Measuring the Quality of Care in Different Settings

Abstract

As HCFA initiates the next generation of health plan performance measures, the agency must address challenges associated with measuring the quality of care in all of the settings in which Medicare and Medicaid beneficiaries obtain care. One such challenge will be to integrate health plan performance measurement and health care quality measurement initiatives, which have been proceeding separately. Of equal importance is the challenge to improve coordination across the diverse, setting-specific quality measurement initiatives now in various stages of development or implementation by HCFA. Finally, HCFA must address the challenge of improving the collection, reporting, and analysis of data needed for health care quality measurement. This article describes these challenges and suggests steps HCFA might take in addressing them.

Introduction

The science of measuring health plan performance and health care quality has advanced rapidly over the past 10 years and is now commonly applied in monitoring systems (Brook, McGlynn, and Cleary, 1996; Eddy, 1998). But despite what can only be described as a plethora of burgeoning measurement initiatives, health care purchasers and consumers still lack answers to their questions about the quality of most types of care furnished by particular health plans and providers. At the same time, providers and plans increasingly decry what they see as unduly burdensome new data collection and reporting requirements levied by purchasers, private accrediting bodies, and regulators, and often do not see the benefits of quality measurement for themselves and their patients. Some refocusing and redirection of today's quality measurement efforts is needed to ensure that future initiatives are better able to serve the needs of all stakeholders.

As a dominant purchaser of both health care and health benefits and as a de facto Federal regulator of the U.S. health system, HCFA plays a major role in defining and implementing systems for measuring and reporting on health care quality. In fact, HCFA has taken the lead in establishing quality measurement systems to evaluate the performance of several types of health care providers and has ensured that the predominant health plan performance measurement systems take into account the health care quality concerns of the Medicare and Medicaid populations. HCFA's strategy and future activities, therefore, will do much to shape the next generation of health care quality measures.

This article begins with an overview of HCFA's strategy for measuring health care quality and also describes three key challenges the agency faces as it builds a sound infrastructure for assessing, assuring, and improving the quality of care furnished in all settings. The article offers a number of suggestions for HCFA to consider as it develops new quality measurement initiatives designed to meet those challenges.

HCFA'S Quality Measurement Strategy

HCFA has determined that measuring the quality of care provided to Medicare and Medicaid beneficiaries is an essential step toward meeting its goal of improving the care beneficiaries obtain (Gagel, 1995). The agency has, therefore, developed a strategy to create health care quality measurement tools and information systems that support improvement across the full range of Medicare services and most Medicaid services (Jencks, 1995). In implementing this strategy, HCFA has proceeded along two fronts.

First, it has worked to create health plan performance measurement systems that can be used to evaluate the quality of care provided in Medicare and Medicaid managed care plans, as well as in plans for special populations. Examples of performance measurement systems include:

• The Medicare and Medicaid components of the Health Plan Employer Data and Information Set (HEDIS^®), which provide measures of the underuse of preventive services and a global measure of Medicare beneficiary outcomes, known as the Health Outcomes Survey.

• The Consumer Assessment of Health Plans Survey (CAHPS^®), which permits evaluation of health care quality from the consumer perspective.

• The Outcome-Based Continuous Quality Improvement System for the Program of All-Inclusive Care for the Elderly (OBCQI for PACE), which is being designed to assess a variety of health care outcomes relevant to the nursing home eligible population enrolled in the program.

Secondly, HCFA has worked to develop health care quality measurement systems that focus on particular service settings. For a small minority of provider types, quality monitoring systems have been developed, tested, and instituted. These include the nursing home quality indicator system, which uses patient assessment data from the minimum data set (MDS) to evaluate long-term care (LTC), and the outcome-based quality improvement indicators for home health care, developed for use with the Outcome and Assessment Information Set (OASIS) (Zimmerman et al., 1995; Shaughnessy et al., 1994). For a somewhat larger minority of provider types, such as renal dialysis centers and intermediate care facilities for persons with mental retardation, clinical performance measurement systems are now in development (Health Care Financing Administration, 1999). For most other types of providers, including acute-care hospitals, implementing standardized systems for health care quality monitoring at the facility level is still a longer-term goal.

To date, health plan and provider-oriented quality measurement initiatives have not been particularly well coordinated in their design or implementation in the Medicare and Medicaid programs. For example, most health care quality measures included in HCFA's health plan performance measurement systems have not yet been successfully applied for use in the traditional Medicare program. Similarly, setting-specific quality indicators have not yet been integrated into health plan performance monitoring systems. Furthermore, the setting-specific measurement initiatives vary among themselves in scope and orientation. The result of this lack of coordination has been to limit the utility of the information produced by the measurement systems and to fuel concerns by health plans, providers, and policymakers about equity of oversight, particularly regarding the differential burden associated with meeting new data collection and reporting requirements.

Measuring Quality of Care in Different Settings

As HCFA initiates the next generation of health plan performance measures, the agency must address challenges associated with measuring the quality of care in all of the settings in which Medicare and Medicaid beneficiaries obtain care. Among those challenges are three that are particularly critical:

• Reconciling ongoing efforts to monitor health plan performance with those designed to evaluate the quality of care providers furnish.

• Improving coordination of the diverse, setting-specific quality measurement initiatives now in various stages of development or implementation by HCFA.

• Reducing the burden health plans, providers, and beneficiaries bear in collecting, reporting, and analyzing data needed for health care quality measurement.

Reconciling Plan Performance Measurement with Health Care Quality Measurement

One key challenge for HCFA will be to integrate health plan performance measurement and health care quality measurement initiatives. To meet needs for comparative information on quality in traditional Medicare and Medicare+Choice, the agency needs to ensure that new health care quality measures are implemented in such a way as to provide information relevant to health plan performance, as well as provider performance. It must similarly ensure that quality measures developed as part of health plan performance initiatives are applied to assess the care obtained by beneficiaries through the traditional, fee-for-service (FFS) Medicare and Medicaid programs.

Failure to meet this challenge will mean continued limits on the utility of information developed through performance measurement as well as continued concerns about equity of oversight between providers participating in the traditional Medicare program and health plans participating in Medicare+Choice. Equally important, failure to reconcile measurement systems is likely to result in even greater data collection and reporting burdens for providers who need to supply information to meet both types of requirements.

HCFA's original decision to proceed with separate development of provider- and plan-oriented quality measurement initiatives is understandable, given the initial context for quality and performance measurement and technical constraints. The initial decision may have reflected the following considerations, among others.

A first consideration is the evolution of thinking about accountability for performance. The notion of demanding accountability for health care quality arguably grew as a result of health care purchasers' relationships with managed care plans. HEDIS^®, therefore, represented the first standardized set of measures of health care quality to obtain widespread use and was applicable only to managed care plans. Efforts to develop quality measurement and reporting systems to track provider performance are relatively new.

A second consideration is the complicated relationship between health plan performance and health care quality. From one perspective, health care quality is but one of several aspects of health plan performance—along with such considerations as financial stability and member relations—that is of potential interest to purchasers and other stakeholders. From another perspective, health plan performance is but one of several potential determinants of health care quality, together with factors such as practitioner skill, facility characteristics, and patient compliance.

A third consideration is the difference between HCFA's role in the traditional Medicare program versus the former Medicare risk (managed care) program. In the traditional program, HCFA is a purchaser of services, and the health care provider is the party responsible for furnishing them. Under the risk program, however, HCFA purchased managed care. The health plan was accountable for providing this product, which was presumed to offer greater value than the provision of services on an uncoordinated basis.

Yet a fourth consideration relates to technical differences between traditional Medicare and the former risk program. HCFA has traditionally had very different types of data available for oversight of the traditional Medicare and managed care programs. In particular, the agency lacked information about services provided to beneficiaries who received care under risk arrangements. Under traditional Medicare, claims data provided some such information. However, with the establishment of patient encounter data reporting requirements under Medicare+Choice, established for purposes of risk-adjusting payments, comparability of administrative data has increased.

Maintaining separate quality measurement systems for managed care plans and providers participating in the traditional Medicare and Medicaid programs makes less sense now than it once did. The rationale for maintaining disparate performance measurement systems has dissipated greatly in recent years, while the need for integrated measurement systems has increased.

One important, recent change has been growth in the range of health plan choices for Medicare beneficiaries. This growth calls into question the value of distinguishing quality measures for use in health plan performance monitoring from those applied to evaluate providers. In Medicare, for example, quality measures designed to assess care provided either under traditional, FFS arrangements or tightly managed health maintenance organizations participating in the risk program must now be expanded or revised to allow for Medicare+Choice program participation by more loosely organized preferred provider organizations and private FFS plans. Increasingly, what is needed are measures of quality that can be applied at the health plan or the provider level as appropriate, irrespective of how care is paid for or organized.

HCFA's efforts to institute performance measures for the PACE program exemplify the problems with developing separate, uncoordinated performance measurement systems for health plans and providers. PACE is designed to offer integrated delivery and financing of primary, acute, and LTC services for a frail (nursing home eligible) population. Because 18 of the 25 PACE sites are licensed as home health care agencies in the States in which they operate, these sites are now evaluated using OASIS-based quality and performance measures. However, HCFA is currently in the process of developing a performance measurement system designed specifically to address PACE patient care issues. It is important that these provider-and plan-oriented performance measurement systems be coordinated with one another, to avoid duplication of effort, potentially conflicting reporting requirements, or unnecessarily burdensome reporting requirements for those PACE sites that are also designated as health care providers.

Improving Coordination of Setting-Specific Measurements

A second challenge HCFA faces in instituting performance measures that represent the full spectrum of beneficiary care is to improve coordination of setting-specific measurement initiatives. This is particularly important for those service settings in which patient mix, conditions treated, and services provided are comparable. Developing separate setting-specific quality measures and information systems to support measurement creates a lack of comparability across settings that diminishes the utility of the information generated by the measures.

The settings in which beneficiaries obtain services change over time. Furthermore, growth over the course of the recent decade or so in the number of service settings—particularly, though not exclusively, in the subacute and post-acute care arenas—means that similar patients are increasingly obtaining similar care for similar conditions in more than one type of setting. For example, certain gastrointestinal endoscopy services are provided to Medicare beneficiaries in outpatient departments, ambulatory surgery centers, and physicians' offices, as well as in the acute-hospital inpatient setting (Medicare Payment Advisory Commission, 1999). Similarly, rehabilitation care is furnished in a variety of settings, including rehabilitation units of acute-care hospitals, rehabilitation hospitals, and skilled nursing facilities. Developing separate and different quality measures for each type of care furnished in each setting limits the ability to compare how patients fare across different settings.

Burden of Data Collection, Reporting, and Analysis

A third challenge HCFA faces systems is minimizing the data reporting requirements associated with quality and performance measurement systems. Addressing this challenge is essential if the agency is to obtain and maintain support from providers. Reducing burden also offers a potential advantage in freeing up resources that could be spent for quality improvement purposes.

Quality measurement systems can be implemented without developing new data reporting requirements, where claims or other administrative data can serve. However, claims data have limitations as input for quality measurement. Very often, information from patient medical records, patient assessments, or patient survey data must be used. Thus, implementation of quality measurement systems often entails new data reporting requirements for providers or plans. Sometimes these reporting requirements pose burdens for patients or health plan members as well.

For most types of health care providers, HCFA has not yet implemented quality measurement and reporting requirements, but already providers are feeling pressured by health plans, private accrediting bodies, and others to produce data to document quality and performance. Health plans, at present, bear much of the direct cost associated with performance measurement, particularly because HEDIS^® reporting requirements call for submission of measures—processed, rather than raw, data—which necessitates investment in both data collection and analytic support staff. One analysis, reported by Eddy (1998), found that the cost to health plans of complying with an earlier version of HEDIS^® ranged from $20,000 to $700,000 per measure. Providers may, in principle, support the notion of accountability for performance, particularly as compared with the alternative—structural requirements, specifying how care is to be organized or delivered, which are seen as more prescriptive and constraining. However, additional requirements for data reporting that are perceived as duplicative or lacking value to the provider and patient are unlikely to gain acceptance.

Building A Comprehensive Quality Monitoring System

Steps HCFA might take in continuing its effort to build a comprehensive quality monitoring system for Medicare and Medicaid are described in the following section. Ways in which taking such steps could help to address the challenges previously described are also discussed.

Establish a Conceptual Framework for Measurement

As HCFA moves forward to initiate the next generation of performance measures, it could greatly benefit from establishing a common conceptual framework for quality measurement. Such a framework could serve as a roadmap to guide the agency's work to develop measures representing the full range of services used by beneficiaries. It would be valuable in establishing priorities among potential projects, reducing duplication of effort, and promoting an understanding of how multiple efforts work together to achieve HCFA's goals for quality measurement.

A conceptual framework for quality measurement would provide answers to a number of questions that must be addressed in designing any measurement initiative, such as:

• Who will use the information generated and for what purpose?

• How is quality defined for purposes of measurement?

• What aspects of quality will be measured?

• What types of measures will be used?

• What is the appropriate scope of measurement?

HCFA has addressed these questions on a case-by-case basis, by deciding on and carrying out the current quality measurement efforts. However, articulating a conceptual framework that could serve as a reference for future decisionmaking would support greater commonality across the multiple initiatives that are required to measure quality across the full range of beneficiary care.

HCFA's measurement strategy incorporates the Institute of Medicine's definition of health care quality (Jencks, 1995). Quality of care is “…the degree to which health services for individuals and populations increase the likelihood of desired health care outcomes and are consistent with current professional knowledge…” (Lohr, 1990). This definition can serve as a guide in determining which aspects of quality are important to address in quality measurement initiatives. For example, HCFA might consider that accessibility and acceptability of care to patients and improvements in patients' health and functioning constitute the categories of “desired health care outcomes.” Similarly, HCFA could look to standards defining underuse, overuse, and misuse of services to evaluate the extent to which health care is provided in a manner consistent with current professional knowledge.

HCFA's conceptual framework should next address the types of measures to be used. There is considerable debate as to the relative value of outcome, process, and structural measures (Brook, McGlynn, and Cleary, 1996). Outcome measures are attractive in that they permit direct evaluation of the subject of interest. However, their use in establishing accountability for quality is challenging in that many outcomes are subject to influence by factors outside the control of the health care provider or health plan and that risk adjustment techniques are inherently imperfect (Iezzoni, 1997). The use of process and structural measures must therefore be considered, but, similarly, the link between many such measures and relevant outcomes has not been established. Use of evidence-based practice guidelines in measures development provides one way to address this concern. In fact, HCFA took this approach in developing clinical performance measures for dialysis centers based on guidelines from the National Kidney Foundation.

HCFA might choose to select measures of structure, process, and outcome to serve different purposes. For each type of care furnished to program beneficiaries, HCFA needs to identify the most important clinical and non-clinical outcomes. The next step is to identify the structures and processes associated with health care delivery that are most strongly associated with the outcomes of interest. For example, a variety of measures of appropriateness of service use or technical proficiency in administering care might be used if they were found to be important determinants of certain desired outcomes.

One important consideration will be to determine “how much” measurement is appropriate, given that resources expended for quality measurement are resources that cannot be employed for other purposes, such as patient care or quality improvement. HCFA could rationalize measurement efforts by limiting use of clinical performance measures to those that meet a designated threshold, such as having a large expected impact on beneficiaries' health. Siu and colleagues (1992) developed an approach that HCFA and its contractors could use to select among potential measures for use. This approach draws on data on (or estimates of) the burden of disease, such as the annual number of deaths associated with the condition, efficacy of available treatments, and the quality of care currently being provided. Use of such an approach would likely result in more extensive use of quality measurement for certain types of care, with minimal use of quality measurement for the types of care in which quality already approaches established benchmarks or fewer opportunities for improvements in beneficiaries' health are available.

Design Measurement Initiatives for Multiple Needs

Systems to measure the quality of care furnished in all of the settings that serve Medicare and Medicaid beneficiaries must be designed to be capable of meeting the multiple demands that will be placed on them, including:

• HCFA's needs as a health care regulator to ensure that providers furnish care that meets minimum standards for safety and quality.

• HCFA's needs as a value-based purchaser of care to assess health plans' performances in delivering high-quality care to enrollees.

• Providers' needs for information to direct and support quality improvement efforts.

• Consumers' needs for information to select health plans and providers.

Few health care quality measures will be able to address all of these needs, not to mention those of policymakers, researchers, private accreditation bodies, and others who are contributing to a growing national appetite for information on quality. But to increase efficiency and effectiveness of measurement, it is important that more quality measures and quality measurement initiatives be designed with multiple purposes in mind. Equally important is that quality measures and measurement initiatives be coordinated to fill existing gaps and to avoid duplication of effort.

For a variety of reasons, specific quality measures may not fill all needs equally well. For example, some measures may be too technical for use by consumers and patients lacking clinical expertise. Others may be too narrowly focused to be of interest to these end users. Certain outcome measures, such as global measures of health and functioning, may be too broad to be used by providers for quality improvement purposes. However, HCFA may not need to develop different measures to serve each purpose. For instance, measures that are too narrow or too technical for consumer use might be combined in an index or summary measure designed to address a particular aspect of quality in a broader way. HCFA could further such work by funding the development of index measures or other approaches for adapting quality measures to serve multiple purposes.

Another critical way in which needs for information on quality vary across key end users is in their different interests in the relevant unit of accountability. HCFA requires information about quality of care aggregated at the health plan level to assist in oversight and management of its contractors. Policymakers, on the other hand, might find information on the quality of care obtained by particular populations, such as disabled beneficiaries, beneficiaries residing in particular geographic areas, or beneficiaries enrolled in particular types of health plans, to be most useful for program decisionmaking purposes. Several key stakeholders are potential users of information aggregated at the facility or group practice level. Beneficiaries, many of whom do not appreciate the potential influence of the health plan on quality of care, may find provider-level information to have greater utility in their own decisionmaking (Jewett and Hibbard, 1996). Providers, too, are likely to find information that allows for comparisons of their performance with that of their peers or established benchmarks to be of greatest use for quality improvement purposes. Finally, HCFA also increasingly requires information at this level as it endeavors to move its Medicare and Medicaid program conditions of participation for health care providers away from requirements designed to ensure the capacity to provide quality care in favor of those relating to demonstrated performance (Medicare Payment Advisory Commission, 2000b).

Standardize Quality Measures and Underlying Components

One much-needed step to increase synergy across different quality measurement initiatives is to standardize certain key health care quality measures and underlying components of measurement (such as data collection tools). Standardization could assist greatly in efforts to coordinate different quality measurement initiatives and to generate comparable information through them. It could also foster the efficient use of resources by directing them to new development, rather than continually reinventing measures and methods of measurement. Perhaps most importantly, standardization could assist in reducing the multiple, conflicting demands for information faced by health plans and providers that draw on resources that could otherwise be expended to improve quality of care (Quality Commission, 1998). Areas that could benefit from standardization include metrics (such as rating scales), definitions and terminology, data collection methods and tools (such as survey questionnaires and patient assessment instruments), as well as specific types of measures with broad application, such as measures of satisfaction, health and functional status, and patient conditions.

An example of a sector where there is a lack of this type of standardization is subacute care, where HCFA has made progress in implementing quality measurement systems. In the LTC and post-acute care arenas, HCFA is building new systems for prospective payment and quality measurement that employ data from standardized patient assessment instruments. LTC facilities report the MDS, which is used to determine skilled nursing facility payments under Medicare and to evaluate the quality of care furnished by a facility. Home health agencies similarly report the OASIS for use in determining Medicare payment amounts and assessing quality of care. The information collected by these tools is quite different, not only in terms of the types of information collected, but also in the way similar items are framed and in the rating scales used.

The items in the MDS and OASIS relating to patient bathing status provide an illustration of such differences. They differ considerably in how they define bathing, what about bathing is of interest (documenting what actually occurs or the perceived ability to undertake the activity), and the number and nature of response codes. The MDS defines bathing as “how resident takes full-body bath/shower, sponge bath, and transfers in/out of the tub/shower (exclude washing of back and hair),” while the OASIS considers “patient ability to wash entire body (exclude grooming, washing face and hands only).” The MDS provides 11 response codes to the bathing item; 6 are for coding patients' bathing self-performance and 5 for coding staff-supported bathing activity. The OASIS offers seven response codes that range from full patient independence to complete dependence on another person for bathing.

Lack of comparability in patient assessment data collection across settings is problematic for several reasons. First, the differences limit the use of the data to make comparisons across settings, even in cases in which the patients and the care furnished may be comparable in important respects. They also create different data reporting burdens on providers and patients across settings and serve as a limiting factor in moving toward more coordinated delivery and payment arrangements for post-acute care.

A number of entities could play a role in improving standardization. For example, the Agency for Healthcare Research and Quality (AHRQ), formerly known as the Agency for Health Care Policy and Research, could identify candidates from among the various components of measurement that would benefit from standards development, convene experts and interested parties to assess options, and fund research and development in areas identified as requiring additional work.¹ The National Quality Forum, a private-sector group formed in response to the recommendations of the President's Advisory Commission on Consumer Protection and Quality, expects to serve a similar role by selecting and promoting use of core sets of quality measures in particular areas. The organization has announced plans to begin its work in this area by developing a core set of measures for assessing the quality of acute-hospital care (National Quality Forum, 1999).

HCFA could promote standardization in a number of ways. One is to work closely with AHRQ and the National Quality Forum to ensure that their efforts benefit from the lessons learned through Medicare and Medicaid quality measurement. This task is already being accomplished, in part through HCFA's representation on the Board of Directors of the National Quality Forum² and through HCFA's involvement in the Administration's Quality Interagency Coordination Taskforce, a group developed to coordinate the activities of Federal Government agencies responsible for purchasing, providing, regulating, or studying health care services. In addition, HCFA could take steps on its own to standardize the tools it uses in Medicare and Medicaid quality measurement. A key area for such work is in the patient assessment data collection tools used for payment and quality measurement in post-acute care (Medicare Payment Advisory Commission, 2000a).

Organize Measures Development Around Types of Care

HCFA should focus new quality measurement initiatives on particular types of health care used by program beneficiaries, rather than on particular settings in which care is furnished or on the arrangements under which it is financed and delivered. This approach would maximize the utility of the information generated by the measures by ensuring comparability and would provide flexibility for the measures to be adapted to a continuously evolving health care financing and delivery environment. Core measures for each type of care should apply across all of the relevant sites of service and across health plans of various types, including traditional Medicare. In designing these measures, HCFA should take into account the various settings in which the care is provided, so as to identify sources of data that could be collected comparably across settings.

Apply Measures Consistently

Once HCFA has identified core measures of quality for each of the types of care beneficiaries receive, the agency must strive to apply those measures uniformly across all of the relevant sites of service, and in all of the available financing and delivery arrangements. Uniform application is a necessary first step; however, alone it will not ensure comparability of the information generated through performance measurement. An important, outstanding technical challenge relates to the need for risk adjusters to account for relevant differences in the mix of patients or enrollees. Addressing the challenge should be a key element in HCFA's research agenda.

Rationalize Data Collection

HCFA has taken some steps to ease the burden associated with new data reporting requirements to support quality measurement, but must make it a priority. One approach used by HCFA has been to make available public domain software designed to assist in standardized data collection and reporting of information. The agency has released software for reporting the MDS (used to determine payments and measure quality of nursing facility care) and the OASIS (used to determine payments and measure quality of home health care).

Another critical step is to reduce data reporting requirements. This could be accomplished by reviewing existing requirements to identify and eliminate items that are not needed for payment, quality monitoring, or other administrative purposes. To the extent possible, data collection should be designed as an integral part of the delivery of high-quality health care services, rather than as a secondary consideration. In addition, HCFA should examine whether collecting some information from only a subset of patients or providers could reduce certain data requirements. Finally, HCFA should determine whether data reporting requirements could be reduced by substituting use of information that can be generated with a minimal burden on plans, providers, or patients. For example, HCFA might seek to make additional use of patient medical records data abstraction or administrative data, such as claims.

Yet another important step that HCFA could take to reduce the burden associated with meeting needs for data would be to encourage health plans and providers to adopt automated clinical information systems. The value of such systems for reducing the burden of information collection and reporting has been noted by numerous industry experts and other observers (Schneider et al., 1999).

Analyze Variations in Quality

Because a variety of factors can and do influence health care quality and because the relevant unit of analysis varies across the many uses to which performance data are put, HCFA should design its quality measurement initiatives to permit analyses of variations in quality across health plans, providers, and beneficiaries. HCFA should strive to develop and report information about the quality of care furnished by particular providers and health plans (including traditional Medicare) and obtained by program beneficiaries with various characteristics. This reporting will allow HCFA and other stakeholders to better understand the nature of any quality problems or failures to achieve goals for improvement. Is large variation in provider or health plan performance a contributing factor? Or does variation in the quality of care obtained by different groups of beneficiaries appear to be relevant? Understanding the nature of the problem is likely to be helpful in determining which avenues for improvement are most promising.

To maximize analytic flexibility and minimize the burden of data processing on providers and plans, HCFA will need to structure any new data reporting requirements to collect data at the individual observation level. In general, this implies that providers should report unprocessed data that can be aggregated in various ways to assess the performance of health care providers, health plans, and the Medicare and Medicaid programs overall, as well as the quality of care obtained by various subgroups of beneficiaries. It also implies that the data need to be readily linked to information on the characteristics of patients, providers, health plans, and payers. The health plan, therefore, may or may not be the best locus of accountability for data reporting requirements pertaining to quality of care. In terms of measuring quality of care, it might be better to think about health plan performance measurement as just one of many relevant ways to analyze available data, rather than as an end goal for health care quality measurement.

Conclusion

HCFA has played an important role in health plan performance and health care quality measurement. The agency also serves as a model for other large purchases of health care and health benefits in demonstrating ways in which this type of information can be used to benefit program beneficiaries. Although sizable challenges remain to be addressed, the remarkable progress that has been made in the short period of time since HCFA began its efforts to measure quality systematically suggests the likelihood of significant and speedy progress in addressing those challenges.