How to Assess Quality of Primary Research Studies in the Medical Literature?

Abstract

Accuracy studies are the cornerstone of the evaluation of new diagnostic tests. In an accuracy study, patients with a clinical suspicion of disease undergo both the new tests that is being evaluated, and the reference test or "gold-standard" test for the disease. Patients are stratified according to the test result and to whether or not the diagnosis was confirmed by the reference test. Main accuracy indices include the sensitivity (proportion of patients who test positive among those who have the disease), specificity (proportion of patients who test negative among those without the disease), as well as the positive and negative predictive values (proportion of patients with the disease among those with a positive test, and of patients without the disease among those with a negative test, respectively). In appraising an accuracy study, the reader should check that the study design and analysis follow methodological standards. The study population should be representative of the population in which the test will be used in practice. The new test should be interpreted in a blinded and independent fashion from the reference standard to avoid expectation bias, and should not be used to establish the diagnosis (incorporation bias). The reproducibility should be verified. Interpretation criteria and technical aspects should be described with enough details to allow replication. Provided that these conditions are met, the next step is to decide whether the test may be used for patient care. The clinical setting in which the test will be used, and the corresponding pretest probability of disease, will determine how best can the test be used in practice.