Introduction
Did the authors include an appropriate spectrum of patients? Is there any spectrum, work-up, expectation or observer bias?
Bias is a systematic difference between the groups under study. Ask yourself whether the two groups were treated in the same way at all the stages in the study.
Spectrum bias results when the subjects under study are from the extreme ends of the patient spectrum. For instance, if they had only included cancer patients with huge swollen, asymmetric legs and teenage patients who had walked in, but excluded slightly swollen legs in older patients, the reported sensitivities and specificities would be artificially inflated. A diagnostic test should be at its most useful when there is real diagnostic uncertainty.
Expectation bias exists when the researcher recording the information is aware of all the clinical details influencing the diagnosis. Ideally, people who perform assessments of clinical features should be blinded to all the other features of the individual.
Observer bias results when the people recording the data are influenced by their own knowledge.
Work-up bias exists when the gold standard test is not applied to subjects who test negative and positive.