Annemarei Ranta, Mark Weatherall, John Gommans, Murray Tilyard, Des Odea, Susan Dovey

AIMS Many transient ischaemic attack (TIA) patients receive initial assessments by general practitioners (GPs). In a randomised controlled trial (RCT) we showed that BPAC Inc. TIA/stroke electronic decision support (EDS) for GPs improves patient outcomes and guideline adherence. This secondary analysis assesses the impact of trial associated enhanced GP access to radiological investigation. METHODS Post-hoc analysis of a multi-centre, single blind, parallel group, cluster RCT comparing TIA/stroke EDS guided GP management with usual care to assess whether imaging requests and their appropriateness differed between study groups. RESULTS GPs requested 15/291 (5.2%) carotid ultrasounds and 19/291 (6.5%) computed tomography (CT) head scans. Scans were obtained more frequently in the intervention group (ultrasound cluster adjusted OR (95% CI) 1.41 (0.44 to 4.49), P = 0.56 and CT 13.8 (1.7 to 110.7), P < 0.001). All CTs were clinically appropriate. More ultrasounds were appropriate in the EDS group (cluster adjusted OR (95% CI) of 8.4 (0.39 to 92.3), P = 0.18). Overall investigation costs did not differ between groups (P = 0.83). Some apparent avoidable imaging duplication occurred where patients were subsequently assessed by secondary services. CONCLUSION In the setting of a RCT assessing GP electronic decision support, frequency of GP initiated imaging requests was low and largely appropriate especially in the setting of EDS use. Thus enhanced GP imaging access as part of the EDS tool did not result in inappropriate or excessive GP imaging requests. However, some duplication occurred and practitioners need to ensure that test referrals and results are adequately communicated between sectors.