Chapter 003. Decision-Making in Clinical Medicine (Part 8)

Statistical Prediction ModelsBayes theorem, as presented above, deals with a clinical prediction problem that is unrealistically simple relative to most problems a clinician faces. Prediction models, based on multivariable statistical models, can handle much more complex problems and substantially enhance predictive accuracy for specific situations. Their particular advantage is the ability to take into account many overlapping pieces of information and assign a relative weight to each based on its unique contribution to the prediction in question. For example, a logistic regression model to predict the probability of CAD takes into account all of the relevant independent factors from...
Nội dung trích xuất từ tài liệu:
Chapter 003. Decision-Making in Clinical Medicine (Part 8) Chapter 003. Decision-Making in Clinical Medicine (Part 8) Statistical Prediction Models Bayes theorem, as presented above, deals with a clinical predictionproblem that is unrealistically simple relative to most problems a clinician faces.Prediction models, based on multivariable statistical models, can handle muchmore complex problems and substantially enhance predictive accuracy for specificsituations. Their particular advantage is the ability to take into account manyoverlapping pieces of information and assign a relative weight to each based on itsunique contribution to the prediction in question. For example, a logisticregression model to predict the probability of CAD takes into account all of therelevant independent factors from the clinical examination and diagnostic testinginstead of the small handful of data that clinicians can manage in their heads orwith Bayes theorem. However, despite this strength, the models are too complexcomputationally to use without a calculator or computer (although this limit maybe overcome once medicine is practiced from a fully computerized platform). To date, only a handful of prediction models have been properly validated.The importance of independent validation in a population separate from the oneused to develop the model cannot be overstated. An unvalidated prediction modelshould be viewed with the same skepticism appropriate for a new drug or medicaldevice that has not been through rigorous clinical trial testing. When statistical models have been compared directly with expert clinicians,they have been found to be more consistent, as would be expected, but notsignificantly more accurate. Their biggest promise, then, would seem to be tomake less-experienced clinicians more accurate predictors of outcome. Decision Support Tools Decision Support Systems Over the past 35 years, many attempts have been made to developcomputer systems to help clinicians make decisions and manage patients.Conceptually, computers offer a very attractive way to handle the vast informationload that todays physicians face. The computer can help by making accuratepredictions of outcome, simulating the whole decision process, or providingalgorithmic guidance. Computer-based predictions using Bayesian or statisticalregression models inform a clinical decision but do not actually reach aconclusion or recommendation. Artificial intelligence systems attempt tosimulate or replace human reasoning with a computer-based analogue. To date,such approaches have achieved only limited success. Reminder or protocol-directed systems do not make predictions but use existing algorithms, such aspractice guidelines, to guide clinical practice. In general, however, decisionsupport systems have shown little impact on practice. Reminder systems, althoughnot yet in widespread use, have shown the most promise, particularly in correctingdrug dosing and in promoting adherence to guidelines. The full impact of theseapproaches will only be evaluable when computers are fully integrated intomedical practice. Decision Analysis Compared with the methods discussed above, decision analysis represents acompletely different approach to decision support. Its principal application is indecision problems that are complex and involve a substantial risk, a high degree ofuncertainty in some key area, or an idiosyncratic feature that does not fit theavailable evidence. Five general steps are involved. First, the decision problemmust be clearly defined. Second, the elements of the decision must be madeexplicit. This involves specifying the alternatives being considered, their relevantoutcomes, the probabilities attached to each outcome, and the relative desirability(called utility) of each outcome. Cost can also be assigned to each branch of thedecision tree, allowing calculation of cost effectiveness. Typically, the data topopulate a decision model are derived from the literature, from unpublishedsources, from expert opinion and from other secondary sources. Third, thedecision model must be evaluated to determine the net long-term health benefitsand costs of each strategy being considered. Fourth, the incremental healthbenefits and costs of the more effective strategies must be calculated. Finally,extensive sensitivity analyses must be used to examine the effects on the results ofvarying the starting assumptions through plausible alternative values. An example decision tree created to evaluate strategies for screening forhuman immunodeficiency virus (HIV) infection is shown in Fig. 3-3. Up to 20,000new cases of HIV infection are believed to be caused each year in the UnitedStat ...