The development of a risk-adjustment strategy to benchmark emergency medical service (EMS) performance in relation to out-of-hospital cardiac arrest in Australia and New Zealand

• Published in: Resuscitation 2023-07, Vol.188, p.109847-109847, Article 109847
• Main Authors: Howell, Stuart, Smith, Karen, Finn, Judith, Cameron, Peter, Ball, Stephen, Bosley, Emma, Doan, Tan, Dicker, Bridget, Faddy, Steven, Nehme, Ziad, Swain, Andy, Thorrowgood, Melanie, Thomas, Andrew, Perillo, Samuel, McDermott, Mike, Smith, Tony, Bray, Janet
• Format: Article
• Language: English
• Subjects: Adult; Australia - epidemiology; Benchmarking; Cardiopulmonary Resuscitation; Cohort Studies; Emergency Medical Services; Heart arrest; Humans; New Zealand - epidemiology; Out of hospital cardiac arrest; Out-of-Hospital Cardiac Arrest - therapy; Registries; Resuscitation; Risk Adjustment
• ISSN: 0300-9572; 1873-1570
• DOI: 10.1016/j.resuscitation.2023.109847

The aim of this study was to develop a risk adjustment strategy, including effect modifiers, for benchmarking emergency medical service (EMS) performance for out-of-hospital cardiac arrest (OHCA) in Australia and New Zealand. Using 2017–2019 data from the Australasian Resuscitation Outcomes Consortium (Aus-ROC) OHCA Epistry, we included adults who received an EMS attempted resuscitation for a presumed medical OHCA. Logistic regression was applied to develop risk adjustment models for event survival (return of spontaneous circulation at hospital handover) and survival to hospital discharge/30 days. We examined potential effect modifiers, and assessed model discrimination and validity. Both OHCA survival outcome models included EMS agency and the Utstein variables (age, sex, location of arrest, witnessed arrest, initial rhythm, bystander cardiopulmonary resuscitation, defibrillation prior to EMS arrival, and EMS response time). The model for event survival had good discrimination according to the concordance statistic (0.77) and explained 28% of the variation in survival. The corresponding figures for survival to hospital discharge/30 days were 0.87 and 49%. The addition of effect modifiers did little to improve the performance of either model. The development of risk adjustment models with good discrimination is an important step in benchmarking EMS performance for OHCA. The Utstein variables are important in risk-adjustment, but only explain a small proportion of the variation in survival. Further research is required to understand what factors contribute to the variation in survival between EMS.