Clinical accuracy data presented as natural frequencies improve dentists' caries diagnostic inference: Evidence from a randomized controlled trial

The authors assessed whether dentists' diagnostic inferences differ when test accuracy information is communicated using natural frequencies versus conditional probabilities.A parallel, randomized controlled trial with dentists was carried out in Rio de Janeiro, Brazil. The dentists received a question on the probability of a patient having interproximal caries, given a positive bite-wing radiograph. This question was asked using information that was formulated into either natural frequencies or conditional probabilities.Only 14 (13.9%) of the dentists gave the correct answer; 13 in the natural frequencies group, and 1 in the conditional probabilities group (P .001). There were 7 nearly correct answers in the natural frequencies group and none in the conditional probabilities group (P = .005).Representing diagnostic test accuracy in natural frequencies substantially helped dentists make diagnostic inferences. Nearly twice as many dentists overestimated the presence of interproximal caries when given information in conditional probabilities.Our study findings show information shared using natural frequencies may be more accurately interpreted by dentists than that based on conditional probabilities. Patients will probably receive different standards of care depending on the format in which dentists receive diagnostic test accuracy information.