reforming health care financing and regulation to achieve better, higher-value care for each person. It describes a specific series of steps building on current initiatives to improve the way care is delivered in each part of our health care system, including Medicare and Medicaid, the employer and individual insurance markets, as well as antitrust enforcement and other regulatory reforms. The estimated net savings in the overall plan are around $300 billion at the federal level over the next decade (2014 – 2023). After gradual implementation of the proposed reforms over the coming decade, long-term savings from better care and sustainable spending growth will exceed $1 trillion over 20 years. While this framework focuses on lower costs through supporting reforms in health care delivery, it can be combined with other reforms to achieve additional reductions in health care costs.

The report recognizes that we live in a time of unprecedented breakthroughs in genomics, systems biology, and other biomedical sciences that are leading to better prevention and to innovative combinations of treatments based increasingly on each person’s characteristics and preferences. Furthermore, improvements in wireless technologies and other non-medical technologies make it possible to prevent complications, and deliver care at home and in other settings different from traditional medical care. To take advantage of these opportunities to improve care, health care financing must shift away from paying on a fee- for- service basis for specific medical services, and toward paying for coordinated care that meets each patient’s needs.

Focusing on person-level quality of care as the fundamental strategy for addressing health care cost growth is in some ways new, but it builds on promising ideas and trends throughout our health care system. Our group is by no means the only ones who have reached this conclusion. A broad variety of recent reports, from the Simpson-Bowles Commission, the Bipartisan Policy Center, and others, all agree that the most important thing that policymakers can do now to improve health care quality is to make feasible changes in health care payments and benefits so that they can better support patient-centered care. Improving quality in health care is difficult, it must be done carefully to avoid unintended consequences, and the quality and cost problems won’t be solved overnight. But so long as providers are generally paid more for more services rather than better quality, and so long as patients get more benefits and more financial support when they use more services rather than take steps toward better health and using care more effectively, our policies are not providing the needed support and momentum for solving these problems.

The same principle applies to achieving better quality measures. Just as it is hard for patients to understand prices when they have to add up dozens or hundreds of specific fees for specific services, it is hard to get a meaningful picture of the overall quality of care at the patient level from quality measures that have to accompany dozens of specific services across different types of providers. While these specific aspects of care all matter, what really matters to most patients is how these specific services or aspects of care come together for their specific needs. For a knee replacement, putting all of these services and processes of care together, were they treated well, did they avoid any safety problems and complications, and is their knee function improved? For patients at risk of complications from a chronic disease like diabetes or high blood pressure, are they using the medications that minimize the chance of the disease progressing, and are they getting support in making the changes in their lifestyle, so that they are really reducing their risk of disease progression? For a life-threatening condition like cancer, were they and their family included in the process for making decisions about care, and did their many treatment decisions and up to good results and the best possible experience of care based on their particular circumstances and preferences? Especially if they have multiple health issues, as Medicare beneficiaries often do, patients also care about how they can get the best quality of life overall for themselves and their caregivers. Finally, are these results being achieved at the lowest cost? These are multidimensional, complex, and highly personal issues that cannot be measured perfectly and that, in the end, depend on health care providers being able to focus on the needs and goals of each individual patient. If health care financing and regulation could be better aligned with the aims of clinicians and their patients, there would be more support and better incentives both for developing and using measures that matter, as well as for actually improving care.

Next Steps for Improving Quality of Care and Quality Measurement

            As I have noted, the problem of quality is not just or mainly a problem of quality measurement. It is a problem of providing better support for the hard work and reforms in care delivery needed to improve quality and lowering costs as a result. Despite the challenges, many clinicians and health care organizations are making progress, reflecting their fundamental professional commitment to better care and better health for their patients, as well as progress to date in policy reforms to support better care. But without further steps to support better care at the patient level, progress on both quality improvement and quality measurement will be slow.

I have four recommendations for the Committee:

  1. Take further steps to transition payment systems in public programs to case-and person-level payments.

To support quality measures that really matter to patients, further payment reforms building on recent trends in the public and private sectors are needed. Medicare has taken some notable steps toward payments that focus on the episode or person level of care. These include diagnosis-related group (DRG) payments for hospitals and the recent penalties for readmissions, person-level payments and quality measures in the Medicare Advantage program, person-level payments in accountable-care organization (ACO) payment reforms like the Medicare Shared Savings Program and the Pioneer pilot program, and past and current episode-based payment pilots. But most Medicare payments are still siloed and based on fee-for-service. Our recent report on “Person-Centered Health Care Reform ” describes a way to transition to greater use of episode- and person-level payments in Medicare. The basic idea is that it can happen gradually, starting now, with a modest element of case-based or episode payments in physician payment, post-acute care payment, and other payment systems that currently pay primarily or entirely on a volume and intensity basis, rather than on the basis of patient need and quality. This will help focus the development and improvement of performance measures that reflect the outcomes, experiences, and other key aspects of quality of care at the episode and person level.


Even though current quality measures at the episode or person level are far from ideal, providers and patients can still benefit from the shift of a component of their payment from fee-for-service, because it gives providers more ability to provide individual patients with what they need, rather than just what’s covered in fee-for- service. A wide range of physician specialties have identified specific ways in which a limited amount of their fee-for-service payments could be shifted to episode- or person-level payments in the near term to get better results while reducing overall costs. For primary care physicians, this is the payment reform idea behind the patient-centered medical home. The same idea applies to specialists as well. For example, clinical leaders from the American College of Cardiology, the Society of Thoracic Surgeons, and others have described the concept of the “Heart Team” working across specialties like cardiology, interventional cardiology, and cardiac surgery to more effectively identify which patients with different types of heart disease should be referred for specialty care, what tests should be performed to support their care, what information they should receive to make the best decision about care in their own circumstances, and how all of these experts can best work together for each patient. Shifting a part of the payments that specialists receive for performing procedures that are not well coordinated now into a case-based payment for their patient – and providing an opportunity to share in the overall health care cost savings that could occur – would provide better support for the Heart Team approach. Similarly, some oncologists have started to implement “Oncology Medical Homes” for their cancer patients, which provide ongoing tracking for the patient’s status and use of evidence-based treatment, and more staff support for preventing emergency room visits and hospitalizations (e.g., after-hours access to a member of the patient’s oncology team), among other things. But unless some of the payment for oncologists shifts from current fee-for-service activities, such as the intensity of chemotherapy use and use of imaging, it is very difficult for oncologists to put the resources into these activities that can improve quality of care and prevent costly complications. Radiologists and other physicians who provide technical services in collaboration with other providers could also benefit from such a payment, as it would enable them to devote more effort toward making sure that the right tests are being used – and used well – rather than just being reimbursed based on volume.

Case- or patient-level quality measures have not yet been fully developed to support the adoption of such payment systems for all health care providers, and sudden major changes in payment could disrupt needed care. But starting an incremental transition in this direction would provide a strong and predictable foundation for making more progress on the development and use of meaningful quality measures – and more importantly, it could help improve quality of care. As the Finance Committee considers legislation in the near future on Medicare physician payment, and possibly other areas such as payment for post-acute care and other Medicare services, including steps away from payments for specific services and toward patient- or case-level payments would provide important momentum for achieving better patient care.

2. Take further steps to implement case- and person-level quality measures in public programs.

In conjunction with clear steps to implement payment reforms that enable providers to focus more on quality of care, outcomes, and experience at the patient level, reforms should also include clear and predictable expectations for meaningful progress on accompanying quality measures. A growing number of performance measures are in use in both the public and private sectors, in part due to the momentum provided by their inclusion in Medicare’s payment systems. But as I have noted, relatively few measures address outcomes that matter for patients and patient experience. Although available outcome and experience measures have important limitations, describing a clear path by which they will be incorporated in Medicare’s payment system and beginning to do so will provide momentum for their further refinement and endorsement.

A growing set of case- and patient-level measures are becoming available, or could transition into more widespread use. For example, patient experience measures have been developed and endorsed for a wide range of settings of care, and for the overall care experience of many types of patients, yet the use of patient experience measures outside of the hospital and ACO setting is limited. With respect to cardiovascular disease risk, process measures like “Body Mass Index-Screening and Follow-Up” are in relatively widespread use. A more meaningful, outcome-oriented measure is the result of screening and followup: a person’s long-term cardiovascular disease risk and changes in that risk. Many such measures have been developed, such as the cardiovascular risk assessment of the National Cardiovascular Education Program and Heart Health Risk Assessment used by Kaiser Permanente. An even broader measure that is being implemented in some health care organizations is a ten-year mortality predictor developed by Drs. Elliott Fisher, Chris Murray, and colleagues at Dartmouth Medical School and the University of Washington. This measure incorporates twelve major health and behavioral risk factors (e.g., smoking and blood pressure) and can be used to counsel and engage patients in addition to track risk reductions. Because improvements in measures like these will significantly improve outcomes that matter to patients, they are much more “patient-centered.” For elective joint replacement for osteoarthritis of hip or knee, post-operation complication rates like readmissions are coming into more prominent use. Stronger next steps would be to include measures of patient experience and functional outcomes, both reported by patients themselves. Such measures are being used in some programs, and are being further developed and implemented through quality improvement initiatives like the High-Value Healthcare Collaborative. Reflecting the need for further refinements in these measures, as well as the fact that even the most effective providers cannot control all or most of the factors that influence important patient outcomes, these outcome-oriented performance measures need only have a limited role in payment, at least initially. In fact, they might not be tied directly to payment amounts at all; many Medicare payment systems have first used measures for quality improvement, and only later for public reporting or performance-based payment.

This emphasis on key outcome and experience measures could help drive greater use of many other supporting measures of quality. In many quality improvement initiatives today, to support providers’ efforts to achieve better outcomes, entire sets or systems of measures have been developed, including many evidence-based clinical processes of care or “structural” features of care systems. These detailed measures assist health care providers in identifying specific ways to improve outcomes, and also help develop new evidence on ways to achieve better outcomes in the future. For example, to support improvements in outcomes, the Society of Thoracic Surgeons’ National Database on Coronary Artery Bypass Grafting includes NQF-endorsed measures of risk-adjusted mortality and morbidity after surgery (and for some patients, the database includes longer-term outcomes). It also includes many details on perioperative medications and operative care processes, which have been used both to help surgeons identify opportunities for improving care and to better understand which processes of care lead to better outcomes. The American College of Cardiology and other medical specialties have also supported the development of clinical registries with detailed measures of processes of care and an increasing emphasis on the outcomes that these processes are intended to affect.

Many providers and collaborations across providers are moving forward on implementing patient registries and tracking systems with detailed quality assessments, as part of a strategy to implement or prepare for reimbursement systems in which payment is moving to the patient level. These payment reforms include ACOs, bundled payments, and other types of capitated payments. For example, many ACOs are working with expert advisers and implementing reforms in their information technology systems to develop increasingly sophisticated clinical and operational “dashboards” of measures that enable them to improve patient outcomes and reduce overall costs. Under those payment systems, quality improvement that leads to improvements in patient experience and outcomes can receive much more financial support. Other organizations – such as the Joint Commission, the Premier Healthcare Alliance, and the High-Value Healthcare Collaborative, as well as many consulting groups – are also implementing systems of measures to help hospitals, healthcare systems, and other providers improve care. Implementing meaningful patient outcome and experience measures can help leverage all of these important activities.

3. Support the NQF and a streamlined process for developing, endorsing, and incorporating more meaningful quality measures into public programs.

NQF serves very important roles in helping to improve quality through better quality measurement. It has identified priority areas of measure development, it “endorses” quality measures, and it helps provide guidance for the implementation of measures in public programs. As Dr. Chris Cassel has noted, the NQF endorsement process helps assure that quality measures are consistent and to prevent the administrative burdens and difficulties of comparisons across measures that intend to assess the same aspect of quality but are specified in somewhat different ways. Through input from a wide range of stakeholder groups, the NQF process also helps assure that measures are both feasible and can significantly improve quality. NQF is working on ways to continue to improve the efficiency and impact of its prioritization and endorsement process.

The Clinician Workgroup of the NQF’s Measure Application Partnership (MAP) has helped illustrate how this can be done. Among other things, the MAP is responsible for making recommendations on which quality measures should be adopted in Medicare’s payment systems for clinicians, which should be rejected, and which need further development. Because of the diversity of measures needed for the vast array of specialties and clinical care in the Physician Quality Reporting System (PQRS), the MAP has had to make recommendations related to adoption for literally hundreds of quality measures each year. To manage this workload with limited time and staff budget, the MAP developed a set of principles to guide our recommendation process and to make it predictable for interested stakeholders. I would like to highlight three general considerations to employ moving forward to enable a flexible, faster mechanism for achieving consensus around meaningful measures:

(1) Identify a core set of endorsed outcome-oriented measures that are relevant to almost all clinicians, regardless of specialty. The core set should focus on patient experience and engagement, outcomes related to care coordination like readmissions, measures of important safety complications, and measures of population and preventive health. The core set should also include patient-reported outcomes and other key outcomes; the relevant outcomes will vary by condition. While more endorsed measures in these areas are needed, many measures have been endorsed (e.g., patient experience measures, surgical and hospital complications, outcomes for common ophthalmologic procedures, etc.).  

(2) Align measures across multiple programs, to reduce administrative burdens and achieve greater impact. For example, clinician quality measures for Meaningful Use, the Physician Quality Reporting System, and Value-Based Payment Modifiers should be as consistent as possible; the equivalent programs for hospitals and other providers could also be aligned. Integrating these multiple payment adjustments into a simpler, more comprehensive system like a case-based or person-based payment as I have described (Recommendation #2) would help achieve this goal. Further steps toward measure alignment in public programs would reduce the cost and complexity of the endorsement and adoption of meaningful measures.

(3) Provide a lower-cost pathway for promising but less-developed measures to transition into more widespread use and NQF endorsement. In particular, the Clinician MAP in some cases has supported the use of measures that are not yet endorsed, if they have begun to be used, appear likely to meet the key criteria for endorsement, fill an important gap in the available quality measures, and they are expected to be submitted for endorsement. Such measures could be used for quality improvement (as in the PQRS, in which measures have not been publicly reported) as experience accumulates before inclusion in public reporting or used to adjust payments.

These principles, along with continuing support for the measure endorsement process, could help promote the more rapid development and endorsement of high-priority, outcome-oriented measures, reduce the complexity of measure development, and provide a means for refining key measures and demonstrating their effectiveness. By providing a more predictable pathway toward using non-endorsed measures where endorsed measures are not yet available, this approach would likely generate more private interest and support for the measure development process, thereby limiting the need for public funding. By focusing on the most important patient-level measures, it is likely that health care providers and organizations will continue to develop and adopt more sophisticated internal performance measurement “dashboards” to back up these key measures.

4. Support collaborations to implement quality measures using existing and emerging electronic data systems.

Even with payment reforms to support the development and use of endorsed, outcome-oriented quality measures, getting them into effective use to achieve quality improvements will still be challenging for many providers and their partners in improving care, such as electronic health record vendors, and insurers and other organizations providing decision support services. Effective implementation of quality measures – obtaining accurate data needed for performance measures, doing so in a timely and reliable way, and finding ways to improve on performance while still paying attention to all of the other pressures of clinical practice – is difficult and occupies much of the effort of quality improvement collaborations around the country. A number of steps in public programs could make quality measurement and quality improvement easier for providers.

Quality measures should be designed so that they can be implemented from data systems used in the actual delivery of care. This is not only less costly than requiring providers to do after-the-fact chart abstractions or other data collection that can distract from a direct focus on meeting the needs of their patients. It would also help make sure providers know where the gaps in quality are, so they can take more timely and informed steps to close the gaps, rather than getting a surprise after the fact when it is too late to help patients or perhaps even to correct e