# Improving Utstein accuracy: concordance of bystander CPR reporting by paramedic documentation vs. telecommunicator audio review

## TL;DR

This study compares how accurately paramedics and dispatchers record whether bystanders performed CPR during cardiac arrests, finding significant discrepancies.

## Contribution

The study reveals moderate agreement and highlights the need for standardized reporting of bystander CPR in emergency systems.

## Key findings

- Audio review documented higher bystander CPR rates than patient care reports (74.8% vs. 57.0%).
- Moderate inter-rater agreement was found between the two documentation methods (κ = 0.402).
- Initial shockable rhythm and advanced life support units were associated with higher documentation rates.

## Abstract

Bystander cardiopulmonary resuscitation improves outcomes for out-of-hospital cardiac arrest. While registries typically rely on patient care reports for bystander CPR documentation, the accuracy of reporting is unknown. This study aimed to determine the agreement in bystander CPR reporting between patient care reports and audio review of public safety answering point emergency calls.

In this retrospective study, we analyzed paired patient care reports and dispatch audio files for bystander CPR documentation. Cases dispatched by a secondary public safety answering point, emergency medical services-witnessed arrests, or those with missing data were excluded. We compared documented rates of bystander CPR from each source, calculated inter-rater agreement, and identified factors associated with reporting.

The concurrence in reporting was 72.2 %, with a moderate inter-rater agreement between the two methods (κ = 0.402, 95 % CI 0.341–0.463). Audio review documented higher bystander CPR than patient care reports (74.8 % vs. 57.0 %). An initial shockable rhythm (adjusted odds ratio (aOR) 1.68, 95 % CI 1.18–2.40 in patient care reports; aOR 1.57, 95 % CI 1.04–2.36 in audio files) and advanced life support unit first on scene (aOR 1.55, 95 % CI 1.15–2.07 in patient care reports; aOR 1.51, 95 % CI 1.08–2.11 in audio files) were associated with higher documentation.

We found moderate agreement in bystander CPR documentation between audio files and patient care reports, with a higher incidence of bystander CPR recorded in dispatch audio. These findings suggest inconsistencies in bystander CPR documentation across the emergency response system, highlighting the need for standardized reporting to ensure accurate data collection.