Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
SUMMARY 14 even if they do, it is unlikely that appropriate data on patient outcomes and program costs will be available and analyzed. Instead of a full-fledged evaluation, the agency can more reasonably be expected to provide a report on its evaluation plan, the steps being taken to implement the plan, and any preliminary evidence of impact. DIVERSITY IN CLINICAL PRACTICES AND GUIDELINES In its discussions, the committee repeatedly returned to questions of diversity in clinical practice and inconsistency among guidelines. Diversity in clinical practice can be acceptable or unacceptable. It may be reasonable when the scientific evidence to support different courses of care is uncertain. In addition, some degree of diversity may be warranted by differences in individual patient characteristics and preferences and variations in delivery system capacities related to locale, resources, and patient populations. However, even though practice variation based on scientific uncertainty or differences in values may be acceptable, both science and values are open to change. Thus, what is perceived as acceptable diversity in clinical practice may change over time. Diversity in practice is unacceptable when it stems from poor practitioner skills, poor management of delivery systems, ignorance, or deliberate disregard of well-documented preferable practices. It should not be tolerated when it is a self- serving disguise for bad practices that harm people or waste scarce resources. Guidelines can clarify what is acceptable and unacceptable variation in clinical practice, but that clarification itself has limits that may lead different groups to different and even inconsistent guidelines. Weak evidence is still weak evidence, although the processes described in Chapter 3 should allow the best use of whatever evidence is available. Nonetheless, these processes still leave room for differences of expert opinion about such issues as whether a flaw in research design "matters" or whether differences in results between two treatment alternatives are "clinically important" or only "statistically significant.'' Inconsistency among guidelines can also arise from variations in values and tolerance for risk. People may simply differ in how they perceive different health outcomes and how they judge when benefits outweigh harms enough to make a service worth providing. One way to approach this kind of variation is to try to establish practitioner and patient attitudes toward different benefits and harms and then identify what is known about the probabilities of those different outcomes. In some cases, the developers of guidelines may take the further step of applying their own values, but others considering the guidelines later could look at the same information and perhaps come to different conclusions. Also, for some services and clinical