Access and timeliness of acute stroke reperfusion after cluster-wide protocol standardisation: an interrupted time-series study in Hong Kong

BACKGROUND: Protocol standardisation is widely used to improve the quality and consistency of acute stroke reperfusion, but evidence from routine public hospital practice remains limited, particularly on whether gains are sustained during later periods of service stress. We evaluated changes in access to and timeliness of intravenous thrombolysis after cluster-wide protocol standardisation in a Hong Kong public hospital cluster and examined whether these performance domains changed during a later Omicron-related service shock. METHODS: In this interrupted time-series study, we analysed acute ischaemic stroke care in the Kowloon West Cluster, Hong Kong, from 1 January 2017 to 31 December 2024. Prespecified breakpoints were implementation of a unified reperfusion protocol in July 2019 and the Omicron-related service shock in February 2022. Primary monthly outcomes were intravenous thrombolysis rate, median door-to-needle time, and the proportion treated within 60 minutes. Secondary process outcomes were door-to-computed tomography time and imaging-to-needle interval. Secondary quarterly patient-centred outcomes were early neurological improvement at 24 hours and 3-month functional outcome measured using the modified Rankin Scale. RESULTS: Protocol implementation was associated with an immediate increase in intravenous thrombolysis access. Median door-to-needle time showed a more complex pattern, with an immediate worsening after protocol implementation followed by a favourable post-implementation trend. The proportion treated within 60 minutes improved after protocol implementation. During the Omicron period, thrombolysis access remained stable, but treatment timeliness deteriorated progressively. Decomposition of the reperfusion pathway suggested that this later erosion was driven mainly by prolongation of the downstream imaging-to-needle interval rather than by deterioration in initial imaging access. Early neurological improvement showed post-protocol improvement but attenuated during sustained service stress. Favourable 3-month functional outcome remained broadly stable, with sensitivity analyses yielding similar estimates despite declining follow-up completeness in the post-shock period. CONCLUSION: Cluster-wide protocol standardisation was associated with improved access to intravenous thrombolysis and a mixed pattern in treatment timeliness. Implementation effects on timeliness were not linear and included early disruption, subsequent improvement, and later deterioration during prolonged service stress. Preserved treatment access should not be assumed to indicate preserved treatment speed. Quality evaluation of acute stroke services should distinguish access from timeliness and use pathway-level measurement to identify workflow segments most vulnerable to disruption.