Aims:
Various models of telestroke care have been implemented in Aotearoa NZ (NZ) and internationally. This study aimed to assess the impact of telestroke models of care on door-to-needle times and patient outcomes.
Methods:
This is a registry-based, multicentre, observational target trial emulation. Data was derived from the NZ National Stroke, Telestroke, and health administrative data registers. We used logistic regression to assess the effect of video versus audio-only assessment and explored differences in image viewing platforms, workflow order, computed tomography perfusion (CTP) usage and hub hospital characteristics, aiming to define an optimal model of telestroke care.
Results:
Our main analysis included 975 patients treated within one of NZ’s four telestroke networks between 1 March 2022 and 31 December 2024. Patients presenting to video-supported telestroke hospitals (n=457) were 2.29 (1.29-4.04) and 1.98 (1.06-3.71) times more likely than those presenting to audio-only hospitals (n=518) to achieve a door-to-needle time <60 minutes and independence at 3 months, respectively. Video telestroke hospitals had a thrombolysis rate of 16.7%, nearly double that of audio-based hospitals at 8.8%. Telestroke care provision by stroke physicians from thrombectomy centres, access to CTP imaging, and access to full sequence Picture Archiving and Communication System (PACs) were associated with better patient outcomes.
Conclusion:
Compared with audio-only telestroke services, video-based telestroke networks achieved faster treatments and improved patient outcomes. Patient outcomes were further enhanced through access to CTP, thrombectomy centre telestroke support, and access to full PACs imaging. These results should inform the development of future telestroke models.