[Decision making analysis and the diagnosis of approximal caries]

A decision tree approach to dental caries diagnosis has been described, using information obtained from the literature on caries prevalence and on the accuracy of dentists when using a mirror and probe or bitewing radiographs for diagnosing doubtful approximal carious lesions. In this example, three different diagnostic pathways are compared with the pathway 'no dental examination'. The expected value of clinical information was calculated for the three pathways as were the predictive values, positive and negative, for the various test combinations. In a sensitivity analysis the caries prevalence and diagnostic accuracy were varied for the above calculations within realistic levels. Extrapolations from the results indicated that the procedure 'probe and look' is counter productive at low caries prevalence, since, for all pathways, diagnosing 'sound' surfaces as having dentine caries becomes an increasing problem. This decision making approach highlights the fact that more information on the diagnostic process is needed before recommendations about changing diagnostic criteria, pathways or techniques are made. In particular the diagnostic accuracy of dentists using radiographs and especially the weighting they put on different diagnostic outcomes should be investigated.