To treat or not to treat? The decision to administer or withhold treatment based on test results could have far reaching effects on the health of a patient. The evidence supporting the accuracy of medical tests introduced in practice should therefore be credible and generated with valid and robust methods. Eleanor Ochodo looks at how the recommended methods of diagnostic accuracy have been applied in reports of diagnostic test accuracy. She specifically focuses on the methods of reporting of primary studies and on methods for assessing risk of bias, performing meta-analyses and investigating publication bias in systematic reviews of diagnostic test accuracy.