Use of artificial intelligence in the out-of-hospital care settings: a scoping review.

Background: Out-of-hospital services face significant challenges, including growing patient demand, workforce limitations and evolving care pathways. Artificial intelligence (AI) technologies offer potential solutions, but their application in out-of-hospital settings remains inconsistently implemented and poorly understood. Objective: To identify the types of AI technologies being applied in out-of-hospital settings, explore their purposes and implementation contexts and examine associated outcomes. Methods: Six electronic databases were searched for English-language studies published between 2013 and 2024. Eligible studies involved AI technologies in the out-of-hospital emergency services setting. Data were synthesised according to six implementation domains: system level, dispatch zone, response zone, on-scene zone, onward prognosis and inferential (insights). Results: From 236 publications, we identified diverse AI applications across the care pathway. System-level implementations (46 studies) featured AI for demand forecasting, optimal resource allocation and strategic facility location, with demonstrated improvements in coverage efficiency of 10-20%. In the dispatch zone (32 studies), AI-enhanced emergency call triage and ambulance allocation reduced response times by up to 10-20%. Response-level applications (43 studies) included intelligent traffic management and real-time route optimisation, reducing travel times by 15-30%. On-scene zone implementations (75 studies) supported clinical decision-making with cardiac arrest rhythm detection, achieving an area under the curve (AUC) values exceeding 0.90 and acute coronary syndrome prediction sensitivities of 85-90%. Onward prognosis models (19 studies) predicted patient outcomes with some AUC values of 0.80-0.90 for survival forecasting, enabling better resource allocation and early intervention. Further inferential analysis applications (21 studies) were also identified that provide higher-level insights through secondary analyses of out-of-hospital data. Conclusions: AI demonstrates significant potential across the care pathway, from operational optimisation to clinical decision support. Future development should focus on real-time adaptive systems, ethical implementation, improved data integration across the care continuum and rigorous evaluation of real-time patient outcomes. Cross-disciplinary collaboration and standardised reporting of AI implementations will be essential to realise the full potential of these technologies in improving out-of-hospital care delivery.