Background: Identification of the underlying cause of stroke and transient ischaemic attack (TIA) is important so that preventative therapy can be used to reduce the risk of recurrence. Transthoracic echocardiography (TTE) and transoesophageal echocardiography (TOE) are diagnostic tools used to identify those cardiac sources of stroke that may respond to treatment. Objectives: (1) Undertake systematic reviews to determine (a) the prevalence of cardiac sources of stroke and TIA and (b) the diagnostic accuracy of echocardiography; (2) undertake a survey to ascertain which guidelines and management strategies are used by UK stroke centres; and (3) evaluate the cost-effectiveness of the addition of TTE to the routine assessment of patients who have had a first-episode diagnosed stroke or TIA in the UK. Data sources: Bibliographic databases including MEDLINE, EMBASE, the Cumulative Index to Nursing and Allied Health Literature, PsycINFO and the NHS Economic Evaluation Database were searched from inception to December 2010 (prevalence) or September 2011 (diagnostic accuracy). Bibliographies of related papers were screened and experts were contacted to identify additional published and unpublished references. Review methods: The systematic reviews were undertaken according to the general principles recommended in the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) statement. A decision-analytic model was developed to estimate the costs and quality-adjusted life-years accrued by each potential echocardiography strategy in the management of stroke and TIA for patients aged 45, 55 and 65 years. The model took a lifetime horizon and a NHS perspective. Costs and health benefits were discounted at an annual rate of 3.5%. Evidence to enable modelling was found for left atrial thrombus only. The cost-effectiveness of echocardiography is therefore based on all stroke patients being tested but only those with a left atrial thrombus receiving the benefits and harms of treatment. To describe current NHS stroke management practice we provided a questionnaire to the lead clinician of all stroke units in the UK. Results: The searches identified 17,278 citations for the systematic review of the prevalence of potential cardiac sources of stroke and TIA, of which 65 studies were included. Patent foramen ovale was the most frequently reported pathology, followed by atrial septal aneurysm and mitral valve prolapse, with prevalence ranging from 0.25% to 73%, from 0.4% to 28% and from 0% to 31.6% respectively. For the systematic review of the diagnostic accuracy of echocardiography, 16,504 citations were identified, of which 51 studies were included. The pooled sensitivity to detect left atrial thrombus in three studies using transthoracic echocardiography in second harmonic imaging mode (TTEh) was 0.79 [95% credible interval (CrI) 0.47 to 0.94], with a pooled specificity of 1.00 (95% CrI 0.99 to 1.00) compared with TOE. Differences in the diagnostic accuracy of tests occurred mostly in their sensitivity to detect cardiac sources of stroke. No adverse events data were reported. Our principal economic finding is that TTEh is a cost-effective use of NHS resources compared with TOE when clinicians deem it the most appropriate test. The survey showed that the decision-making process for the management of stroke and TIA is very complex and varies considerably by site. It is clear that to accurately describe current management practice a very sophisticated questionnaire would be required. Limitations: The prevalence review highlights the difficulties that clinicians face when identifying the cause of cardioembolic stroke (the limitations of the tests, the confounding comorbidities and the inherent mobility of blood clots). The diagnostic accuracy review was limited by the small number of studies reporting data or because studies included too few participants with a cardiac pathology, leaving a large degree of uncertainty about the underlying diagnostic accuracy. The economic model has limitations because of the limited data available for important parameters such as the efficacy of treatment in reducing stroke recurrence. Conclusion: The economic analysis indicates that, in those cases in which TTEh is deemed the most appropriate test, it is a cost-effective use of NHS resources. However, this analysis has highlighted a lack of evidence in several areas and the results of the economic evaluation should therefore be treated with caution. There is a need for further evaluation of current echocardiography technologies, the causal associations between potential risk factors and stroke and whether or not anticoagulation therapies prevent recurrent stroke. Studies attempting to establish the prevalence of cardiac sources of stroke should identify all potential risk factors, rule out those that are not relevant and grade the findings according to risk. Research is also needed to reduce the uncertainty around the estimates of the sensitivity and specificity of TTEh and TOE, singly and in combination, in detecting treatable cardiac abnormalities compared with the 'gold standard' in each pathology.