- Question 1: Is this test potentially relevant to my practice? (Greenhalgh, 1997)

Sackett and colleagues call this the utility of the test Evenif this test were 100% valid, accurate, and reliable, wouldit help me? Would it identify a treatable disorder? If so, wouldI use it in preference to the test I use now? Could I (or my patients orthe taxpayer) afford it? Would my patients consent to it? Wouldit change the probabilities for competing diagnoses sufficientlyfor me to alter my treatment plan?

- Question 2: Has the test been compared with a true gold standard? (Greenhalgh, 1997)

You need to ask, firstly, whether the test has been comparedwith anything at all. Assuming that a "gold standard" test hasbeen used, you should verify that it merits the description,perhaps by using the questions listed in question 1. For manyconditions, there is no gold standard diagnostic test. Unsurprisingly,these tend to be the conditions for which new tests are most activelysought. Hence, the authors of such papers may need to developand justify a combination of criteria against which the newtest is to be assessed. One specific point to check is thatthe test being validated in the paper is not being used to definethe gold standard.

- Question 3: Did this validation study include an appropriate spectrum of subjects? (Greenhalgh, 1997)

Although few investigators would be naive enough to select only,say, healthy male medical students for their validation study,only 27% of published studies explicitly define the spectrumof subjects tested in terms of age, sex, symptoms or diseaseseverity, and specific eligibility criteria. Importantly, thetest should be verified on a population which includes mildand severe disease, treated and untreated subjects, and thosewith different but commonly confused conditions.Although the sensitivity and specificity of a test are virtuallyconstant whatever the prevalence of the condition, the positiveand negative predictive values depend crucially on prevalence.This is why general practitioners are skeptical of the utilityof tests developed exclusively in a secondary care population,and why a good diagnostic test is not necessarily a good screeningtest.