Introduction: Transient neurological symptoms are a common presentation in emergency departments (ED). Access to stroke specialists and advanced imaging to identify acute cerebrovascular syndrome is resource intensive. Diversion to specialized TIA units improves logistics but incurs delay. Observational studies suggest efficacy of TIA units but are limited by sample size. We hypothesized that management through a centralized TIA service reduces risk of recurrent stroke.
Methods: Vancouver Island (VI) introduced a centralized TIA unit in 2004 to which over 15,000 referrals (ED & GP) have been made. Long-term follow-up is possible as all VI residents (∼750,000) have one electronic health record for ED visits and hospital admissions. Large referral volumes and variable unit capacity have subjected patients to a wide range of delays-to-treatment. We used variation in secondary prevention delays to examine unit efficacy. We assessed hospitalized stroke and stroke death in all ED-referred patients comparing those seen in the unit within 90 days of symptom onset with those not seen within that time.
Results: Between 2005 and 2013 there were 11,330 referrals, of which 4,017 were from the ED and referred within 2 weeks of symptom onset. The transition times between symptom onset, referral, arrival at the unit and stroke (or censoring time) were modeled using a multi-state model (Putter et al. 2007). Age, ABCDD, gender and the unit intervention were found to be associated with transition times. The hazard of stroke for patients who attended the unit was estimated as 30% that of those who did not attend (p=0.098). Predicted stroke-free survival curves are shown for a 70 year old woman. ARR at 90-days ranged from 4.2% downto 1.4% for High to Low ABCDD respectively.
Conclusion: This large observational study reinforces published studies suggesting TIA units reduce risk of recurrent stroke within 90 days.