On August 22, 2006, President Bush issued an Executive Order—Promoting Quality and Efficient Health Care in Federal Government Administered or Sponsored Health Care Programs—calling on all federal agencies and those who do healthcare business with the government to engage in collaborative efforts to incorporate the cornerstones of valuedriven healthcare (Table 1). The 4 cornerstones of value-driven healthcare are1:
- Interoperable health information technology
- Measure and publish quality information
- Measure and publish price information
- Promote quality and efficiency of care.
Local Collaboration Key to Value-Driven Healthcare
Department of Health and Human Services Secretary Michael O. Leavitt has embarked on a campaign to make these 4 cornerstones a reality. A key piece of this campaign, known as the Value-Driven Health Care Initiative, is encouraging the public and private sectors to work collaboratively at the local level. Regional or local public–private collaboration among payers, health plans, providers, and consumers is essential to the success of this initiative.
Healthcare systems are local. The different environments for healthcare delivery differ in the range of populations served, resources available, and in the characteristics of the local marketplace. Thus, broad-based efforts to improve quality of care need to be driven by local leaders who represent the various stakeholders and who are willing to pool their resources toward achieving common goals. In addition, quality initiatives need to allow for local input so that local market conditions can be taken into consideration.
Although fostering the growth of locally led collaboratives is crucial to the success of the Value-Driven Health Care Initiative, national coordination is also needed. Patients should expect and be able to receive good quality care regardless of where they live. Similarly, all providers should be able to meet certain basic, agreed on standards of care regardless of where they practice. The need for national coordination is particularly salient for the Medicare program, which provides health insurance coverage for approximately 43.9 million beneficiaries across the United States.2 Because of Medicare's national scope, the Centers for Medicare & Medicaid Services (CMS) needs to ensure that Medicare beneficiaries residing in Florida can receive information on quality of care that is consistent with the quality-of-care information provided to Medicare beneficiaries residing in California.
A National Approach to Physician Performance Measurement
A key piece to measuring and reporting quality information is measuring and reporting information on physician performance. Typically, efforts to measure and report on physician performance have been done in a piecemeal manner.3 Physicians, who usually see patients covered by a variety of public and private insurers, often receive information about their performance only as it relates to patients covered by a single insurer. Since physicians typically receive information on only a subpopulation of their entire patient population, physicians typically have no idea how their practice, as a whole, is performing against various quality measures. As a result, consumers looking for information on physician performance to help them select a physician or to review treatment options typically do not have a comprehensive picture of a physician's performance either.
Often, the insurance plans measure the physicians on a different group of measures as well. For example, a physician could potentially receive a report from one entity that only looked at the preventive services provided by the physician, while a second report from a different entity looked at the services the physician provided to his or her patients with diabetes. Even if the quality measures on the 2 reports appeared to be the same, it is possible that the methods used by the 2 entities to generate the physician's performance rate may have differed. For example, a physician receives 2 reports from 2 entities notifying the physician of the percentage of his or her patients with diabetes whose hemoglobin A1C levels are under control. However, the threshold used for determining whether a patient's hemoglobin A1C level is under control may differ between the 2 entities. Consequently, the information that the physician received from the 2 entities would not be comparable. Since the various reports provide the physician with inconsistent information that may even be conflicting at times, it becomes difficult for the physician to act on the information provided in these various reports to improve the quality of care he or she provides.
CMS' BQI Project
In an attempt to explore how to provide physicians who treat Medicare beneficiaries with more meaningful, comprehensive information of their performance, CMS launched a new project in late 2006. In keeping with the central principles of the Value-Driven Health Care Initiative, this project, known as the Better Quality Information (BQI) to Improve Care for Medicare Beneficiaries Project, also leverages the experience of local multistakeholder collaboratives to explore a national, coordinated approach to physician performance measurement.
CMS contracted with the Delmarva Foundation for Medical Care (CMS's Quality Improvement Organization for Maryland) to provide overall project administration and management for the BQI Project and to conduct analyses on the Medicare administrative data. Delmarva awarded subcontracts to 6 local collaboratives, or pilot sites (Table 2), to receive, aggregate, and analyze Medicare administrative data along with their existing datasets. Delmarva and the 6 BQI pilot sites were tasked with testing:
- Methods to aggregate Medicare administrative data with data from commercial health plans and, in some cases, Medicaid
- The use of the aggregated data to calculate and report on quality measures for physician groups and, in some cases, for individual physicians practicing in each of the pilot sites' local communities.
The BQI Project will be completed at the end of October 2008. The following criteria were used to select the pilot sites:
- Strong physician leadership engaged in creating the coalition
- Multiple employer participation
- Experience in measuring and aggregating physician level data
- Experience providing feedback reports to physicians
- Presence of a public website for consumers to access relevant information
- Demonstrated capacity and interest to accept additional tasks
- Willingness to work with a viable health information network if available
- Capability and infrastructure to begin data collection within a short time frame.
The 6 local collaboratives selected to participate in the BQI Project each brings a unique set of characteristics and experiences that have resulted in some differences in approach to implementation. One pilot site, for example, receives all payer data (including Medicare and Medicaid) directly from its member physician practices, while another pilot site relies solely on administrative data received from participating health plans. Similarly, some pilots have access to clinical data, including laboratory results, whereas other pilots only have access to administrative data, such as claims. Some pilots have a wealth of experience in publicly reporting healthcare performance information. Other pilots have no experience with public reporting. Some pilot sites are statewide in terms of population covered and scope, whereas others cover specific areas within the state in which they operate. We believe, however, that the uniqueness of each pilot site will prove beneficial for gathering lessons for the development of a national strategy for physician performance measurement.
Goals for the BQI Project
The BQI pilot sites all have had some previous experience linking data from different sources together. In some cases, a pilot site may even have previous experience using the linked, or aggregated, data to produce healthcare performance information on the providers in their respective communities. Before their participation in the BQI Project, however, few of the pilot sites had the ability to incorporate Medicare data with the pilot sites' other data sources. In seeking to test methods for aggregating Medicare administrative data with other data sources to produce a more comprehensive picture of the quality of care provided by physicians to Medicare beneficiaries through the BQI Project, CMS has provided many of the BQI pilot sites the opportunity to incorporate the experiences of Medicare beneficiaries into their local efforts for the first time.
For many BQI pilot sites, the BQI Project represents the first time that Medicare data have been combined with other data sources for the purpose of generating meaningful information on physician performance. In most cases, the addition of Medicare data represents a significant addition of information about a physician's practice that many of these local collaboratives had been missing. In 2005 alone, approximately 33 million people received a reimbursed service under Medicare fee-for-service, including roughly 32.7 million people who received reimbursable physician services.4
The BQI pilot sites will be working until the end of October 2008 to aggregate Medicare administrative data (eg, Medicare inpatient claims, outpatient claims, carrier claims, enrollment databases, and provider databases) with data from other payers to produce quality measure results. Since the BQI pilot sites will be using data from multiple payers, including Medicare, to produce these measurement results, they will be able to provide a more comprehensive picture of the quality of services being provided by physicians in their communities who treat Medicare beneficiaries. Each BQI pilot site will be reporting on 12 measures. The measures selected by each BQI pilot site are derived from the AQA Alliance (formerly the Ambulatory Care Quality Alliance) starter set of measures for physician performance (Table 3)5 and other quality measures endorsed by the National Quality Forum. In selecting measures for the BQI Project, a number of considerations were taken into account, including the data sources available to each BQI pilot site, to calculate the measure as well as the measure's relevance to the Medicare population, and the local community.
Before CMS can develop a strategy for measuring the performance of physicians around the country using Medicare administrative data, a number of methodological questions need to be answered. Each BQI pilot site will be aggregating the Medicare data with their own data sources to generate multipayer physician level and/or physician group level performance measurement results on 12 measures, at least 3 times. During the course of conducting this work, the BQI pilots will be capturing and sharing the lessons learned and challenges around the numerous methodological questions that need to be answered, such as measure selection, methods for assigning accountability for a patient's care to a physician, or physician attribution, identifying how individual physicians practice as groups, identifying how information about group composition is captured on payers' administrative data, understanding Medicare data, and the reliability of measure results.
The BQI project also aims to:
- Provide beneficiaries in these 6 pilot sites with healthcare performance information on the physicians who treat them to help select physicians and make treatment choices
- Provide performance information to the physician groups and/or physicians who treat these beneficiaries, which can be used by the physician groups and/or physicians to improve the quality of care they provide.
Thus, the BQI pilot sites will also be capturing and sharing lessons learned around mechanisms for sharing the measurement results with patients as well as their physicians to the extent feasible.
The results from CMS's BQI Project and the shared experiences of the BQI pilot sites could be used to guide future efforts for incorporating Medicare administrative data into the process of measuring the performance of physicians who treat Medicare beneficiaries. Specifically, the lessons learned from the BQI Project could be used to help inform CMS how Medicare administrative data can be aggregated with other sources of data to produce quality measure results that provide a more comprehensive picture of the quality of services being provided by physicians to Medicare beneficiaries across the country.
- US Department of Health & Human Services. Value-Driven Health Care. http://www.hhs.gov/valuedriven. Accessed May 2, 2008.
- US Department of Health & Human Services. Centers for Medicare & Medicaid Services. 2007 CMS Statistics. June 2007. http://www.cms.hhs.gov/CapMarketUpdates/Downloads/2007CMSstat.pdf. Accessed May 5, 2008.
- AQA Alliance. AQA Pilot Information. Ambulatory Care Quality Alliance Announces Pilot Project: Six Sites Will Combine Public and Private Data on Physician Practice. March 1, 2006. http://www.aqaalliance.org/pilot.htm. Accessed May 30, 2008.
- US Department of Health & Human Services. 2007 CMS Statistics. June 2007. http://www.cms.hhs.gov/CapMarketUpdates/Downloads/2007CMSstat.pdf. Accessed May 5, 2008.
- AQA Alliance. Recommended Starter Set of Clinical Measures for Physician Performance. http://www.aqaalliance.org/files/RevisedStarterSetApril2006.doc. Accessed May 12, 2008.