There is little evidence on how best to present diagnostic and treatment effect information to dentists and whether this makes any difference to their understanding of the information given and clinical management. The objective of this study is to assess the influence of different formats for presenting diagnostic test results and treatment efficacy on dentists' understanding and on their reported clinical management. The hypothesis is that presenting diagnostic test and treatment effectiveness results using natural frequencies influences the threshold for treatment and improves interpretation of tests results compared to conditional probabilities and percentage differences. A group of dentists recruited from postgraduate courses were randomized to two of four arms, being two arms of a hypothetical realistic diagnostic scenario (a) and two arms of a hypothetical realistic treatment scenario (b): a1) diagnostic test results presented in conditional probabilities; a2) diagnostic test results presented in natural frequencies; b1) treatment efficacy presented in percentages; b2) treatment efficacy presented in natural frequencies. The main outcome measures were the correct answer as to whether disease was actually present (post-test probability) and the correct answer as the number of patients that would benefit from the treatment (efficacy). Secondary outcome measures included the decision whether to treat or not treat the hypothetical patient.

Inclusion Criteria:

• Graduated dentist undergoing a postgraduate course in a public university in Rio de Janeiro.

Exclusion Criteria:

• None.