Basic Concepts and Propositions

The IOM has defined quality of care as "the degree to which health care services for individuals and populations increase the likelihood of desired health outcomes and are consistent with current professional knowledge" (IOM, 1990i, p. 21). To the extent that guidelines are based on scientific knowledge, estimate expected health outcomes, and delineate current professional judgment, they clearly have a role to play in assessing and assuring the quality of care.

Efforts to ensure high-quality care must prevent or, alternatively, detect and overcome three main problems: (1) overuse of unnecessary care and of inappropriate care, (2) underuse of necessary care, and (3) poor performance (in both the technical and interpersonal senses). Most experts now agree that a comprehensive approach to quality of care must address all three problems, perhaps to different degrees depending on the setting or nature of the care and various local or institutional factors. Good practice guidelines have the potential to contribute in each area.

First, guidelines and review criteria that explicitly and clearly describe appropriate care for particular clinical problems provide a solid base for detecting patterns of overuse or underuse. Second, detailed guidelines may improve the technical provision of care. Certainly, some aspects of technique have to do with physical capacity, skill built on experience and repetition, attention to detail, and similar factors. Good performance, however, depends on a solid understanding of what constitutes appropriate care (e.g., choice of antibiotic) or correct performance of a technical task (e.g., sterile technique). Finally, when guidelines include good estimates of outcomes (risks, benefits, harms), they can contribute to better communication and shared decision making on the part of patients and practitioners. These interpersonal processes lie at the heart of the humanistic "art of care" vision of quality. Traditional programs of quality assurance (QA) (see Donaldson and Lohr, 1990) derive from a conceptual framework advanced more than a quarter-century ago by Donabedian (1966). This conceptualization (further explicated in Donabedian, 1980, 1982, 1985) emphasizes (1) the structure of care—that is, characteristics of practitioners, institutions, and the health care system including regulatory mechanisms; (2) the process of care, meaning at least patient care activities and, in some formulations, the housekeeping, administrative, and other activities that may affect quality of care; and (3) the outcomes of care, including patient health status and functioning, well-being, and satisfaction.

Clinical practice guidelines, medical review criteria, and standards of quality relate more directly to the processes of care than to outcomes because they describe what constitutes appropriate management of specific clinical problems. Performance measures describe the data needed to evaluate whether actual behavior conforms to guidelines, criteria, and standards. Utilization and quality review programs may employ all these instruments to identify and deter unnecessary and inappropriate services. As more and more guidelines explicitly estimate expected benefits and harms of care (that is, possible outcomes), they may more directly contribute to the specification of sophisticated criteria that relate good quality of care to expected (good) patient outcomes. The use and development of review criteria are discussed further in a later section.

At a systems level of quality assurance strategies, the potential role of guidelines in certification and re-certification programs and examinations should be noted. For example, as mentioned in Chapter 4, the American Board of Family Practice has for many years administered a re-certification process that involves review of office records against a predetermined list of performance criteria (Langsley, 1991). In Canada, the College of Family Physicians and McMaster University have been involved in an extensive effort to develop and apply chart audits, tests of basic clinical skills, and other methods. These criteria and assessments do not appear to be based explicitly on published practice guidelines, but such guidelines could certainly be factored into the certification and re-certification processes. As specialty boards continue their work on methods to judge qualifications accurately, attention to practice guidelines as one base for performance evaluation is an obvious step.

Continuous Quality Improvement

The case studies in Chapter 3 reflect the extent to which the use of guidelines to improve quality of care is a subtext in a managerial and policy debate over the relative contributions of traditional QA strategies and newer quality management approaches. These approaches are variously called total quality management (TQM) or continuous quality improvement (CQI). This discussion, which employs the latter term and abbreviation, relies on a composite model of CQI based on principles described in recent literature¹ and on observations drawn from the health care press, committee site visits, and other discussions. (The IOM committee that studied issues of quality assurance in Medicare [IOM, 1990i] compared traditional QA and CQI models in more depth.)

The growing interest in CQI is reflected in statements by the Joint Commission on Accreditation of Health care Organizations that it intends to shift from a QA to a CQI perspective and develop new principles and standards of hospital accreditation to reflect this shift (O'Leary, 1991). In describing the Joint Commission approach, O'Leary (1991) says, "We tend to use CQI . . . because to us the term means a way of life in an organization. Total quality management . . . might imply that there is a single management style that is necessary for all of this change to happen" (p. 74). A recent American Hospital Association survey states that more than 40 percent of reporting hospitals say they are engaged in continuous quality improvement (Utilization Review Newsletter, 1991b). The Maryland Hospital Association's ambitious quality indicator project is cast in a continuous quality framework (Maryland Hospital Association, 1990). In addition, the Robert Wood Johnson Foundation supports a program for "Improving the Quality of Hospital Care"; a quarterly newsletter for that program, entitled Quality Exchange, is produced at the Johns Hopkins University.²

CQI models are generally described in terms of a set of reinforcing principles for implementing change. They aim to make nonpunitive tactics for quality assurance more usable and, it is hoped, more effective than they have been under traditional approaches. Each of these principles has impli cations for the way practice guidelines may be incorporated into the fabric of health care organizations.³ The principles emphasize the following:

• close relationships between so-called customers and suppliers, that is, the partners in any given health care transaction;
• errors being more often the result of defects in systems (e.g., those for reporting test results or scheduling operating rooms) than the consequence of individual deficiencies ("bad apples");⁴ planning, control, assessment, and improvement activities grounded in statistical and scientific precepts and techniques;
• reliance on internal (self-) monitoring—as opposed to external (regulatory) inspection—with mistakes viewed as "treasures" that should be used for learning and for resolving problems rather than as an occasion for punishment;
• standardization of processes (decreasing their variability) to reduce the opportunity for error and to link specific care processes to health outcomes;
• feedback to practitioners of statistical information on how their practices may differ from those of their peers or depart from evidence-based standards for practice;
• visible commitment to quality by the top leadership of the organization and involvement by all parts of the organization in processes of quality improvement; and
• a striving for continuous improvement in contrast to simply achieving preset goals.

Within this framework, sound practice guidelines and medical review criteria have several possible uses. First, to the extent that guidelines become more sensitive to patient preferences and participation in decision making, they should improve patients' informed consent, their participation in decision making and, ultimately, their satisfaction with both the processes and outcomes of care. Guidelines could also help identify important patient outcomes to incorporate in patient satisfaction surveys and other instruments designed to improve or assess "customer-supplier" relationships. Second, guidelines and review criteria could play a role in identifying possible quality problems arising from underuse, overuse, or incompetent provision of care. They may be particularly useful in instances in which short-term health outcomes (those that are most readily employed) may not be good indicators of long-term results.

Third, to the extent that guidelines identify how compelling is the evidence for certain clinical practices, they will help in determining priorities for improving or standardizing specific patterns of clinical care and in sorting out competing claims for funding for biomedical and outcomes or effectiveness research. Fourth, participation by clinicians in the review, critique, and improvement of practice guidelines can help bring the science of medicine more forcefully into the equivalent ''cycles" emphasized by CQI.

Just as guidelines have the potential to contribute to continuous quality improvement, the application of CQI principles and processes can support the effective implementation of practice guidelines. Schoenbaum notes (1990, p. 102) that between the step in an algorithm (guideline) directing a colonoscopy and the actual procedure lie "dozens or hundreds of steps," such as locating, notifying, scheduling, and preparing patients. At all of these points things can and sometimes do go awry in ways that may undermine the successful application of guidelines for appropriate care. A number of CQI techniques are designed to uncover such process flaws and to structure activities to correct them. Moreover, because CQI generally tends to have a "quantitative" emphasis and to be detail oriented, it may encourage those who design practice guidelines and medical review criteria to be more explicit, specific, and comprehensive with respect to the clinical content of guidelines and to better anticipate and confront the practical problems that may face prospective users of guidelines. In addition, the CQI stress on cycles of planning, testing, evaluating, and revising of procedures—the "Plan/Test/Check/Act" cycle—can encourage the processes of guideline assessment and improvement called for in this report and elsewhere.

Although practice guidelines have a potential role in CQI, some aspects of CQI models or strategies may deflect attention from guidelines. To date, CQI efforts have tended to focus more on nonclinical than on clinical issues. Quality problems arising from poor clinical performance and decision making thus have not been highlighted (Causey, 1991). In discussing a series of quality management projects, Berwick and colleagues (1990, p. 24) reported that few of the project teams tackled clinical processes; rather, they focused on problems similar to those found in other industries, problems that were on the "comfortable fringes" of clinical processes. Typical of the problems to which these and other projects gave attention were the length of time needed to get X-rays posted to the medical record department, the late arrival of patient meals, and delays in patient discharge or admission. Williamson (1991) cites a study of TQM prepared for the federal Office of Management and Budget that reports that "nearly 40 of 44 examples [almost 90 percent] of improvement achieved focused specifically on administration process variables as opposed to outcomes, especially health outcomes" (p. 55).

How much this near-total emphasis on administrative rather than clinical processes will change as CQI becomes more institutionalized is unknown. Also unclear is how much the focus should change given the unknown magnitude of clinical versus nonclinical quality programs and the relative susceptibility of each to improvement. The focus on nonclinical issues is, in any case, understandable. CQI is still relatively new, users are looking for easier rather than harder targets of opportunity, and nonclinical problems are likely to be easier to solve from both technical and behavioral perspectives.

Nonetheless, even when clinical outcomes are at issue, practice guidelines and outcome measures do not appear to be widely perceived or explicitly applied as benchmarks for informing or assessing performance. For example, in a recent article examining CQI concepts and applications for physician care, the only reference to practice guidelines occurred in a discussion on the need for chart review and other monitoring activities to supplement CQI (Kritchevsky and Simmons, 1991).

Among the hospitals visited during this study, several were using or trying to develop clinical protocols or pathways that specified the sequence and timing of various interventions for different clinical problems; the object was to standardize practice and reduce errors. (Appendix A presents a pathway for coronary artery bypass surgery.) Most pathways in these hospitals and elsewhere appear to be built on implicit clinical judgments and local statistical data rather than on systematically developed practice guidelines (Coombs, 1991). This reflects the highly operational environment of pathway development and the relatively immediate opportunities for incremental action, monitoring, and adjustment. The focus of most pathways appears to be when and how to undertake a particular intervention (such as cardiac monitoring or respiratory therapy), not whether the intervention is appropriate.

Similarly, in confidential materials reviewed by the committee, several organizations structured their feedback to physicians almost entirely in statistical terms, in part because they feared that substantive guidelines might antagonize physicians, at least initially. An individual physician might be informed how often he or she performed a particular procedure compared with patterns of care for the same procedure by peers—without an explicit accompanying statement of the appropriate indications for the procedure.

Notwithstanding that the principle of variation reduction may be generally sound, specific reductions in practice variation and moves toward consistency are not inevitably shifts toward more appropriate care. Quantitative analyses of variation ideally should consider explicit practice guidelines as one reference point, given that soundly developed guidelines represent the best current understanding of which health interventions will produce desired results (Coombs, 1991). This understanding, in turn, can be evaluated and revised based on a systems study of the link between practice choices and short- and long-term changes in health outcomes.

To some degree, reliance on statistical feedback may reflect the lack of credible, relevant, and specific guidelines for problems targeted by a quality improvement effort. However, other explanations are also likely to apply. One anonymous reviewer of this report observed that many advocates of CQI see practice guidelines as "outdated." This view may reflect an identification of guidelines with the procedures for external review that have been widely criticized by CQI proponents, both on grounds of general principle and on grounds of poor administration. In stressing internal strategies for improvement and rejecting outside review, advocates of CQI may disregard the legitimate interest of other parties in external oversight and systems-level monitoring through retrospective, concurrent, and prospective review programs.

In this context, it is important to recall the observation in the IOM report (1990i) on quality of care in the Medicare program that the superiority of specific CQI techniques, although plausible, has not been demonstrated. Further, and consistent with the sense of the committee, Williamson (1991) reports that "many [CQI] advocates whom I have queried acknowledge that they have had difficulties adopting industrial [quality improvement] methods to clinical outcomes" (p. 55). As a practical matter, people are still trying to grasp what CQI means, how to use it, and how to assess its impact.⁵ As one hospital manager said, "We are only at mile two on a never ending journey" (Causey, 1991, p. 3). The stress on guidelines is even more recent than the promotion of CQI and postdates the initiation of CQI projects in many institutions. As credible and relevant practice guidelines become more available and more widely known, they should be perceived as more relevant to those attempting to implement quality improvement models in health care settings.

Medical Review Criteria

Although CQI programs may emphasize professionalism and internal quality improvement rather than regulation and external inspection, both approaches are needed for a public program such as Medicare and for private health plans that are accountable to employers and others for their performance. How well a continuous improvement mentality can dovetail with outside monitoring, periodic audits, and externally developed review criteria is an important—and unanswered—question. The characteristics of specific medical review criteria and the processes for developing and applying them can make coexistence either easier or more difficult.

Directions for Quality Assurance Strategies

The committee believes that well-developed, scientifically based practice guidelines have an important role to play in assessing and assuring the quality of health care services provided in this country. Clear, specific guidelines and associated review criteria should help deter or remedy problems of overuse of care, underuse of care, and poor technical and interpersonal provision of care. Guidelines accepted by those responsible for providing care, those responsible for financing it, and those responsible for monitoring care in the public interest are one means of bridging the chasm between internal and external quality assurance strategies.

With respect to models of quality assurance as discussed earlier, the committee urges that their focus on systems problems, on improvement of average performance, and on variation reduction be more systematically and explicitly joined with an effort to apply and improve sound guidelines for clinical practice. Specifically, the committee urges the following:

• Guidelines, medical review criteria, and other evaluative tools should be used both to improve average performance and—as is still important—identify substandard performance.
• Analyses of how individual practice patterns differ from average patterns should go beyond statistical analysis to consider relevant practice guidelines as benchmarks for performance.
• Both the statistical information from such analyses and the pertinent guidelines should be part of educational feedback on practice patterns.
• Evaluations of performance and outcome data should seek to determine the sources of poor outcomes and deviations from guidelines so that systems problems can be corrected, information efforts strengthened, and, if necessary, impaired practitioners dealt with through counseling, limiting of privileges, or other appropriate mechanisms.
• Evaluations of performance and outcomes data should also be used to determine whether practice guidelines ought to be updated or revised.
• Developers of guidelines and health care institutions should convene educational conferences to acquaint practitioners with specific guidelines and provide an opportunity for them to discuss and plan setting-specific applications.
• Institutional activities to develop guidelines or adapt national practice guidelines should aspire to the attributes for guidelines described in Chapter 1 of this report.

The committee suggests that the popularization of continuous quality improvement in the health arena may have underemphasized a principle that was clearly articulated by Deming and others in their original discussions of CQI in the industrial sector. This principle is that an organization must integrate the science of its field into its day-to-day workings. When applied to health care, the principle brings the role of science-based guidelines more to the fore. In other words, each of the activities listed earlier is a vehicle for bringing science-based guidelines into efforts at quality management and improvement.

In addition, the committee recognized the controversy that has developed over the use and content of medical review criteria, and consequently identified several desirable attributes of such criteria. These eight attributes are analogous to those described in Chapter 1 for clinical practice guidelines. They are listed in Table 5-1. Both sets of attributes build on the IOM report on quality assessment and assurance for the Medicare program, which identified 23 desirable attributes of what it termed appropriateness indicators, case-finding screens, and evaluation and management criteria.

TABLE 5-1

Desirable Attributes of Medical Review Criteria.

Generally, the committee would prefer carefully devised "national" review criteria to those developed locally, for reasons that are discussed more thoroughly in Chapter 7's examination of local adaptation of practice guidelines. In addition, like guidelines, review criteria should be accompanied by documentation of the procedures followed, the participants involved, and the evidence or guidelines used as a basis for designing them. By providing such information, review organizations can respond to some of the more serious criticisms of their credibility.

Coverage Policy and Administration

As described in the second case study in Chapter 3, health benefit plans are a clear "market" for guidelines and review criteria. These plans apply guidelines in various ways to limit their liability for particular expenses and to influence practitioner or patient behavior (IOM, 1989a). When existing guidelines do not meet their needs, health plans may undertake their own development initiatives.

In this context, health benefit plans include traditional indemnity insurance and Blue Cross/Blue Shield programs, self-insured employer plans, preferred provider organizations (PPOs), and health maintenance organizations (HMOs), as well as the various organizations that may provide these plans with such services as claims administration and utilization management. Although public plans such as Medicare and Medicaid may operate under special legal and other constraints, much of this commentary applies to these plans as well.

Credentialing, Selective Contracting, and Related Strategies

In theory, guidelines may inform decisions made by health care institutions, HMOs, and other organizations about the selection and retention of practitioners. Such decisions may involve the initial credentialing of physicians and ongoing delineation of clinical privileges in hospitals and other institutions. They may also involve hiring and contracting decisions in a variety of contexts. Most selection or credentialing decisions are based on educational qualifications, licensure, board eligibility or certification, previous positions, willingness to abide by organizational policies, and similar factors. However, quality objectives, liability concerns, and cost-management objectives are providing the impetus to develop and apply more direct measures of medical competence and performance (Gosfield, 1991b). The specifics of this process, including the role of practice guidelines, are still being defined (Langsley, 1991).

Health benefit plans and health care institutions that somehow select or "credential" practitioners have several opportunities to encourage the application of practice guidelines.¹³ For example, they can make employment, participation, or privileges contingent on a practitioner's prior agreement to practice in accord with the organization's clinical policies. Among the sites visited by the study committee was a group of primary care clinics that employed physicians with the understanding that they would practice in conformity with the clinic's primary care manual and that their practice would be evaluated annually. The group's primary care manual is a collection of both what this committee considers systematically developed guidelines and protocols and policies that reflect the system's patient population, resources, and objectives.

Guidelines can be worked into credentialing and other decisions in several ways. Depending on the access of a health benefit plan to relevant data sources, a plan can examine or profile the practice patterns of candidates for new or continued employment, contracts, or privileges to determine whether they already practice in accordance with selected guidelines. This review of practice may be based on claims data, medical records, and on-site observation (Stocker, 1989). Currently, however, nonclinical factors, such as geography and the willingness of providers to agree to health plan terms, may play the major role in selection and deselection decisions. The committee knows of no good evidence on the subject, but it suspects that profiling based explicitly on clinical practice guidelines is less common than simpler profiling based on utilization rates or levels.¹⁴

To the extent that credentialing, selective contracting, and similar strategies work well and are supported by ongoing health plan structures and processes, they presumably should bring into health plans and institutions those practitioners who are already committed to the desired practice patterns and then help to maintain that commitment. To maintain or redefine desired performance, other strategies are necessary. These may be both negative (e.g., financial penalties or even dismissal from the plan for unacceptable performance) and positive (e.g., education and feedback consistent with quality improvement models). Guidelines may contribute to the definition of what constitutes acceptable clinical performance for purposes of education, feedback, or evaluation.

Interestingly, one major initiative that is very strongly linked to the concept of selective contracting or purchasing, InterStudy's Outcomes Management System, makes little explicit reference to practice guidelines (InterStudy, 1991). This system is built on four main elements: (1) protocols and instruments for collecting data on patient satisfaction, functional status, demographic and other characteristics of patients, insurance status, and treatment setting; (2) condition-specific measures of changes in patient status over a course of treatment; (3) construction of a national data base; and (4) practitioners and analysts working in participating organizations to analyze information and promote change.¹⁵ With InterStudy acting as coordinator, two consortia, one involving six large group practices and another (in the planning stage) involving 24 employers, are engaged in a major effort to test this system.

One appealing feature of selective contracting is that it can reach beyond the hospital and large group practice to office-based practitioners in small group practices, at least when it is used by independent practice associations (IPAs), PPOs, and similar groups. Thus, it can encourage conformity with good guidelines in settings less readily reached by other strategies. However, one question about the use of guidelines in closed panel systems with incentives for physicians to control costs is what information should be provided to consumers about the incentives, guidelines, or other components of these plans. This question is explored in the next chapter.

Other Economic Incentives

Many health plans and public programs use explicit financial incentives to influence practitioner or patient behavior without explicitly attempting to encourage specific appropriate care and discourage specific inappropriate care (Brook, 1991). For example, in and of itself, typical patient cost-sharing in the form of deductibles and coinsurance applies equally to appropriate and inappropriate care. Likewise, capitated and per-case methods for paying health care practitioners and providers do not, in themselves, explicitly differentiate between appropriate and inappropriate care. When, however, such provider payment methods are used in conjunction with monitoring, educational, and other programs to detect poor quality care and to promote specific types of appropriate care, they may play an important strategic role in encouraging the application of sound guidelines.

Although explicit financial incentives have been rejected by many as philosophically objectionable or simply too controversial, some employers and insurers have designed positive or negative financial incentives to change specific consumer or patient behaviors that are inconsistent with recommended health practices and that are thought to lead to higher health care costs (New York Business Group on Health, 1990; Becker, 1991; Sipress, 1991; Terry, 1991). For example, some payers employ such incentives as insurance premium rebates (positive) or higher insurance premiums (negative) for individuals based on the results of blood, urine, and other tests related to recommended cholesterol and blood sugar levels, blood pressure, weight, and smoking. Some health plans cover or provide counseling for individuals who "fail" the tests; others do not. An unknown number of employers go further and refuse to hire individuals who are thought to be at higher risk of illness or injury because of their behavior or genetic inheritance.

Taking a different but also controversial approach, some theorists have proposed that consumers could be discouraged from demanding marginally beneficial care if they were offered choices among health plans that offered more or less generous levels of care as defined by different practice guidelines (Havighurst, 1990a, 1991a; Blumstein, 1991). To be covered for care for which the expected benefit is slight or questionable, consumers would buy a more expensive plan governed by more generous guidelines; if a less generous plan were chosen, consumers could, in principle, pay for such care on their own. Potentially, the guidelines could be incorporated into the health plan contract itself and referenced by specific name or they could be developed according to a process defined in the contract.

This contract-based strategy would offer direct economic incentives to the consumer in the first instance, with provider behavior affected as a consequence.¹⁶ Whether policy makers would agree that patients and consumers are able to make informed choices about these important matters and that they should be held to their choices is an open question that demands thorough debate. Research and experience suggest that people frequently find some basic provisions of existing plans difficult to comprehend (McCall et al., 1986; National Association of Private Psychiatric Hospitals, 1991). Even if people understand their choices, is it fair to expect healthy people to make good choices about health insurance that would apply in the future, should they or their family members become ill? Aaron (1991) notes that "health insurance and health care pose special problems [for economic analysis because] health insurance typically is purchased by healthy people, while most health care is consumed by sick people. Society normally has little interest when people gamble and lose. But when the gamble concerns events that change basic preferences and that affect the life and health of oneself and one's family, it is not clear why past consumer decisions deserve priority over new preferences" (p. 17). The more drastic the consequences of the choices people make about health care, the more compelling are these kinds of questions and arguments. Similar issues arise, for example, with respect to decisions involving living wills and advance directives about the use of life support in cases of terminal illness or injury. Judgments on such matters require a mix of empirical evidence (which may not be available), ethical judgments, and practical wisdom.

Directions for Cost Management

On both philosophical and strategic grounds, this committee believes that thoughtfully designed and applied programs to encourage cost-effective use of health care have an important role to play in supporting the wider application of guidelines for clinical practice. Such programs need guidelines and related materials that provide information on the cost-effectiveness of alternative ways of managing particular clinical problems. They also need to be supplemented by explicit programs that employ guidelines and medical review criteria to monitor the quality and appropriateness of care.

Those who develop review criteria should be guided by the attributes for such criteria that were discussed in the section on quality assurance and improvement. Review organizations that allow these attributes to govern their work on criteria will make their review activities as manageable and as unintrusive as possible for both patients and practitioners. They will also make their review criteria available to practitioners and others. Furthermore, these organizations will provide an explicit process for appealing negative decisions that is free from unreasonable complexity, delay, or other barriers. If a review organization identifies quality-of-care problems, it should have procedures for, at a minimum, discussing these problems with the practitioner or provider involved and, perhaps (with due regard for legal risks), raising the matter with the relevant PRO.

It is the committee's hope that economic incentives and quality review mechanisms will, in the future, reduce the need for so-called micromanagement of professional and institutional behavior. Utilization review still may have a role in monitoring practice and targeting problems, but many payers will readily admit that they would prefer to rely more on effective self regulation by practitioners and providers. Consistent with the principles of quality improvement, they can stress education and feedback to physicians aimed at improving practice rather than punishing missteps.

Risk Management

Risk management programs attempt to evaluate and decrease liability risks related to clinical care, housekeeping functions, management decisions, and other sources. Liability claims may arise from adverse events experienced by patients, visitors, staff, and others—even by prospective patients turned away from care. Across health care institutions, risk management programs and quality assurance programs are generally distinct organizational functions with varying degrees of interaction (Donaldson and Lohr, 1990). The wider availability and application of clinical practice guidelines should provide a stimulus for closer coordination.

For risk management programs, clinical practice guidelines offer several potential benefits, as shown in Figure 5-1. First, the wider application of good guidelines should improve clinical performance and thereby reduce the number of adverse events. These outcomes have been one objective of guidelines for anesthesiology and emergency room care. An important corollary aim has been to have malpractice insurers recognize adherence to such guidelines as an indicator of lower risk and a basis for lower premiums (Eichorn et al., 1986; Holzer, 1990).

FIGURE 5-1

Guidelines and Malpractice Incidence and Costs.

Second, practitioners and health care institutions should find well-developed, condition-specific or treatment-specific guidelines useful in communicating to patients the risks and benefits of treatment alternatives, thus helping patients to make informed choices among alternative courses of care. From a risk management perspective, among the objectives of such communication should be the reduction of litigation inspired by poor communication and of disappointment resulting from unrealistic patient expectations of perfect or guaranteed outcomes.

Third, guidelines that define appropriate care for specific clinical conditions can help in determining whether identified adverse events are the result of poor care rather than the unfortunate consequence of medical uncertainty. An adverse drug reaction following first-time administration of a drug to a patient with no ascertainable contraindications would fall in the latter category; a second use of the drug followed by another adverse reaction would, absent qualifying circumstances, likely be judged negligent (Brennan, 1991b). Sound guidelines should provide a strong basis for distinguishing negligent from nonnegligent care, which, in turn, should help deter unjustified malpractice claims, resolve justified claims earlier, and improve decisions for cases that go to trial.

Fourth, if practitioners have confidence in particular clinical practice guidelines and expect that documented conformance with these guidelines will do much to protect them against unwarranted claims (and findings) of malpractice, the anxieties that give rise to defensive medicine should diminish and the willingness to apply guidelines should increase. Given that defensive medicine occurs when practitioners provide or order services that are unnecessary except as potential malpractice protection, a reduction in defensive medicine should both improve quality of care and help contain health care costs.¹⁷

Medical Malpractice

The second, third, and fourth points above bring clinical practice guidelines directly into debates about medical liability, standards of care, and malpractice reform. Most proposals for malpractice reform concentrate on such issues as limits on punitive damages and use of alternative dispute resolution mechanisms; many, if not most, make no mention of practice guidelines (GAO, 1987; Macchiaroli, 1990; U.S. House of Representatives, Committee on Ways and Means, 1990; Miller, 1991). Nonetheless, the increasing interest in guidelines has prompted several analyses of their potential for improving decision making in cases of alleged malpractice. Although opinions about this potential are mixed, the precept that "good medicine is good law" is widely accepted (Miller, 1991, p. 1).

The legal theory of liability provides a deterrent to wrong behavior as well as a form of redress for such behavior when it occurs. To influence behavior, the liability strategy relies on economic penalties and fear. In this context, the role for guidelines is not to motivate but rather to assist decision making—prospectively in the case of clinicians and retrospectively in the case of those evaluating claims of malpractice.

The potential impact of guidelines on the incidence and resolution of cases of medical negligence depends on many factors that have already been discussed and questioned here. These factors include the soundness of guidelines (e.g., their validity, reliability, precision), the degree to which practitioners accept and effectively employ them, and the extent to which guidelines exist for particular types of care—in this case, care that is likely to incur negligence and give rise to legal claims of negligence. On this last point, one recent article estimated that existing guidelines could provide relevant evidence for 20 percent of the medical injuries identified in a major study of adverse medical events and related litigation (Garnick et al., 1991). Since the availability of practice guidelines is growing, this percentage should also grow somewhat—even if malpractice concerns are not the key factor motivating the development of most guidelines.

In addition, for guidelines to influence legal decision making, courts have to accept guidelines as important evidence of the standard of care. Some argue that the potential of guidelines to reduce defensive medicine, improve decisions in liability cases, and discourage unwarranted claims cannot be adequately realized unless courts accord clinical practice guidelines more weight than they currently do in determining the standard of care to be applied in assessing claims of malpractice (Hall, 1989; Brennan, 1991b; McCormick, 1991). This committee agrees generally with that proposition. The rest of this section presents definitions and describes options for granting guidelines more weight in malpractice decision making. The next section presents the committee's assessment.

According Guidelines Greater Weight in Determining the Standard of Care

Although the definition of standard of care in medical liability law has been changing—moving away from local toward national standards and permitting published standards to be cited as evidence—change based on case law tends to be neither quick nor predictable. Some statutory or administrative mechanism may well be needed to give guidelines based on sound, documented scientific evidence and analytic processes greater weight in litigation than they now have. In particular, such a mechanism may be needed to provide immunity from malpractice liability for performance in accord with practice guidelines. Immunity means that a conforming clinician, if sued, would not be held liable absent some other basis of liability.

Assuming the existence of good guidelines that can be used in everyday medical practice and as a basis for evaluating performance, what else may be necessary to accord guidelines greater weight in determining the legal standards of care? Table 5-2 lists some options and considerations for action. They relate to (1) how guidelines might be granted special recognition, (2) what authoritative status guidelines might have, (3) what parties could use the guidelines in the prosecution or defense of a malpractice suit, and (4) what weight might be granted guidelines.

TABLE 5-2

Options For Recognition of Guidelines in Medical Malpractice Law.

Proposals for legislation that would confer malpractice protection on clinicians who practice in conformance with guidelines have a precedent dating back to the 1972 law (Public Law 92-703) that created professional standards review organizations (PSROs); the relevant provision continues to apply to the successor peer review organizations (Gosfield, 1975). That provision declares that if a doctor has exercised due care, he or she cannot be held liable for actions that conform to norms developed by PSROs. It appears, however, that this "immunity" provision has never been used as a defense against a malpractice claim (Gosfield, 1989). Perhaps the due care reference raises the specter of continued argument about whether the standard of care was met, thus making the immunity promise appear dubious. More important may be the general low representation of Medicare patients among litigants and the lack of knowledge among attorneys of this provision of the Social Security Act and its amendments.

More explicit and more controversial than the PRO immunity provision is the approach taken by the state of Maine. The state legislature has initiated a 5-year demonstration project to produce "standards of practice designed to avoid malpractice claims and increase the defensibility of the malpractice claims that are pursued" (Edwards, 1991, p. 3).²² Unlike the federal legislation that created AHCPR, the Maine statute does not explicitly require that the guidelines that are developed be based on the best available scientific data.

The process was designed by the Maine Medical Association and is defined by statute. The state's Bureau of Insurance and Board of Registration in Medicine oversee the project, which started with three medical specialty advisory committees composed of physicians and public members serving 3-year terms. The advisory committees have, as directed by statute, focused on anesthesia, obstetrics and gynecology, and emergency medicine; radiologists recently asked to be included, and legislation to that effect is pending. Smith (1990), in explaining the legislature's action, states that the first two clinical areas were selected because of the "well-established standards that have already been promulgated nationally" (p. 2).

Once guidelines and protocols are developed, the Board of Registration will adopt them as rules under the state's Administrative Procedure Act. If 50 percent of the physicians in the state agree to practice according to the guidelines, then physicians can cite the guidelines in their defense in malpractice cases. Plaintiffs cannot use the guidelines unless they are first cited by the defense or unless the provisions they cite are identical to those in some other independently developed set of guidelines. At this point, it seems that most of the Maine guidelines will incorporate guidelines already issued by other organizations; one exception may be provisions regarding preoperative tests for patients facing relatively uncomplicated surgery (G. Smith, Maine Medical Association, personal communication, June 10, 1991).

Questions have been raised about the constitutionality of the "defendant use only" aspect of the Maine project. Indeed, one malpractice insurer, nervous over the questions raised about the law, has backed away from the effort. Nonetheless, the legislature does not appear inclined to change this aspect of the law, which was a significant element in the careful, lengthy process of negotiation among involved parties that led to the demonstration project (State Representative C. Rydell, Maine House of Delegates, personal communication, August 5, 1991).

In 1986, the legislature of Massachusetts took a different approach. It created a risk management unit in the state's medical licensing and discipline agency and required physicians to participate in a quality assurance program as a condition of licensure. Guidelines feature prominently in the quality assurance program, and the statute includes certain requirements for their content and for the development process. Physicians who participate in the program are entitled to reductions in malpractice insurance premiums from the state's liability insurers. Two specialties, anesthesiology and emergency medicine, have developed guidelines that initially qualified participating physicians for a 20 percent discount in liability premiums (McGinn, 1988).

Whether the experience of Massachusetts can be generalized to other types of guidelines is not clear. For example, the anesthesiology guidelines have several characteristics that are not common to most guidelines. Specifically, they were developed as a direct response to problems recognized as sources of liability; they are precise, particularly in their key recommendations for the use of pulse oximeters; and professional support for them was virtually unanimous. (It should be noted, however, that these guidelines do not cite or analyze any scientific evidence of effectiveness.)

Statutes that would allow judicial and jury discretion in determining the admissibility of guidelines or that provide for their consideration as merely some evidence of the applicable standard of care accord little more stature to guidelines than is granted under existing case law. Hall (1991), however, suggests that statutes specifically allow judges to consider, without the jury present, the stature of guidelines offered by defendants. Thus, without being bound by the technical rules of evidence, judges could determine the weight to give guidelines and could, in the case of "authoritative and indisputably applicable" guidelines, direct a verdict for the defense (p. 135). Authority would be assessed according to "the respectability of the issuing organization and its process of promulgation, [separate] from the inherent appropriateness of the standard itself" (p. 140).

At the federal level, several legislative proposals link guidelines to medical liability reform (McCormick, 1991). The most far-reaching would enmesh guidelines in the competition strategy for health care reform (described earlier), using contract law and federal tax policy as leverage (Havighurst, 1990a, 1991a; Blumstein, 1991). This approach would encourage consumers to choose among health plans that could offer different standards of care. Health plans could incorporate (or cite) particular practice guidelines in their contracts, or the contracts could provide for the application of guidelines developed according to specific procedural and substantive criteria.²³ How courts would reconcile this approach with the mostly unitary standard of care that has been developing through medical liability case law is far from clear (Blumstein, 1991; Schulman, 1991). This uncertainty has prompted a proposal that would essentially require, first, that health plans and their participants use an arbitration procedure for dealing with malpractice and, second, that the arbitration process apply the standard of care defined in the health plan contract (Havighurst and Metzloff, 1991).

Directions for Risk Management and Liability Standards

Once again, the committee stresses the desirable attributes of guidelines that were identified in Chapter 1. Regardless of specific statutory developments, guidelines that are based on scientific evidence and judgment and that are clear, specific, and developed by a reputable organization and process should carry greater weight in malpractice decision making than vague, nonspecific, and undocumented guidelines; they should certainly be accorded more weight than isolated expert testimony (Miller, 1991; Hirshfield, 1990b). To the extent that guidelines document the strength of the evidence, the importance of the risks and benefits, and how compelling is the case for a particular intervention, they should help courts distinguish required care from optional or unindicated care.

The committee vigorously debated whether to recommend that federal or state legislation grant immunity to practitioners acting in conformance with practice guidelines. Although some members argued quite forcefully that such a recommendation was warranted now, the committee on balance concluded that it would be premature.

Legislation that provides for immunity, the committee believes, ought to specify either operational criteria for the organizations developing guidelines or particular criteria for the guidelines themselves. Criticism of the variability and weaknesses of guidelines developed by PROs made the committee reluctant to accept organizational imprimatur alone as a sufficient basis for a grant of immunity. More broadly, existing publicly and privately developed guidelines and related organizational processes for developing guidelines vary widely in quality; some are clearly inferior. As discussed in Chapter 8 (and Appendix B), an entity to assess the soundness of guidelines might provide the foundation on which legislation could rest. Such a body could help users of guidelines, including attorneys and courts, to evaluate the merits of different guidelines, and hence the merits of lawsuits that employ guidelines as evidence.

Although the committee understands that allowing only the defense to cite authoritative guidelines in court might reassure physicians and speed acceptance of guidelines, such a restriction is unfair and probably constitutionally suspect (McCormick, 1991). Equally important from the committee's perspective, such restrictiveness is inconsistent with the basic concept of science-based guidelines for appropriate care. Both the evidence and any further credibility offered by a sound process of guideline development should be available to defendants and plaintiffs alike.

The lack of a legislative imprimatur does not mean that guidelines cannot and will not be increasingly used to define the standard of care (Hall, 1991). For example, judges are in a position to consider the reputability of organizations that develop guidelines. As the development and assessment of guidelines become more firmly grounded and more recognized in practice, judges and juries should eventually give them greater weight than isolated expert witnesses, less firmly grounded testimony, or older medical treatises.

The committee urges AHCPR to continue to support scientifically meritorious research on medical liability that could simultaneously examine topics related to guideline development. One question the agency might consider is whether clinical practice guidelines for services of unclear or very marginal benefit could be phrased and explained in ways that would reduce the likelihood of defensive medicine. For example, are judges or juries (or, for that matter, patients, lawyers, or practitioners) likely to make different decisions depending on whether guidelines (a) explicitly state that, given the scientific evidence, a practitioner can prudently forego a specific service under specific circumstances, (b) explicitly state that a practitioner may want to provide the service although evidence does not support a specific recommendation, or (c) are silent on the issue? Chapter 6 suggests how guideline developers might construct recommendations that would be

The committee noted one serious potential problem that medical liability law raises for the CQI model of quality assurance. Current liability procedures and standards could jeopardize implementation of the CQI vision of mistakes as "treasures" to be uncovered, analyzed, and used for learning. In an ideal world, if such mistakes resulted in harm to patients or others, institutions and practitioners would admit liability and compensate the injured party. This practice operates now to some degree, but the financial, reputational, and other potential harms of this aspect of the CQI strategy must be recognized as a powerful disincentive to such behavior.