Modelling emergency response times for Out-of-Hospital Cardiac Arrest (OHCA) patients in rural areas of the North of England using routinely collected data

Abstract Background National response time targets for ambulance services are known to be more strongly maintained in urban areas compared to rural. That may mean that responses in rural areas could be less immediate which can in turn affect survival of those experiencing cardiac arrest. Thus, analysis of variation in response times using routinely collected data can be used to understand which rural areas have the highest need for emergency intervention. In this study we have focused, given the heterogeneity of demographic make up, on a specific area of the North of England. Some areas in North England have shown to have a large proportion of cardiac arrests occurring in a rural setting, specifically, in the anonymised study region this was almost half of the cases at 46.3%. Response times to these areas were found to be over 3.5 minutes slower than for urban areas making it worthy of further exploration. Methods A retrospective observation analysis was conducted on routinely collected data from regional ambulance services for areas within the North of England from April 2016 to March 2021. Information was collected on service and geographic characteristics for 1915 incidents. A multivariable linear mixed effect regression model was used to understand the association between geographical, service factors and response times to cardiac arrest patients. To advance previous research which up to now only used visualisations to analyse ambulance response times, the study used a mixed effects model with a variety of predictors, capturing geographical variation alongside service characteristics. Results From the cases analysed it was found that the mean response time to scene was 9.1 minutes, with a standard deviation of 6.4 minutes. After adjustment for geographic variation and incorporating robust standard errors into the model: distance to the nearest ambulance station (coefficient = 0.61, 95% confidence interval [CI]: 0.56-0.66), urgency of the call (Category 2, second most urgent, compared to the most urgent coefficient = 1.66, 95% CI: 1.13 - 2.18), location of the nearest ambulance station to the incident and the type of crew who attended the incident (Advanced Paramedic when compared to just Paramedic, coefficient = -0.70, 95% CI: -1.24 - -0.16) were all factors which affected response times to scene. Conclusion For each extra km the incident was away from an ambulance station, the response time to scene increased by 37 seconds. The ambulance station which displayed the largest increase in response time, Station L was 170 seconds (95% CI: 79, 261) longer than Station N, which had a median performance across all stations, as measured by median survival rate to return of spontaneous circulation (ROSC). The rural geography of the North of England means that lots of cardiac arrest incidents occur a considerable distance away from the stations, emphasising the need to use alternative emergency services technologies within these rural areas to attend to patients sooner.