Outcomes of out-of-hospital cardiac arrest in Ireland 2012-2020: Protocol for an observational study

Out-of-Hospital Cardiac Arrest

Out-of-hospital cardiac arrest (OHCA) describes the sudden loss of mechanical function of the heart and the absence of blood flow around the body, occurring outside of a hospital setting^1,2. OHCA is the most time-critical medical emergency, with survival depending on the prompt actions of the community where OHCA occurs, emergency medical services (EMS) and in turn, appropriate follow-on hospital care³. Internationally, the annual incidence of OHCA treated by emergency medical services (EMS) is estimated to be 30 – 97 individuals per 100,000 population, median age ranges from 64 to 79 years and more than half of victims are male⁴. OHCA with attempted resuscitation has an incidence of 56 per 100,000 population per year in Europe and is considered the third leading cause of death⁵. Survival to/ immediately after hospital discharge ranges from three to twenty percent internationally and is eight percent in Europe⁶. Of those that do survive to discharge in Europe and North America, approximately seventy percent are still alive at three years⁷. In Ireland there are now almost 3,000 OHCAs with attempted resuscitation each year of which six to seven percent involve survival to hospital discharge⁸.

The Chain of Survival

The ‘chain of survival’ describes the critical actions that can link an OHCA victim with survival⁹. This chain involves early access to care, early cardiopulmonary resuscitation (CPR), early defibrillation and post resuscitation care¹⁰. CPR and defibrillation are exquisitely time sensitive interventions and thus health systems face a significant challenge in ensuring OHCA patients have timely access to these treatments¹¹. Thankfully, the simplicity, affordability and increasing awareness of these interventions mean that they can be cascaded to the community and performed by ‘bystanders’ when cardiac arrest occurs¹². Thus, at population level these represent the most important OHCA interventions^13,14. In circumstances where bystanders have not initiated CPR, dispatch assisted CPR (where the EMS emergency call taker provides CPR instruction to bystanders over the phone) can further improve survival following OHCA¹⁵.

CPR and defibrillation exist alongside several other OHCA interventions that can be delivered in the community and enroute to hospital by EMS. Randomised controlled trials have demonstrated that delivery of medications such as adrenaline and amiodarone may have additional benefits for some patient groups, while trials of airway management and mechanical CPR devices have failed to show significant survival improvements^16–19. An additional novel but resource intense hospital (or hospital outreach) treatment ‘eCPR’ (extracorporeal cardiopulmonary resuscitation) uses a mechanical device to temporarily replace heart and lung function. eCPR has shown significant promise in specific OHCA sub-populations²⁰. However, the degree to which this treatment can be made available in an equitable fashion at population level is yet uncertain^21,22. In terms of follow on in-hospital care, international guidelines now recommend that OHCA patients are cared for in specialist cardiac arrest centres, while acknowledging limited evidence for this recommendation outside of specific patient subgroups^22,23.

Out of Hospital Cardiac Arrest & Quality Improvement Registries

The science underpinning the ‘chain of survival’ is unquestionably of critical importance in OHCA care. However, scientific research alone will not be sufficient to improve OHCA outcomes and must be accompanied by a focus on real world implementation and quality improvement²⁴. To this end the European Resuscitation Council recommend that ‘health systems should have population-based registries which monitor the incidence, case-mix, treatment and outcomes for cardiac arrest’ and that these data ‘should inform health system planning and responses to cardiac arrest’²⁵.

An internationally agreed ‘Utstein’ registry dataset has been devised which facilitates national and international comparison of OHCA²⁶. A recent pan European survey reported that six countries (including Ireland) had an OHCA registry with full population coverage, fourteen had partial population coverage and seven countries reported not having an OHCA registry²⁷. Comprehensive data on the Irish experience of OHCA is provided by the Irish Out-of-Hospital Cardiac Arrest Registry (OHCAR). OHCAR achieved national coverage using the Utstein dataset in 2012, and by 2020 this contains over 20,000 cases. In terms of the period considered in this study OHCAR does not include OHCA with EMS response but without attempted resuscitation. We anticipate that approximately fifty percent of cases recorded in OHCAR are witnessed (European average 66%), and twenty percent have an initial ‘shockable rhythm’ (European average 20%)⁶. The primary sources of OHCAR data are patient care records and dispatch data from Ireland’s statutory ambulance services. OHCAR has provided invaluable continuous quality improvement data on an annual basis to ambulance services. To date however, no temporal analyses have been conducted to assess the impact of national interventions on survival in Ireland. Nor has the database been interrogated to determine whether international trends in OHCA (including increased bystander CPR and associated survival improvements^28–30) are mirrored in the Irish data.

Emergency medical services in Ireland

Ireland’s population is now in excess of five million people, for whom emergency medical services (EMS) are provided by the National Ambulance Service (NAS)³¹. In Ireland’s capital Dublin, Dublin Fire Brigade (DFB) are contracted to provide an emergency ambulance service alongside the NAS. Each year these services respond to in excess of 300,000 ambulance calls³². The National Emergency Operations Centre (NEOC) coordinates statutory emergency service responses to OHCA for most of the Irish state. In the Dublin metropolitan area EMS response is coordinated by both NEOC and the DFB east regional communications centre (ERCC). EMS care across the Irish state is provided by Paramedics or Advanced Paramedics who number approximately 2,500 & 700 respectively³³. All front-line ambulances must be staffed by at least a paramedic grade practitioner. A small number of EMS physicians supplement this care on a voluntary as available basis; however, they are not a core component of the statutory response.

Paramedic and Advanced paramedic scope of practice is determined by a statutory agency, the Pre-Hospital Emergency Care Council (PHECC). PHECC publish clinical practice guidelines and maintain a practitioner register³⁴. Only Advanced Paramedics are permitted to provide intravascular medications or to perform endotracheal intubation. Mechanical CPR devices are now widely available on all frontline ambulances. eCPR is generally not available in Ireland. In addition to statutory EMS providers, Ireland also has an extensive network of voluntary community first responders (CFRs) who can be dispatched by NEOC to OHCA to provide early CPR and defibrillation^35,36.

OHCA Health System Developments in Ireland 2012–2020

Over the period 2012–2020 that OHCAR has been in existence, Ireland has undergone several health system developments that are pertinent to OHCA care. Figure 1 summarises a timeline of these developments. The Pre-Hospital Emergency Care Council national Citizen CPR or Call-Check-Compress programme was launched in 2010³⁷. This public awareness campaign involved a series of national roadshows combined with a major national television, cinema, on-line and transport advertising designed to increase bystander CPR. Then in 2012 the NAS launched their ‘one life’ quality improvement programme³⁸. This ongoing programme focuses on several key aspects of OHCA including community interaction and public education, NEOC call taking and dispatch, EMS quality care on scene, and finally quality data management and audit processes. In 2014 the Irish Health Information and Quality Authority (HIQA) published a health technology assessment of public access defibrillation in Ireland³⁹. This was commissioned by government to inform decision making around proposed legislation to mandate public defibrillator availability. The health technology assessment estimated the clinical and cost effectiveness of a range of potential Irish public access defibrillation configurations, ranging from comprehensive to targeted³⁹. It estimated that between two and ten additional OHCA survivors could be achieved annually; however, none of the models achieved the threshold for cost effectiveness. It advised that targeted AED deployment in higher incidence locations in combination with an EMS-linked AED register and increased public awareness could potentially render the programmes’ cost effective³⁹. Ultimately the proposed legislation was abandoned, and a comprehensive national EMS-linked AED register is yet to be established.

Figure 1. OHCA Health System Developments in Ireland 2012–2020.

NAS: National Ambulance Service, QI: Quality Improvement, PAD: Public Access Defibrillation, CPR: Cardiopulmonary Resuscitation, CFR: Community First Responder, GP: General Practitioner, OHCA: Out-of-Hospital Cardiac Arrest.

In 2015 and 2016 the National Ambulance Service transitioned from a system of multiple regional independent control centres to a single national control centre (NEOC). This significant change allowed an enhanced level of resource co-ordination and further allowed dispatch assisted CPR to be fully embedded as a standard of care at national level. Between 2016 and 2020 the Irish Heart Foundation ‘CPR 4 schools’ programme trained 1,827 teachers in 531 schools to perform CPR training for a potential 288,197 students. Between 2016 and 2020 lay community first responder groups increased from 100 to 175 nationally. In 2017 the NAS created four new Community Engagement Officer posts to support CFR activities and expansion. Also between 2016 and 2020 GP (general practitioner) first responder numbers doubled to almost 200 individuals⁴⁰. The final year of this study period 2020 is exceptional in that it represented the first year of the Covid-19 pandemic, the first wave of which occurred between February and July 2020⁴¹. The Irish health service responded by introducing a range of public health measures including travel restrictions, schools and business closure, cessation of large indoor gatherings, and other social distancing measures⁴². In terms of OHCA, internationally Covid-19 is known to have detrimentally affected systems of care and was associated with prolonged EMS response and worse short-term outcomes including survival compared to pre-pandemic periods⁴³. The effects of COVID-19 on OHCA in Ireland have not yet been systematically evaluated; however similar trends are probably likely. Thus, it is necessary to consider this as a significant OHCA development in the 2020 period.

Focus of this study

This study will examine data from the Irish Out-of-Hospital Cardiac Arrest Registry (OHCAR) to describe OHCA incidence and its care processes in Ireland over the period 2012 to 2020. We will further interrogate this data to identify key OHCAR variables that predict key outcomes. We anticipate approximately 18,000 relevant observations will be available for analysis. We will consider survival to hospital discharge, bystander CPR, and bystander defibrillation to be key outcomes of interest over the study period. We hypothesise that of the timeline of OHCA developments described above, two key developments are likely to be associated with a significant change in outcomes. Given the scale of reorganisation and centralisation of EMS control in 2015 and 2016 on the backdrop of the ongoing NAS ‘one life’ project we hypothesise that the NAS transition from a system of multiple regional independent control centres to the single national control centre (NEOC) would be associated with improvements in bystander CPR and survival to hospital discharge. In turn, given Covid-19’s detrimental effect on OHCA internationally we hypothesise that COVID-19 would be associated with significant reductions in bystander CPR, bystander defibrillation and survival to hospital discharge in Ireland.

Study aims

The study will have four aims.

Study design

Key trends in available OHCAR data from the period 2012 – 2020 will be described and evaluated using descriptive summaries and graphical displays. Multivariable logistic regression will be used to model predictors of ‘bystander CPR’, ‘bystander defibrillation’ and ‘survival to hospital discharge’ and to explore the effects (if any) of system level developments in 2015/2016 and the COVID-19 pandemic on these outcomes. The R software for statistical computing will be used for analysing the data.

Study population, setting and time frame

The population for this study will be patients of all ages who suffered un-witnessed, or bystander witnessed OHCA during the time period 2012 – 2020 and are included in the Irish national OHCAR cardiac arrest registry database. Patients who had an EMS witnessed OHCA will be specifically excluded from this current study. Follow on work will consider this distinct excluded group separately.

Variables and categories of concern

All analysis will be based on the variables shown in Table 1. Variables 1–19 will be obtained from the OHCAR. Variables 20, 21 and 22 represent component time periods that are a priori considered to be potentially significant in the context of either system level developments or key population health challenges (Covid-19) during the study time period. Variables 20,21 and 22 will be dummy variables created using the ‘Year’ variable. Variables 20 and 21 will be linked dummy variables derived from the same 3-level categorical indicator of period and will be simultaneously included in regression models. We do not expect collinearity of these predictors, as transition and post-transition period correspond to years 2015 and 2016, and 2017 through 2020, respectively. The Covid period dummy variable (variable 22) will be an indicator of 2020 only, and thus should allow us to estimate any shift in survival in the final year of the study data. Table 1 highlights the independent variables that will be explored for each outcome of interest. Where an original OHCAR variable has multiple potential associated categories, we will collapse these categories to avoid decreased statistical power from analysis of an excessive number of potentially sparse categories. Original and collapsed categories are shown in Table 1. Season will represent the ‘winter’ healthcare period (October to March) as compared to other as ‘winter’ is a period of increased demand for the Irish healthcare service. The three time categories chosen are adopted based on previous OHCA research⁴⁴. Home (versus other) location was chosen as home represents the most common location for OHCA in Ireland. ‘Year’ (variable 9) will be treated as a continuous variable to conserve degrees of freedom and statistical power.

Statistical model building and interpretation procedures

For each outcome we will build a logistic model: initially a full model with all relevant predictors will be fitted. A refined model will then be built using a stepwise model selection procedure (STEPAIC function in R). This procedure builds several models from all possible combinations of the predictors by sequentially adding and dropping predictors and finally selects the model with the lowest AIC. The stepwise model will then be further improved by examining addition of pairwise interaction variables and retaining any interactions which improve fit. We will evaluate each model (full, stepwise and with interactions) based on AIC. In addition, the model deviance and a Hosmer-Lemeshow Goodness of Fit (GOF) test will be inspected for each model. To assess linearity between continuous predictors and the log odds of the outcome we will rely on good model fit as indicated by the Hosmer-Lemeshow goodness of fit test. We will also use logistic regression with higher order terms (polynomial terms) to explore if the relationship between the outcome and the continuous predictors is indeed non-linear. If higher-order terms are significant, these will be retained in the model, otherwise they will be dropped from the model. We will also attempt to ameliorate fit by including pair-wise interaction terms between predictors. We will summarise the effect of each individual variable in the final model using odds ratios and 95% confidence intervals. The effect of any significant interactions in the final model will be explored graphically. After selecting the final model from the three models, we will evaluate the predictive ability of the final best fitting model using 10-fold cross-validation and evaluate the prediction accuracy for the model.

Missing data and sensitivity analysis

We anticipate some missing data in both the outcome and predictor variables. We will document the amount of missing data for each variable and graph missing data patterns across the entire dataset. To conduct sensitivity analysis, results from complete case analysis and multiple imputation will be presented. Multiple imputation will be done using the mice package in R, ten imputations will be derived, and the selected model will then be fit to these datasets, and the results will be compared with the complete case dataset.