The effectiveness of a modified version of the Wilkinson questionnaire in screening for TIA and minor stroke in the United Kingdom.

BACKGROUND: Transient ischaemic attacks (TIAs) and minor strokes are important to diagnose as there are effective secondary preventive interventions. Significant under-reporting by patients occurs, but general practitioners tend to over-diagnose the condition, contributing in part to long clinic waiting lists. An accurate screening test could address both these problems.
METHODS: A modified version of a questionnaire designed to detect TIA was tested against the gold standard of specialist diagnosis in two vascular outpatient clinics in Leicester, UK.
RESULTS: The questionnaire was sent by post with the clinic appointment and completed by 136 participants. In 99 cases the same questionnaire was administered by a clinic nurse. Overall levels of agreement (kappa, 95% CI) with specialist diagnosis of TIA, stroke or neither were 0.32 (0.15, 0.48) and 0.31 (0.12, 0.50) for postal and administered questionnaires, respectively. When the diagnoses of TIA and stroke were combined, agreement rose to 0.38 (0.23, 0.53) for postal and 0.38 (0.20, 0.57) for administered versions. For this outcome, the postal version had a sensitivity of 0.56 (0.43, 0.68) and specificity of 0.81 (0.71, 0.90). Equivalent figures for administered questionnaires were 0.61 (0.46, 0.76) and 0.76 (0.63, 0.87).
CONCLUSION: The questionnaire has potential in prioritising outpatient referrals but is not sufficiently specific to be used for research or population screening.