Background A variety of tests are used in clinical practice to help the diagnostic process and so improve patient care. Many aspects of stroke management depend on accurate and rapid diagnosis. Brain imaging, including CT or MRI, is necessary to identify the location and extent of the cerebral lesion, and to determine the pathological type of stroke and its likely cause. Current treatments - such as thrombolysis - for ischaemic stroke have increased the need for clear evidence on which imaging test is optimal for diagnosis in the acute phase of stroke. Systematic reviews of diagnostic test accuracy may provide evidence on the best use of a diagnostic test in clinical practice and help clinicians to decide among alternative tests. The Cochrane Collaboration has recently included systematic reviews of diagnostic test accuracy within its remit. However, to prepare Cochrane systematic reviews of diagnostic test accuracy is challenging because the methods for such reviews are still in a state of flux. Materials and methods The research work undertaken for this thesis addresses four relevant methodological aspects of such reviews and, I hope, will contribute to informing the development of the methods for Cochrane systematic reviews of test accuracy: i) I assessed the quality of reporting of imaging studies in stroke medicine published between 1995 and 2008 with the current STAndards for the Reporting of Diagnostic accuracy studies (STARD) criteria; ii) I assessed the magnitude of publication bias in diagnostic accuracy studies in stroke medicine, by reviewing all diagnostic abstracts presented at two international stroke meetings between 1995 and 2004 and so evaluating the characteristics and findings of the identified abstracts; iii) I have evaluated the methods for preparing reviews of test accuracy by undertaking a pilot review according to the draft recommendations of the Cochrane Diagnostic Test Accuracy Working Group; iv) I conducted a survey to assess a) how well clinicians and health professionals interpret findings of Cochrane systematic reviews of diagnostic accuracy presented in summary documents; and b) what is the best format for summarising findings of Cochrane reviews of diagnostic accuracy. Conclusions In conclusion, methodological issues concerning the validity and reliability of findings of studies included in systematic reviews of diagnostic accuracy remain of fundamental importance. More empirical evidence is needed to address potential biases such as reporting bias and publication bias. To allow dissemination of diagnostic reviews findings in clinical practice better ways of communicating main characteristics and key results of systematic reviews of diagnostic accuracy should be considered. In the current literature, the quality of reporting and methodological quality of imaging studies for the diagnosis of stroke is less than satisfactory and leaves room for improvement. This is worrying, especially if current health imaging policies are in fact based on poor quality evidence and hence scarce health resources may not being deployed as effectively as they could be.