Inferential statistical methods for strengthening the interpretation of laboratory test results

Laboratory test results for the diagnosis of psychiatric illness usually are reported descriptively despite the ready availability of appropriate inferential statistics. A test's sensitivity, specificity, and diagnostic confidence are conditional probabilities. Confidence intervals may be calculated for these probabilities in any given study. Statistical tests for comparing the results of several studies use techniques for planned and posterior comparisons applied to contingency tables. These established statistical methods aid in the interpretation of laboratory test findings.