Practice guidelines are being introduced throughout medicine, but expectations about their impact on patient care depend on whether one is a clinician, patient, payer, administrator, or politician. Proponents hope that guidelines will enhance the knowledge, attitudes, and behavior of practitioners and will optimize health outcomes, costs, and malpractice decisions, but scientific evidence of these effects is limited. There are also concerns that guidelines could harm patient care. Clinicians worry that guidelines will promote "cookbook medicine," decrease their autonomy and income, and increase medicolegal liability. A particular concern relates to the expansion of enforcement programs that require clinicians to follow guidelines or face financial or other penalties. Guidelines can rarely define optimal care with certainty, due to poor science, imperfect analytic processes, and differences in patients. Recommendations are often worded in highly specific language that achieves clarity at the expense of scientific validity. Rigid enforcement of such guidelines could harm patients, interfere with the individualization of care, increase costs, and promote unfair judgments against clinicians who deviate from them for good reasons. A model that links the intensity of enforcement to the scientific and clinical quality of guidelines is proposed.