Evidence-based medicine for diagnostic questions

When searching the medical care literature for evidence on a diagnostic test, three questions should be addressed each time a study is found: (1) Is this evidence about a diagnostic test valid? (2) Does the test accurately discriminate between patients who do and patients who do not have a specific disorder? (3) Can the test be applied to this patient who is right now sitting in front of me? We will discuss hysterosalpingography (HSG) as an example of a valid and accurate diagnostic test to be applied in a general population of subfertile couples to assess tubal patency (specificity 0.83). HSG is an unreliable test for diagnosing tubal occlusion however (sensitivity 0.65). If HSG were normal, other investigations could be pursued and diagnostic laparoscopy (LS) only performed if conception had not occurred by a later date. If HSG were abnormal, LS would be needed to confirm or exclude tubal occlusion. Patients with risk factors for pelvic or tubal disease, including an abnormal Chlamydia antibody test (CAT) and those showing abnormalities at pelvic examination, should proceed directly to LS because they are significantly more likely to have pelvic pathology. A completely different issue would be HSG as a prognostic test for the occurrence of pregnancy. In theory, the occurrence of pregnancy may be considered a gold standard; however, in reproductive medicine, with so many causes of subfertility other than tubal pathology, a diagnostic test for one single disorder, if normal, will never be able to accurately predict the eventual occurrence of pregnancy.