B. Alexander White, D.D.S., Dr.P.H., M.S., and Gerardo Maupome, Ph.D.:
Preceding presentations have reviewed the scientific literature on diagnosis and management of dental caries, indicators of risk, primary prevention of dental caries, and methods of stopping or reversing early carious lesions. For the practicing dentist, however, such data may not address specific clinical questions that arise in everyday practice. The purpose of this paper is to describe a framework�clinical decision-making�and its potential application to diagnosis and management of dental caries. Subsequent papers will use this framework to describe clinical decision-making for coronal caries in the primary dentition and coronal and root caries in the permanent dentition.
Clinical information is imperfect, yet dentists are expected to make decisions about patient care every day. Patients vary in clinically important ways, uncertainty abounds in diagnostic and prognostic information, and the effectiveness of many preventive and treatment alternatives has not been formally assessed. Scientific information is not available�and likely will never be available�to answer all important clinical questions. Clinical decisions therefore will continue to be made based (at least in part) on probabilistic, as contrasted with definitive, information.
Clinical decision-making�explicit use of information to quantify probabilities and outcomes under conditions of uncertainty�can provide a framework to analyze the impact of uncertainty in clinical information. Clinical decision-making is not descriptive, in that it does not seek to identify the ways in which clinicians actually make decisions. Rather, it seeks to identify how clinical decisions should be made to achieve optimal outcomes.
Clinical decision-making in dental caries management involves four basic steps. First, the clinical question must be identified and characterized. In this step, the relevant population for study (e.g., children, adolescents, adults, elderly) and alternative diagnostic, preventive, and management options are identified. For clinical decision-making to be useful, the clinical question must involve choosing between two or more clinical strategies with meaningful tradeoffs. Clinical questions may focus on such topics as caries detection, including diagnostic techniques and clinical examination; characterization of caries risk status; primary, secondary, and tertiary prevention of dental caries; and arresting or reversing a carious lesion.
Second, the decision problem is structured to address the relevant clinical problem. A model or decision tree that represents the logical and temporal sequence of caries management is described. The decision tree should be sufficiently complex to reflect important events and outcomes associated with the clinical problem, yet sufficiently simple to be understandable and useable. A well-defined clinical starting point must be specified, including such dimensions as age and sociodemographic characteristics; caries risk status; prior and current caries experience; behavioral factors; diet; fluoride exposure; and general health status, including use of xerostomic medications and diseases that may affect salivary gland function. The relationship of relevant diagnostic, preventive, and/or treatment strategies should be identified, and important outcomes�biological, clinical, psychosocial, and economic�described.
Third, the information needed to answer the clinical question is characterized. Much of this information comes from systematic reviews of a literature ideally based on randomized clinical trials. An important feature of the information is its probabilistic nature. Here, the probability of different events (e.g., detection of a carious lesion with a particular diagnostic test, reversing a demineralized lesion), the outcomes associated with those events (including patient preferences regarding the outcome), and the degree of associated uncertainty, are quantified.
Finally, a preferred course of action is chosen, based on the decision tree structure and relevant probability and outcome data. Synthesis of this information does not identify a "correct" course of action, but rather a "preferred" course of action that would yield the best outcome, given the information. Since uncertainty is associated with the probability and outcome estimates, a sensitivity analysis must be done to assess the impact of uncertainty on the conclusions. In some instances the preferred course of action will be robust over a wide range of probability and outcomes estimates. In other cases the preferred course of action will change within a narrow�but clinically important�range of probabilities and outcomes, suggesting that additional information is needed to more fully characterize the clinical problem.
Detsky AS, Naglie G, Krahn MD, Naimark D, Redelmeier DA. Primer on medical decision analysis: Part 1 - Getting started. Med Decis Making 1997;17:123�5.
Detsky AS, Naglie G, Krahn MD, Redelmeier DA, Naimark D. Primer on medical decision analysis: Part 2 - Building a tree. Med Decis Making 1997;17:126�35.
Krahn MD, Naglie G, Naimark D, Redelmeier DA, Detsky AS. Primer on medical decision analysis: Part 4 - Analyzing the model and interpreting the results. Med Decis Making 1997;17:142�51.
Naglie G, Krahn MD, Naimark D, Redelmeier DA, Detsky AS. Primer on medical decision analysis: Part 3 - Estimating probabilities and utilities. Med Decis Making 1997;17:136�41.
Richardson WS, Detsky AS, for the Evidence-Based Medicine Working Group. Users� guides to the medical literature. VII: How to use a clinical decision analysis. A. Are the results of the study valid? JAMA 1995;273:1292�5.
Richardson WS, Detsky AS, for the Evidence-Based Medicine Working Group. Users� guide to the medical literature. VII: How to use a clinical decision analysis. B. What are the results and will they help me in caring for my patients? JAMA 1995;273:1610�3.
Sox HC, Blatt, MA, Higgins MC, Marton KI. Medical decision making. Boston: Butterworth-Heinemann, 1988.
Weinstein MC, Fineberg HV. Clinical decision analysis. Philadelphia: W. B. Saunders, 1980.