Validation of a TIA recognition tool in primary and secondary care: implications for generalizability

In spite of public education campaigns, patients with transient ischemic attack still present to primary care where accurate recognition of transient ischemic attack is essential so that specialist referral can be expedited to reduce stroke risk. The complex task of diagnosing transient ischemic attack is challenging in time-limited settings in primary care yet the potential for a transient ischemic attack recognition tool to improve diagnosis has not been determined.

We set out to examine the potential utility in primary care of the only existing transient ischemic attack recognition tool.

All primary care referrals with suspected transient ischemic attack from a total population of 91,000 people were analyzed over a four-year period from 2002 to 2006. Clinical and research records from the transient ischemic attack clinic and consultation notes and referral letters from primary care physicians were used to populate the Dawson recognition score.

Of 513 referrals, 209 (47%) had a clinic confirmed diagnosis of transient ischemic attack. Agreement between primary care assessments and specialist assessments was greater for speech disturbance (kappa 0·68) than for unilateral facial weakness (0·58) and unilateral limb weakness (0·51). The Dawson score had greater accuracy in diagnosing all transient ischemic attack in specialist assessments than in primary care assessments (c statistics 0·80 vs. 0·70, P < 0·0001) and performed particularly poorly in primary care for detecting posterior circulation territory transient ischemic attack with a c statistic (95% confidence interval) of 0·52 (0·43-0·61).

The Dawson transient ischemic attack recognition score is less accurate in primary care than in its derivation setting of specialist care. Improving the recognition of transient ischemic attack by providers of first contact health care requires derivation of rules in the clinical setting in which they are to be used.