4
Implementing Guidelines: Conditions and Strategies

It was not enough to produce satisfactory soap, it was also necessary to induce people to wash.

Joseph Schumpeter, 1939

Clinical practice guidelines may be meticulously developed, sound in content, clearly presented, and widely known, but they are without value if they are not successfully applied. Indeed, the resources consumed in producing and disseminating such guidelines are wasted if the guidelines are not employed to improve health or achieve other desired outcomes.

At the point of clinical decision making, the key actors are patients and practitioners. Over time, guidelines can improve that decision making by strengthening its science base, increasing its consistency across similar patients and problems, and explicitly identifying how compelling is the case for particular interventions. These steps require the projection and description of benefits and risks of alternative courses of care in terms relevant to patients.

However, even when specific, well-founded guidelines exist, patients and practitioners require a broad range of supportive conditions and organizations to secure their effective use. The creation and maintenance of these conditions will require resources and strong leadership by senior clinicians and managers.

This chapter begins by briefly examining the environment and the philosophical or strategic considerations that can shape how these conditions will be structured and how well they will function. The following sections consider how educational activities and computer-based information and decision support systems can encourage the application of guidelines. Chapter 5 discusses how quality, cost, and risk management systems may support and be supported by guidelines for clinical practice.



The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement



Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.

OCR for page 83
Guidelines for Clinical Practice: From Development to Use 4 Implementing Guidelines: Conditions and Strategies It was not enough to produce satisfactory soap, it was also necessary to induce people to wash. Joseph Schumpeter, 1939 Clinical practice guidelines may be meticulously developed, sound in content, clearly presented, and widely known, but they are without value if they are not successfully applied. Indeed, the resources consumed in producing and disseminating such guidelines are wasted if the guidelines are not employed to improve health or achieve other desired outcomes. At the point of clinical decision making, the key actors are patients and practitioners. Over time, guidelines can improve that decision making by strengthening its science base, increasing its consistency across similar patients and problems, and explicitly identifying how compelling is the case for particular interventions. These steps require the projection and description of benefits and risks of alternative courses of care in terms relevant to patients. However, even when specific, well-founded guidelines exist, patients and practitioners require a broad range of supportive conditions and organizations to secure their effective use. The creation and maintenance of these conditions will require resources and strong leadership by senior clinicians and managers. This chapter begins by briefly examining the environment and the philosophical or strategic considerations that can shape how these conditions will be structured and how well they will function. The following sections consider how educational activities and computer-based information and decision support systems can encourage the application of guidelines. Chapter 5 discusses how quality, cost, and risk management systems may support and be supported by guidelines for clinical practice.

OCR for page 83
Guidelines for Clinical Practice: From Development to Use CONTEXT, PHILOSOPHIES, AND STRATEGIES The context in which guidelines are to be implemented is important, involving as it does a cultural shift in American society. The nation is moving away from a tradition of substantial deference to professional judgment and discretion toward more structured support and accountability for such judgment. This shift takes visible and sometimes controversial form when guidelines for clinical practice move from the development to the application stage, especially when application is backed by formal organizational structures and procedures and by forceful incentives. The ways in which practice guidelines can and do operate as instruments for professional support and accountability are affected by the dynamics of a health care system that is changing and evolving, very often with no particular regard for practice guidelines. These complex, ongoing changes involve such fundamental matters as how medical care is organized and monitored how health benefits are provided to individuals and groups how practitioners and providers are paid how patient preferences are treated how information is recorded, manipulated, and retrieved. These changes may both support and undermine practice guidelines. Although policy makers may try to anticipate and avoid mismatched incentives, those managing the health care system inevitably will be left to deal with inconsistencies or conflicts, such as payment systems that reward overuse of care and guidelines that are intended to discourage such excess. In addition, guidelines are affected by the conduct of clinical research—its scope, priorities, and methods. Clinical and health services researchers can play an important role in making guidelines more applicable to operating environments. In particular, if researchers pay more attention than they have in the past to testing the effectiveness of procedures and patient management strategies in real settings as well as in highly controlled clinical trials, developers of guidelines are likely to have a knowledge base with greater practical relevance to practitioners and others. In turn, the greater the number of practitioners and institutions that adopt the outcomes management tools developed by health services researchers, the greater the body of information that will be available to evaluate and revise guidelines to make them still more useful in achieving desired outcomes. Overall, the influence on behavior of the varied and complex operational environments in which guidelines are to be applied cannot be stressed too much. Practitioner knowledge of guidelines and acceptance of their validity are key conditions for their successful application, but acceptance is not equivalent to change. Thus, as a practical matter, it may be better strategi-

OCR for page 83
Guidelines for Clinical Practice: From Development to Use cally or tactically to focus less on knowledge and acceptance and more on what changes behavior in desired directions (Schroeder, 1987; Lomas et al., 1989, 1991). The rationale for this position is that guidelines may be resisted or, if accepted, fail to motivate change, given strong countervailing forces—in particular, habitual practice patterns, malpractice fears, economic disincentives, information overload, and fear of diminished professional autonomy.1 Eisenberg (1985, 1986) has discussed six sets of activities required for successful alteration of physician practice patterns—education, feedback, participation, administrative changes, incentives, and penalties—and advised that a combined strategy is most likely to be effective. Proposals to change behavior generally reflect a mix of philosophical, strategic, and tactical considerations. For example, discussions of the relative importance of regulatory oversight versus market incentives typically reveal philosophical positions as well as practical views about how to achieve particular goals. Likewise, controversy about proposals to motivate individual conformance with dietary and other health promotion guidelines by charging higher insurance premiums or creating other penalties for noncompliers typically reflects disagreements about both what is fair and what is likely to work. In fact, money and fear figure in many behavioral change strategies (especially for practitioners), although proponents of change may downplay this fact in public statements. Furthermore, many attitudinal and "socialization" barriers stand in the way of behavioral change for traditionally educated physicians and, by implication, other health care professionals as well. These obstacles include the tension between professional autonomy and accountability for the quality of care rendered, processes of recruitment, training and socialization of members of the medical profession, and their preference for informal rather than formal quality assurance interventions (Donabedian, 1991). External barriers include the alienating effect (from the physician's perspective) of formal quality assurance efforts that emphasize identification of individual malfeasance and the near-total unfamiliarity of physicians and other clinical professionals with the concepts, methods, and tools of quality assurance or quality improvement. Among the approaches for overcoming these barriers and changing professional behavior are educational interventions, supportive organizational adaptations, directives, and incentives and disincentives of various sorts. Cutting across these factors are variables such as level of institutional resources and the commitment and competence of senior managerial and clinical leaders. The next section of this chapter discusses education and then turns to 1   One anonymous reviewer of this report argued that physician resistance to guidelines was part of a more general resistance to making clinical practice and judgment more regular, and that clinical judgment remains the "inner rampart" of physician autonomy.

OCR for page 83
Guidelines for Clinical Practice: From Development to Use information and decision support systems. Both are foundations on which to build the quality, risk, and cost-management strategies discussed in the next chapter. The discussion of educational strategies is intentionally brief. The committee judged that its efforts were better spent in focusing on strategies that had been less widely discussed. This lack of emphasis should not be taken to imply lack of importance. EDUCATION Education constitutes both a use of guidelines in itself and an essential component of quality assurance, risk management, and most other strategies for the effective application of guidelines. In medical school, residency, and continuing medical education, weaving guidelines into the fabric of educational processes is an important step in weaving guidelines into the fabric of medical practice. At this time, however, incorporation of guidelines into medical education is little documented and still subject to considerable debate (Darby, 1991b). In continuing medical education programs, specialty societies may organize sessions related to practice guidelines they have promulgated. In medical education, the recommendations contained in a set of guidelines may be less important than the literature reviews, descriptions of analytic processes, rationales, and other materials that should accompany them. Guidelines that provide thorough analyses of evidence, projections of benefits and harms for alternative courses of care, and clear rationales for statements about appropriate care offer a powerful teaching tool, more powerful in some cases than textbooks that lack such documentation and such demonstration of the processes of scientific reasoning. Greenfield, for example, argues that such guidelines can be "hyper-educational" in exposing students to physiopathology, pharmacology, literature review, and the translation of information into practice (Darby, 1991b). Exposure to specific guidelines combined with explicit training in how to assess them (and, for that matter, how to assess medical texts) provides opportunities to hone critical faculties in ways that can benefit clinicians throughout their professional careers. To this end, the Johns Hopkins University Program for Medical Technology and Practice Assessment is developing a curriculum to teach physicians such assessment skills (Robert Hayward, Johns Hopkins Medical Institutions, personal communication, 1991; Ackerman and Nash, 1991). As the revision and updating of guidelines become more systematic and as the opportunities for more-or-less instant electronic communication are more fully realized, some guidelines may become, in essence, the textbooks of tomorrow. Current initiatives to improve the assessment of medical competence and performance should be another stimulus to integrate guidelines into

OCR for page 83
Guidelines for Clinical Practice: From Development to Use practice throughout a professional's career (Nuckolls, 1990; Langsley, 1991). Board certification, which follows residency training and examinations administered by the boards, is a statement about physician qualifications at the time of certification, not about continuing competence over the long run of practice. Some medical specialty boards (notably family practice but increasingly others such as internal medicine) have begun to issue time-limited board certification; after 10 years, or some other designated period, the physician needs to reapply for certification. The re-certification process administered by the American Board of Family Practice includes a review of office records using performance criteria that apparently are not based on formal practice guidelines but that could be (Langsley, 1991). Some groups, such as the American Board of Internal Medicine, have been considering a role for clinical practice guidelines in establishing criteria either for re-certification or for eligibility to apply for re-certification, but close links between guidelines and board certification or re-certification almost certainly lie well into the future. For patients, too, corresponding avenues exist for lifetime learning about healthful behavior and problem-oriented decision making. Guidelines-related information can be incorporated in school, employment-related, insurer-based, and other health education activities; a long, albeit not uniformly successful, tradition of such education already exists as a foundation for these efforts. For example, health educators recognize that many patients or consumers—perhaps one in five—lack important reading skills. When the American College of Obstetricians and Gynecologists developed a magazine for parents or prospective parents who might not be able to use its existing publication, it quickly received requests for 700,000 copies; it had expected to distribute 500,000 copies over three years (Rovner, 1991). The American Cancer Society (ACS) has employed a different simplifying strategy in some of its materials. To attract attention and help embed key information in individual memory, the ACS (1990) uses a simple mnemonic device that highlights the first letters of each of the seven warning signs for cancer to spell C-A-U-T-I-O-N. Change in bowel or bladder habits A sore that does not heal Unusual bleeding or discharge Thickening or lump in breast or elsewhere Indigestion or difficulty in swallowing Obvious change in wart or mole Nagging cough or hoarseness Educational strategies for both professionals and laypersons can help build a foundation for specific quality assurance, risk management, and similar programs. Unfortunately, the appeal of educational strategies appears to be offset by uneven and sometimes discouraging information about

OCR for page 83
Guidelines for Clinical Practice: From Development to Use their impact and cost-effectiveness (see, for example, Eisenberg, 1986). The continuing challenge is to make this most commonly used approach for changing behavior more consistently productive. Many individuals and organizations are trying to meet this challenge, building on extensive behavioral research and practical experience (Eisenberg, 1986; Chassin, 1988; Kanouse and Jacoby, 1988; Green, 1991; Siu and Mittman, 1991). The Agency for Health Care Policy and Research (AHCPR), in particular, is committed to an extensive dissemination and education effort to support the guidelines it is developing (AHCPR, 1991). One important feature of educational strategies such as those cited above is their diversity. Education can be informal or formal impersonal or personal one-way or interactive isolated or connected to ongoing relationships knowledge oriented or change oriented sponsored by individuals or organizations of varying credibility. The most prominent educational strategies for practitioners focus on relatively formal, organized activities. These activities include medical school, graduate medical education, and continuing education courses that tend to be impersonal and involve only one-way communication. Computer or other self-teaching modules, on the other hand, are impersonal but can be interactive. Research on the impact of different educational strategies indicates that personal, interactive strategies tend to be more influential in changing practitioner behavior than are more formal or indirect approaches (Avorn and Soumerai, 1983; Eisenberg, 1986; Chassin, 1988; Soumerai and Avorn, 1990; Siu and Mittman, 1991). Programs undertaken by respected authorities in the context of ongoing organizational relationships are also effective, and sometimes the involvement of respected leaders may be the key to success or failure of efforts to modify the clinical practice. Small group education, individualized "academic detailing," and operations-level feedback of information on practice patterns are personal, interactive strategies with both formal and informal aspects.2 All of these activities can vary in the degree to which they go beyond knowledge building to stress behavioral change. Adequate evaluation of strategies for change requires that benefits be 2   For example, Avorn and Soumerai (1983) and Soumerai and Avorn (1990) describe academic detailing as including interviews to establish baseline knowledge and motivation associated with a practice; programs focused on specific categories of physicians and their opinion leaders; clearly stated educational and behavioral objectives; sponsorship by a respected organization; use of authoritative and unbiased information and concise graphic materials, and repetition of essential messages; active participation by physicians; and positive feedback on improved practice. The approach is built on marketing strategies used by pharmaceutical companies.

OCR for page 83
Guidelines for Clinical Practice: From Development to Use weighed against costs. Yet many studies of educational strategies do not report useful data on the cost-effectiveness of the strategies. Eisenberg (1986) notes that despite the appeal of personalized face-to-face feedback, it may not generate savings that exceed its cost. This is a serious problem if the primary object is cost containment rather than quality assurance or some other purpose. Even when cost containment is not the objective of an education strategy, managers need information on benefits and costs of the alternative strategies available to them. Educational strategies for patients or consumers tend to emphasize impersonal and relatively inexpensive mass information campaigns or to rely heavily on the physician-patient relationship, although this reliance is rarely reinforced by specific reimbursement for patient education (Green, 1991). At its best, the latter is personal, interactive, ongoing, and decision oriented (if not change oriented). The interactive videos now being developed and tested for prostatism and other conditions promise an attractive supplement to direct physician education of patients (see Chapter 6). Evaluations of the effectiveness and costs of this tool will be received with much interest. Educational tools for both physicians and consumers are relying increasingly on computers. A recent publication on pharmaceuticals (National Council on Patient Information and Education, 1991) lists an array of products ranging from a self-medication screening program developed at the University of Florida to commercial software that pharmacists can use to generate easy-to-understand educational materials for patients. Williamson (1991) has emphasized the importance of educating physicians on the needs of their patients for better information and education about the rationale for a course of care and the expected or possible physical and psychological consequences of compliance (including side effects) or noncompliance. He describes one specific hypertension treatment program, built on some of the quality improvement principles described in the next chapter, that includes education for physicians (in particular, specific information about patients' beliefs and behavior), an outcome-oriented plan for improving communication, feedback to practitioners on outcomes, and reevaluation of the program and its statement of the maximum acceptable level of patient noncompliance. Informal educational processes (such as telephone consultations with respected colleagues, bedside conversations, or lunchtime discussions) should not be ignored. The power of respected leaders to facilitate the diffusion, acceptance, and application of new information and technologies is undoubtedly felt in both deliberately organized and less formal ways (Eisenberg, 1986; Lomas et al., 1991). This applies as well to patients and consumers. Repeating the point that introduced this section, education is an essential component of most other strategies for effective application of guide-

OCR for page 83
Guidelines for Clinical Practice: From Development to Use lines. This report acknowledges the central role of educational strategies but, consistent with its charge, has attempted only to frame issues and options rather than to explore them in depth. As developers of guidelines improve the documentation that accompanies guidelines, as well as the clinical specificity and the explication of the reasoning that went into their formulation, the task of educators ought to become easier. Likewise, as this happens, those who develop and manage clinical information and decision support systems will find it easier to incorporate guidelines into these systems. INFORMATION AND DECISION SUPPORT SYSTEMS In manufacturing and financial services, computer systems oversee and control millions of individual actions. Such systems are ubiquitous and cover all activities associated with the function of an institution—as they do, for example, banking. In the medical care arena, however, even those settings with the most advanced computer systems have not automated the majority of their core clinical and other activities (IOM, 1991b). To the extent that automated systems support patient care, such support generally consists of clinical data rather than guidance about appropriate care; generally, it is hospital based and does not extend to the physician's office and similar settings. Information and decision support systems are crucial elements in long-term strategies for promoting the application of guidelines, the evaluation of their impact, and the feedback of such evaluation to revise and improve guidelines. The very translation of guidelines into algorithms and computer-based formats can spotlight deficiencies in guidelines (such as lack of specificity) and lead to revisions that will make guidelines more usable (Margolis et al., 1991). Although the following discussion emphasizes computer-based information systems, guidelines should also be available, understandable, and usable in conventional hard-copy forms. Current Systems The committee visited several institutions that already have or are implementing effective clinical information and decision support systems. An example is the Regenstrief Medical Record System (RMRS), used by University of Indiana house staff in the Wishard Memorial Hospital and its outpatient clinics (McDonald, 1976; McDonald et al., 1984, 1988; Tierney et al., 1990). The RMRS includes modules designed to record, retrieve, sort, and display medical encounter, treatment, and diagnostic study data as reports and flowsheets. The system also provides real-time clinical remind-

OCR for page 83
Guidelines for Clinical Practice: From Development to Use ers and alerts (based on patient-specific data) using protocols for pediatric, medicine, and obstetrics and gynecology clinics.3 For instance, reminders can alert physicians to preventive care needs or untreated hypothyroidism in patients seen for other ailments or reasons; they can also describe alternative diagnoses and therapies, calculate medication dosages, and estimate Framingham Risk of Cardiovascular Disease probabilities based on patient laboratory values. Regenstrief has developed more than 1,400 rules and has documented their rationale and scientific base (McDonald et al., 1988). In addition, users can create their own protocols. Some evidence about the impact of such systems is available. For example, at Regenstrief, a large, two-year randomized clinical trial found increases of up to 400 percent in the delivery of preventive care associated with use of the reminder system (McDonald et al., 1984). Other researchers, including those at Latter Day Saints Hospital in Salt Lake City and Massachusetts General Hospital in Boston, report similar results with computer-based reminder and decision support systems (Barnett et al., 1978; Hattwick et al., 1981; Pestotnik et al., 1990; Elliott, 1991; Williamson, 1991). Existing computer-based information and decision support systems differ in significant ways. These differences involve the degree to which the systems come into play automatically or at the discretion of the practitioner or patient-for example, an on-line reminder or surveillance system versus a user-initiated inquiry system; are more or less intrusive, a case in point being computer-based systems for ordering laboratory tests that request only the reason for a test versus systems that also require approval of the reason; emphasize information or control of behavior-for instance, on-line reminders of appropriate practice versus on-line limits on ordering certain services; and link practice guidelines to patient-specific information, a comparison being a general reference system versus an interactive protocol that uses specific information about a particular patient. As discussed later in this section, change is occurring on two fronts—technical and psychological—which should make computer-based information and decision support systems far more useful and attractive to clinicians. Nonetheless, the following constraints still apply to a considerable 3   Computer-supported reminder systems are not limited to practitioner use, although patient applications are still relatively limited and untested. The simplest systems provide medication storage containers with monitors that beep or otherwise alert patients to medication schedules and record use. A telephone-based reminder system for pharmacists is also available (National Council on Patient Information and Education, 1991).

OCR for page 83
Guidelines for Clinical Practice: From Development to Use extent to most of the health care system; overcoming them is one aim of the recommendations in the IOM's 1991 report on the computer-based patient record (IOM, 1991b). First, practitioners, institutional administrators, and others are wary of the expense of the requisite computer hardware and software. The track record (or at least the perceived record) of computer-related technology is one of failing to meet expectations or of becoming out of date rather quickly, both of which intensify investment concerns (Gardner and Perry, 1989; GAO, 1991a; Gardner, 1991). Second, computer hardware and software remain threatening or unappealing to many practitioners and patients. Even computer-literate practitioners object to systems that make it onerous to enter data or to retrieve accurate and useful clinical information and guidance on a real-time, interactive basis (Lundsgaarde et al., 1981; Brightbill, 1990; Fliegel, 1990; Gardner, 1990b). Third, current systems do not generally link all of the many sources of patient data (e.g., the physician office, commercial laboratory, hospital). Even when the source information is computer based, differences in data storage structures, record identifiers, and coding systems may make information exchange difficult across or within patient care settings (see, generally, Brodnik and Johns, 1991, which includes many of the articles cited here). In addition to deficient linkages to support clinical decision making for individual patients, linkages to support outcomes research and guidelines development and revisions are still limited. Fourth, independent of hardware or software limitations, guidelines themselves are often too incomplete for translation into computer-based decision aids (Margolis et al., 1991). A related problem is that computer-supported use of guidelines may require integration of information on specific patients that is not accessible automatically—for example, handwritten notes. Promising Developments Information and decision support systems are advancing on many fronts and in many ways (Gardner, 1990a; Grossman, 1991; IOM, 1991a; McDonald et al., 1991). These advances will make such systems more useful for many purposes, including support for the application of clinical practice guidelines. Several areas of progress can be cited. First, standard definitions and ground rules for transferring and using information from different computer systems are still limited but are emerging in such forms as Health Level 7 (HL 7) and draft standard 1238 of the American Society for Testing and Materials (ASTM, 1991; Hammond, 1991; McDonald et al., 1991). For practice guidelines, the ARDEN syntax provides a formal way to define guidelines so that they can be tested and

OCR for page 83
Guidelines for Clinical Practice: From Development to Use executed automatically and shared across different computer systems (Hripcsak et al., 1990). Second, although most information and decision support systems are now based in hospitals, hospital expansion of ambulatory activities and partnerships with physicians are making computer-based information systems more available to physicians in their offices. Starting at the other end, group practices and health maintenance organizations are beginning to extend their systems to cover hospital care. The Harvard Community Health Plan with its long history of computerization and its link to Brigham and Women's Hospital is an example. These developments make coordination of care easier and collection of data on episodes of care more feasible. Third, data collection and analysis strategies for laboratory, radiology, and pharmacy departments in patient care settings have become increasingly sophisticated. As information and decision support strategies are used in tandem with other kinds of implementation approaches, they can simultaneously present information and shape and control its uses. A case in point is a test- or drug-ordering protocol programmed into a computer-based decision support system, which can both display data and options and limit the orders that will be accepted for certain combinations of clinical problems. Today, most systems appear to be far less directive than this, but they are likely to change as systems become more sophisticated and as the emphasis shifts from merely providing information to producing desired changes in behavior and outcomes. The committee expects that clinical information and guidelines will become more integrated in such forms as expert rules, normal limits, contraindications, drug interactions, and other supports for decision making (Eckman et al., 1991). Fourth, information input and retrieval technologies are becoming less intimidating. Practical voice-recognition systems may be essential to widespread clinician involvement in timely entry of key patient information. Such systems are advancing, although they generally remain unable to handle normal, continuous speech and the large vocabularies required by medicine. Developments in the arena of information and decision support systems are important because they can support the application and, for that matter, the development of guidelines in at least three important ways. They can provide centralized storage, maintenance, and retrieval of guidelines; decision aids for practitioners (and, less commonly, patients) that are based on authoritative guidelines; and means for collecting clinical information for effectiveness, outcomes, and biomedical research that can, in turn, feed into the development or revision of guidelines.

OCR for page 83
Guidelines for Clinical Practice: From Development to Use Access to Information Perhaps the best developed and most easily improved of these three elements is the central clearinghouse function that permits access to good practice guidelines through remote computer links. The National Library of Medicine (NLM), various commercial vendors, and others are increasing the availability of clinical information through on-line literature search systems, floppy disks, and CD-ROM disks4 To date, these systems appear to include guidelines only incidentally, but this state of affairs is changing as guidelines become more visible (Brightbill, 1990; Frisch, 1991). The Omnibus Budget Reconciliation Act of 1989 requires AHCPR to promote dissemination of guidelines through organizations that represent health care providers or health care consumers, and through peer review organizations, accrediting bodies, and other appropriate entities. Among the first steps that the agency took to fulfill this mandate was to begin work with the NLM for inclusion of guidelines in various NLM bibliographic and information systems. The NLM is arranging for easy access to AHCPR-sponsored (and eventually other) guidelines by staffing its recently established Office of Health Services Research Information; developing bibliographic headings (Medical Subject Headings, or MeSH) related directly to guidelines; creating in its indexing system the label ''practice guideline" to identify guidelines as a type of publication; highlighting citations for AHCPR guidelines in GRATEFUL MED (an on-line software package for searching medical literature); providing on-line access (through its LOANSOME DOC system) to full texts of the "short-form" versions of AHCPR guidelines; and allowing on-line requests for mailing of the complete text of AHCPR guidelines and providing instant facsimile transmission of summaries or short versions of guidelines. Decision Support Not a new feature but also not commonplace is the programming of guidelines into management and decision support systems (Adams, 1986). Decision analysis software also has a role to play in the application of practice guidelines. These efforts can take several forms including the following: 4   CD-ROM (which stands for computer disk-read only memory) disks are computer storage hardware that runs on personal computers and has a vast storage capacity. Currently available CD-ROMs can store 500 million bytes of information, which could translate into as much as 250,000 pages of data or up to 1,500 floppy disks. CD-ROM systems also allow the user to consult several "books"—major diagnostic texts, compendia on the use of medical therapeutics or pharmaceuticals, and so forth—at once.

OCR for page 83
Guidelines for Clinical Practice: From Development to Use Reminders and alerts. For example, a system that integrated information from a patient's medical record with guidelines about appropriate care might be able to state, "Patient had Class III Pap smear four months ago. Repeat smear was due last month but was not done" (Adams, 1986). A system that had on-line entry of clinical activities or patient status could report, for instance, "Patient's throat culture is positive for beta streptococcus. No appropriate antibiotic has been initiated." A more forceful alert might state, "Tracheal tube cuff pressure greater than 27 cm H2O. Indicates excessive pressure and potential for destruction of tracheal cartilage, arterial bleeding, and sudden death." (This last example is adapted from Elliott's [1991] detailed description of the respiratory care system at the Latter Day Saints Hospital in Salt Lake City, Utah.) Embedded controls. For example: "Orders for parenteral nutrition are governed by the following protocol (list). Overrides require (procedures listed)." Decision assistance. An example might be the following: "Patient has mildly elevated creatinine and is already on quinidine, so a lowered dose of digoxin should be considered" (Adams, 1986). Or, "Patient has positive hemocult. Steps in workup are (screen displays flowchart)." Risk prediction. For instance: "According to Goldman's computer protocol for patients with chest pain, this patient has state 'K' and therefore is at high risk for myocardial infarction" (drawn from Goldman et al., 1988). Outcomes and Guidelines Revision Several IOM committees have emphasized the importance of revising guidelines in the light of new evidence about health care technologies (IOM, 1989a, 1990c,i). To that end, better clinical information about the efficacy and effectiveness of health care services must be acquired, and a number of organizations are engaged in practical initiatives with that aim. Improved computer-based record systems offer an opportunity to collect, aggregate, analyze, and transmit such data in a more timely and more organized way than ever before (IOM, 1991b). Such data might, in the future, be used to trigger review of an existing guideline or to provide some defensible reason for local adaptation of a guideline. On a broader scale, these data, as they are brought more quickly into the scientific literature, will become the foundation for better guidelines. Integrating Financial and Clinical Management Systems What were originally financial management systems are increasingly becoming clinical management systems as well, driven by changes in reimbursement systems (in particular, diagnosis-related groups), managed care contracting, and quality-of-care concerns. Two examples of these approaches

OCR for page 83
Guidelines for Clinical Practice: From Development to Use are the systems at Intermountain Health Systems (Salt Lake City) and the New England Medical Center hospitals (Boston). Both are moving from retrospective, externally oriented systems toward internal concurrent or prospective systems. Such shifts require that data be integrated from operational systems (e.g., laboratory, radiology, pharmacy, nursing), medical records, and financial systems to determine the costs of patient care "products." Depending on the quality and scope of the basic information, such integration may also allow practitioners and managers to develop budgets, to project how changes in ways of providing care will affect costs and clinical outcomes, and to evaluate departures from projected costs. For example, "if the laboratory is over budget, variances due to excessive test ordering (the responsibility of physicians) or use of more labor than predicted (the responsibility of the lab manager) can be identified" (Grossman, 1991, p. 242). Practice guidelines offer a benchmark for judging the clinical elements so as to identify problems. The convergence of financial and clinical management systems motivated by the economic and other pressures described earlier will be a powerful force for moving guidelines into multiple environments, for facilitating comparisons of provider performance and guideline impact, and for feeding back useful information to both developers and users of practice guidelines. The recommendations in Chapter 7 are designed to facilitate and test this proposition. Directions for Information Systems At this time, no adequate information infrastructure supports the kind of effective, unobtrusive, easy application of guidelines envisioned by continuous quality improvement models, future-oriented utilization management and cost-containment systems, and patient-centered care proposals. Clearly, however, the information technologies of the future will make the application of guidelines much easier, particularly if other conditions support their use.5 The work of the NLM and others to establish some capacity for responding to user-initiated inquiries and dissemination needs should be encouraged. In succeeding years, the NLM may be in a position to expand its 5   The scope of the committee's charge did not permit extensive treatment of information systems. For this discussion the committee drew on the recently published report and background materials of the IOM Committee on Improving the Patient Record (IOM, 1991b) as well as on the committee's site visits and the members' collective experience. The IOM report covers the technical issues of computer-based patient records (CPRs) and CPR systems in some detail.

OCR for page 83
Guidelines for Clinical Practice: From Development to Use core responsibilities and activities beyond the collaboration with AHCPR to other guideline development organizations.6 The committee also favors efforts to foster the translation and movement of guidelines into computerized decision aids of various sorts. However, it believes those efforts should be in conformance with emerging standards in the computer industry that will permit the guidelines (however transformed) to be used on many different types of computer-based equipment and systems (Gabrieli, 1991; IOM, 1991b; McDonald et al., 1991; Megargle, 1991). Although comprehensive computer-based patient records and systems do not at present exist, parts of them have been implemented in a small number of health care institutions around the country. Likewise, sophisticated computer-assisted applications of particular guidelines are not common but can be found in a number of institutions at the level of individual departments, units, or practitioners. More and better use of computer-based information and decision support systems for all these purposes depend on several technical and behavioral developments that are already in motion: improvements in the logical consistency and completeness of guidelines through use of algorithms and formalized formats; technological improvements (e.g., voice-recognition systems to reduce data input chores); integration of records of an individual patient's episodes of care into a single system; rules of syntax, data base structures, and communication links that support multiple users and settings of care; reductions in hardware and software costs; changes in attitudes and skills of prospective institutional and individual users including patients as well as practitioners; resolution of confidentiality, privacy, and security concerns; and demonstrated clinical utility and practicality of the support provided. Predicting the specific rate of change along these various dimensions is beyond the capacity of this committee. The technical developments (including voice recognition and better networking hardware and software) are more predictable, and closer to hand, than the behavioral changes. The 6   The IOM has recently completed a study to advise the NLM on new and expanded services for health services research and technology assessment (IOM. 1991c). These services may lie primarily in the area of core library activities related to creating and maintaining literature-oriented data bases, supplying basic guidance to regional and organization-based libraries on sources of information, and suggesting likely information search and retrieval strategies. Practice guidelines are seen as lying well within these clinical evaluation fields.

OCR for page 83
Guidelines for Clinical Practice: From Development to Use IOM committee studying the computer-based patient record called for concentrated public-private efforts for the remainder of the decade, many of them to be focused on attitudes and behaviors of users of computer-based patient records and systems. By the end of that time, that committee believed, computer-based information systems in health care could be both widespread and ingrained in the clinical life of many practice settings. SUMMARY Among the supporting conditions for the effective application of sound practice guidelines are educational programs and information and decision support systems. The first is tied closely to the dissemination of guidelines but goes far beyond that one role to promote understanding of the evidence base, rationale, and expected consequences of guidelines. Guidelines that are clearly written, specific, based on evidence, and well documented can be powerful tools of medical education, although this application is more potential than real at this time. A critical adjunct to education is to incorporate guidelines into routine information and decision support systems. Both education and decision support have a concrete place in quality, cost, and risk management initiatives. How these programs can support and be supported by practice guidelines is examined next.