Comparing aggregated and nonlinear measures of coordinative behaviors in immediate responder first aid groups: a controlled experimental simulation study

Background When accidents and disasters happen, emergent ad-hoc groups of immediate responders form to provide aid before the arrival of professional medical responders. Little is known about how immediate responders coordinate with each other over time and how that coordination affects the quality of the first aid that is provided. Emergency response is described as complex and chaotic, but methods for analyzing complexity have rarely been applied in the emergency response context. This exploratory study aims to compare recurrence quantification analysis and aggregated data analysis of coordination behaviors by immediate responders to predict performance in groups providing first aid in a simulated context. Methods Sixteen groups of three participants ( N = 48, 58% male, 40% female, and 2% preferred not to answer) were recruited from a large university and the surrounding city in southern Sweden using convenience sampling. They took part in a simulated traffic accident scenario. Of the participants, 56% had prior first aid training, including 6 emergency response professionals who were individually placed in different groups. The first aid performance of each group was rated by three subject matter experts. Only the first 10 min and 28s of each scenario was included in the analysis to allow for direct comparison of the recurrence quantification analysis results between groups. The Co-ACT framework was used to code participants’ coordinative behaviors, and the dynamics of the behaviors’ structure were calculated using a multidimensional recurrence quantification analysis. The aggregated and dynamical structures of the coordinative behaviors (i.e. behaviors of team members coordinating teamwork) were evaluated for their predictive power with regard to first aid performance ratings. Results Two groups were excluded due to incomplete data, and the analyses were conducted on the remaining 14 groups. Rate of recurrence of coordinative behaviors was found to significantly predict first aid performance ( p = .02, β = -0.57 ) while aggregated post-experimental measures of coordinative behaviors did not (Action implicit p = .37, β = -0.31 , Action explicit p = .97, β = <0.01 , Information implicit p = .99, β = 0.01 , Information explicit p = .68, β = -0.14 ). The negative effect of recurrence rate of coordinative behaviors on first aid performance suggests that other behaviors not measured in the Co-ACT analysis (e.g., task work such as chest compressions or bleeding control) had more effect on the quality of the first aid that was provided, at least in the current scenario. Conclusions The study shows the usefulness of controlled scenarios for studying complexity in provision of first aid. Additionally, the results provide initial evidence for the predictive power of recurrence rate on first aid performance in ad-hoc first aid groups. This exploratory study thus shows initial evidence of the usefulness of complexity methods to understand the emergency response context. These contributions have potential significance for both emergency response researchers and practitioners.