for any areas that fall short of JCAHO's requirements. Although JCAHO refuses to accredit an institution with contingencies above a certain threshold, a hospital with fewer contingencies receives accreditation contingent on correcting the deficiencies within a period of time specified by JCAHO. Accredited hospitals receive a certificate for posting, and, upon request, JCAHO makes publicly available information on whether a hospital passes or fails to gain accreditation. JCAHO does not divulge, however, whether a hospital has any contingencies and, if so, the number of or reason for any contingencies (330). Individual hospitals may make that information available, but their policies vary.

Under this option, a hospital that chose to use JCAHO to satisfy conditions of participation

Photo credit: George Washington Medical Center Physicians who practice in the area of their training are likely to deliver higher quality care than physicians without special training in the area.

would have to disclose any JCAHO contingencies, by number and area. HCFA or PROS could compile the information from all hospitals and annually release it, perhaps along with hospital mortality rates and any other similar information. Purchasers of health care, including individuals, third-party payers, and employers, could then question physicians and hospitals about the status of any contingencies, especially in areas related to procedures for which they were seeking care. Precedent supports public disclosure of this information. As part of the licensing process, 38 States require hospitals to give their State health agencies copies of JCAHO survey reports (48). At least one of these States, New York, will release copies of JCAHO reports to the public. The Veterans Administration also has a policy of making information about contingencies and accreditation available on request (146).

Under this option, HCFA could require other quality-related information to be released, such as mortality rates for all the hospital's patients, HHS sanctions recommended by PROS, disciplinary actions by State medical boards, the total volume of certain services, and the training and experience of physicians performing specialized procedures. PROS or HCFA regional offices could gather information on selected indicators and make it available upon request or on some periodic basis. Alternatively, HCFA could rely on State health departments to compile and distribute the information, perhaps as a condition of Medicaid funding.

Like most other possible indicators of quality, JCAHO contingency scores have not been validated for their association with the quality of hospital care, as measured by the process or outcome of patient care. If hospitals were required to disclose the scores, HCFA and the media would have to advise consumers and the press on their appropriate use―i.e., as a guide for further questioning rather than the basis for any definitive conclusions about a hospital's quality. As discussed earlier, organized purchasers and consumer advocacy groups with experts on their staffs would be better able than individuals to interpret qualityof-care information and to exert leverage over providers to resolve any problems identified.

Option 8: Amend the Social Security Act to permit peer review organizations and the Health Care Financing Administration to disclose publicly information that identifies specific physicians.

During review of the medical care that physicians deliver to inpatients, PROS develop information related to the quality of individual physicians. Some information comes from reviewing a 3-percent random sample of the medical records related to Medicare discharges; other information comes from examining medical records selected because the PRO targets for review specific diagnoses, surgical procedures, or other areas. Although upon request PROs must disclose information that identifies hospitals, the Social Security Act forbids disclosure of information that identifies physicians (42 CFR 476.101, 104, 105, 1986 ed.). The only information identifying individual physicians that PROs and HCFA may make public is information on decisions to impose monetary sanctions or exclusions from the Medicare program.

Like the Federal Government, 6 of the 13 States that collect unique physician identification numbers (such as State license numbers) as part of their discharge data systems prohibit the release of physician-identified information.' To date, patient discharge data, with physicians identified, have been released only in Arizona, with no publication as yet (366). Only in Pennsylvania, where data collection began January 1, 1988, does legislation specifically mandate that data relating to the quality of individual physicians must be made available to the public (Pennsylvania Act 1986-89).

This option would change the Social Security Act to permit PROS and HCFA to make public information that identified physicians. Either in response to requests or on a regular basis, the PROS or HCFA could release, by physician, information such as mortality rates by procedure, the volume of certain services performed for Medicare beneficiaries, and the results of PRO re

Of the seven States that do not prohibit release of physicianidentified data, five have operational data collection systems: Arizona, Illinois, Nevada, Tennessee, and Washington. Pennsylvania and North Dakota have not yet implemented their systems (366).

views. If Congress adopted option 2, HHS could coordinate its selection of indicators to assess physicians with public release of data under this option.

Previous discussions about preserving the confidentiality of individual physicians have highlighted the conflict between public and providers' interests (451,579). Public disclosure might enable consumers and providers to identify poor-quality physicians and might prompt physicians and the hospitals where they work to improve their quality. Such disclosure, however, may also hurt the reputations and unfairly jeopardize the livelihoods of individual physicians. If physicians challenged disclosure of physician-identified data through the legal system, the judicial analysis would most likely weigh these as well as other interests.3

Technical problems, however, may overshadow the legal and philosophical issues. Given the current state of quality assessment, data, and the PRO process, statistics on individual physicians could mislead consumers and erroneously discredit physicians. That each physician has a much smaller number of patients and patient deaths than a hospital makes interpreting physician statistics more difficult. The deficiencies of current techniques to adjust for patient characteristics and other factors outside the providers' control apply to physician as to hospital care. But among physicians' much smaller numbers of patients, chance is much more likely than among hospitals to account for patient outcomes.

Even more basic are issues concerning the reliability of data. Current data collection systems. make it difficult for researchers or quality assessors to identify all the patients treated by a particular physician. Physicians typically use different billing codes for the different locations in which they practice, and a group practice often bills for the claims of all its physicians under one code. Only 13 of 28 States with systems to collect hospital discharge data impose on hospitals unique physician identifiers (366). Attributing patients to physicians poses another thorny technical problem. The hospital designates one person

For a full discussion of the legal implications of releasing physician-identified data, see J.B. Simpson, "Release of PhysicianSpecific Quality of Care Information: Legal Issues" (579).

as the attending physician, usually the physician who admitted the patient, but other physicians acting as consultants may play a major role in the patient's care.

To a large extent, the validity of HHS sanctions recommended by PROS against a physician or hospital stems from the multistep process used to arrive at these decisions. This process entails review of records by trained nurse reviewers, several physician reviewers, and PRO committees; the opportunity for the physician or hospital being reviewed to attend a hearing and to provide supplementary information; and finally, review of the case by the Office of the Inspector General of HHS (see ch. 6). Although low levels of interrater and intrarater reliability threaten the validity of the PRO process in particular and peer review in general, the due process accorded to physicians and hospitals under investigation increases one's confidence in the validity of the sanctions that do result. But the findings at interim stages during the PRO/HHS process lack whatever validity is conveyed by the entire sanctioning process.

To Disseminate Information to the Public

Although much information related to the quality of medical care is already in the public domain, many individuals and organizations do not know that it is available. Available information is often not in the right format or timely enough to influence decisions. To be incorporated into consumers' choices, information must be simple and accessible when people are making decisions. Furthermore, people require skills and social support to undertake what for many is new behavior, namely interacting with physicians and raising questions about quality. The options in this section, which could be undertaken separately or together, consider more efficient ways to disseminate information on quality of care to individuals and organizations.

Option 9: Establish a new office in the Department of Health and Human Services that would be responsible for disseminating information on the quality of medical care to individuals and organizations.

As information on the quality of care becomes increasingly available, one question that arises is how to encourage the most responsible and effective provision of such information. This option would establish an office in HHS to disseminate public information on quality.

Such an office would take an active role in informing people about available information and distributing information developed by Federal programs, such as Medicare and Medicaid. These tasks would entail more than communicating knowledge. The office would inform the public about possible differences in quality among providers and interpret information from quality indicators. Over time, the office could educate consumers about the skills necessary to put their new knowledge into practice. To convey information and to engender social support for questioning quality differences, the office would work with consumer groups, medical providers, employers, third-party payers, business coalitions, States, and the media.

With such activities concentrated in one office rather than dispersed throughout HHS, HHS would be better able to apply principles of health behavior and communication in educating the public to use information on quality. Consumers' belief in the reputation of a source increases their acceptance of information. To the extent that people trust the Government as a source of health information, disseminating information through an office in HHS would increase the likelihood that people would incorporate the information into their decisions. In addition, compared with private sources, the Government would be better able to provide continuous access to quality information. Credibility and accuracy could be increased by creating an expert advisory group to review information before it is disseminated.

Congress could expand the responsibilities of the office to include all or some of those outlined in previous options, such as coordinating research on quality assessment techniques (option 1), selecting quality-of-care indicators (option 2), conducting State and local outreach activities (option 4), developing uniform data requirements (option 6), and making publicly available information on Medicare and Medicaid providers (option 7).

Combining these activities would facilitate the development and implementation of a long-term strategy regarding information on the quality of

care.

Integrating and expanding these activities in a new office would require increased funding. It is also questionable whether a single office could have the experts needed to carry out such a wide range of responsibilities. Instead of creating a new office, Congress may wish to rely on existing public and private activities. Offices in HHS already undertake activities to publicize the availability of consumer information, and HCFA and the California PRO have released hospital mortality data. Several private organizations are disseminating information, and their level of effort appears to be increasing. In periodic and special publications, the Public Citizen Health Research Group has a long history of publicizing information related to the quality of physicians and hospitals. Broadcast and print media periodically gather information on indicators of quality and make it publicly available. For example, 1987 issues of Consumer Checkbook in Washington, DC, and the San Francisco Bay area amassed information on a range of possible quality indicators for local hospitals and analyzed how to use it appropriately. As part of their cost-containment efforts, some employers are making information related to providers' quality available to employees (253), and some private business associations are considering making information on physician and hospital quality available nationally (256). Such efforts, although limited to date, might expand as information on indicators of the quality of care becomes more generally available.

Most private groups, however, do not have the resources available to make information available for broad geographic regions, for a comprehensive set of indicators of the quality of care, or on a regular basis from year to year. As alternative approaches, the Federal Government could require State governments to disseminate such information as part of their participation in the Medicaid program or could enter into partnership with a private organization for this purpose.

A drawback of this option is the paucity of knowledge on how individuals and organizations

use available information, how information can most effectively be communicated, and how existing information affects hospital and physician behavior. One would expect greater insight into these matters would permit HHS to formulate a more effective dissemination strategy than would now be possible.

Option 10: Mandate and earmark funds for research and demonstrations on methods to disseminate information on the quality of medical care.

Although one purpose of providing information on the quality of providers' care is to help consumers make more informed choices of physicians and hospitals, no empirical work has addressed whether the availability of quality-of-care information influences consumers' choices of providers. This option would fund research and demonstrations to explore the effects of qualityof-care information on consumers' decisions. These projects could be funded either instead of or in conjunction with the dissemination activities outlined in option 9.

Possible topics for research under this option include how to use the media to present information on indicators of quality so as to influence consumer choices; what type of quality-of-care information consumers find most useful in making health care decisions; what formats are most useful for providing quality-of-care information; how information learned from marketing about attracting consumers' attention can be transferred

Photo credit: American Association of Retired Persons Little is known about how information on the quality of care can be most effectively disseminated to consumers.

to health care decisions; and other topics related to quality-of-care programs in workplace and community settings and in physician's offices. Such research would apply perspectives drawn from several disciplines and could use a variety of methods: policy review, consumer surveys, laboratory experiments, and field experiments in

CONCLUSIONS

Although the indicators of quality examined in this report do not give conclusive evaluations of a physician's or hospital's quality, individual and organizational purchasers of care could use several of the indicators as flags, to point out areas of concern that merit further investigation. Given the current status of the indicators evaluated, formal disciplinary actions by State medical boards provide the most valid information about poorquality physicians. In evaluating a specific physician or hospital, consumers would improve the validity of quality information if they combined the results of more than one indicator and drew information from more than a single year.

With regard to future policy in this area, those indicators that are already being used to evaluate the quality of care merit particular attention. Since governments and other entities are already disseminating hospital mortality rates, for example, the immediate task is to improve the underlying data and techniques for attributing death to prior medical care. Information on adverse events, HHS sanctions recommended by PROS, and physician specialization is also becoming generally available. Efforts to identify and improve the practices of poor-quality providers also deserve particular attention.

Although existing data sets do not allow routine evaluation of physicians' performance outside hospitals, promising efforts are underway in the United States and Canada to assess the quality of office practice across a range of medical conditions. Also promising, but not yet validated, are activities by several specialty societies to certify

naturalistic environments. Initial research could conduct surveys to ascertain the level of knowledge about the quality of care among the general population and specific subgroups. Currently, this gap in information inhibits the development of effective interventions, particularly for targeted populations.

the competence of physicians to perform certain procedures.

Even with valid indicators of quality that are feasible to develop, using such information to guide consumers' choice of providers represents only one approach or one part of an approach to select a physician or hospital. Consumers may also rely on a primary care physician for a referral to a specialist or a hospital, a strategy that individuals often adopt. In recent years, plans that provide comprehensive care to enrollees, such as HMOs and PPOs, have institutionalized the arrangement by linking each enrollee with a primary care physician to manage that person's care. Indeed, giving consumers information on the quality of care complements consumers' reliance on a physician for referrals, because better informed consumers are more likely to be able to communicate their preferences and concerns to physicians.

Informing consumers and relying on their subsequent actions should not be viewed as the only method to encourage hospitals and physicians to maintain and improve the quality of their care. Even well-informed lay people are unlikely to have sufficient technical knowledge to judge all aspects of quality and must continue to rely on experts to ensure the quality of providers. Some experts come from within the medical community and engage in self regulation, while others operate as external reviewers through private and governmental regulatory bodies. Their continued efforts are needed for assessments of the quality of care to continue and to improve.