Abstract

Clinical guidelines often make recommendations on the use of diagnostic tests. Compared with sensitivity and specificity, the use of pre- and post-test probabilities allows a more explicit and rational selection and interpretation of diagnostic tests. Ideally, clinical guidelines relating to diagnosis should routinely incorporate this information to enhance individualised decision making. We report our experience of incorporating pre- and post-test probabilities into a guideline on the investigation of women with postmenopausal bleeding developed by the Scottish Intercollegiate Guidelines Network. Issues relating to their application are highlighted, including the limitations of available evidence on diagnostic tests and prevalence of disease, acceptability to guideline users, and the uncertain impact on actual clinical decision making. Despite these potential difficulties, the incorporation of data on pre- and post-test probabilities into the development and presentation of guideline recommendations may offer an important opportunity to make clinical decision making more transparent for both clinicians and patients.