Rigour in Medical Diagnostic Accuracy

We need how to exactly evaluate the value of radiological and laboratory exams. Three basic definitions are:1. Population on wich the exam has been evaluated;2. Criteria of normal or abnormal results;3. The frame of reference. Main tools are sensitivity, specificity, likelihood ratio and receiving operative characteristics (ROC) curves. However, these tools did not answer to the clinician question: if the exam is abnormal what is the probability that the patient has the disease, and if the exam is normal what is the probability that the patient has not the disease Answers to these questions are supplied by bayesian predicitive values. It is pointed out that:1. Increasing the number of exams increases the statistical risk of type 1 of error;2. It is necessary to evaluate the decisional-making value of exams;3. Physicians must take in account cost-utility.