Stroke and emergency department re-presentation after outpatient TIA management: an interrupted time series study.

Objectives: To assess the effects of a non-admitted management pathway following emergency department (ED) presentation with suspected TIA on: 90-day stroke and ED re-presentations, overnight admission, length of stay (LOS) and costs. Methods: We implemented a management pathway across an Australian regional health service (4 hospitals; 2 rural, 10,000 km²) including ED protocols followed by urgent outpatient review or telemedicine consultation to one rural hospital. Interrupted time series analysis was conducted on linked hospital administrative datasets for all ED TIA diagnoses 5 years before and 2 years after intervention (2015). We assessed whether pathway introduction was associated with immediate change (level) or subsequent rate of change (slope) in outcomes. Results: There were 2031 presentations: 1,467 before, 564 after implementation. Against background declining trends, overnight admissions decreased by 12.4% (95%CI 5.0, 19.7) and total LOS decreased 6 hours (95%CI 1.5, 10.4). Hospital costs reduced by AUD683 per patient with implementation. Outpatient review occurred for 36% at median 5 days (IQR 3, 9), including 19/87 (22%) telemedicine reviews. Pathway adherence was incomplete: 29% had no specialist review. Recurrent stroke increased by 1.3/100 presentations (95%CI 0.6, 2.1) with implementation, then returned to baseline of 0.9/100. ED re-presentations rose at a significant rate after implementation (extra 1.69/100 patients re-presenting/quarter; 95%CI 0.8, 2.6) reaching 32/100. Conclusions: An ED TIA management pathway designed to avoid hospital admission resulted in decreased hospital use and costs; but an initial increase in recurrent stroke and sustained rise in ED re-presentation, possibly related to delayed and incomplete follow-up.