Software annotation of defibrillator files: Ready for prime time?

Ventilation rates measured by capnography during out-of-hospital cardiac arrest resuscitations and their association with return of spontaneous circulation

BACKGROUND

Clinical guidelines for adult out-of-hospital cardiac arrest (OHCA) recommend a ventilation rate of 8-10 per minute yet acknowledge that few data exist to guide recommendations. The goal of this study was to evaluate the utility of continuous capnography to measure ventilation rates and the association with return of spontaneous circulation (ROSC).

METHODS

This was a retrospective observational cohort study. We included all OHCA during a two-year period and excluded traumatic and pediatric patients. Ventilations were recorded using non-invasive continuous capnography. Blinded medically trained team members manually annotated all ventilations. Four techniques were used to analyze ventilation rate. The primary outcome was sustained prehospital ROSC. Secondary outcomes were vital status at the end of prehospital care and survival to hospital admission. Univariable and multivariable logistic regression models were constructed.

RESULTS

A total of 790 OHCA were analyzed. Only 386 (49%) had useable capnography data. After applying inclusion and exclusion criteria, the final study cohort was 314 patients. The median ventilation rate per minute was 7 (IQR 5.4-8.5). Only 70 (22%) received a guideline-compliant ventilation rate of 8-10 per minute. Sixty-two (20%) achieved the primary outcome. No statistically significant associations were observed between any of the ventilation parameters and patient outcomes in both univariable and multivariable logistic regression models.

CONCLUSIONS

We failed to detect an association between intra-arrest ventilation rates measured by continuous capnography and proximal patient outcomes after OHCA. Capnography has poor reliability as a measure of ventilation rate. Achieving guideline-compliant ventilation rates remains challenging.

A sliding-window based algorithm to determine the presence of chest compressions from acceleration data

This publication presents in detail five exemplary cases and the algorithm used in the article (Orlob et al. 2022). Defibrillator records for the five exemplary cases were obtained from the German Resuscitation Registry. They consist of accelerometry, electrocardiogram and capnography time series as well as defibrillation times, energies and impedance when recorded. For these cases, experienced physicians annotated time points of cardiac arrest and return of spontaneous circulation or termination of resuscitation attempts, as well as the beginning and ending of every single chest compression period in consensus, as described in Orlob et al. (2022). Furthermore, an algorithm was developed which reliably detects chest compression periods automatically without the time-consuming process of manual annotation. This algorithm allows for an usage in automatic resuscitation quality assessment, machine learning approaches, and handling of big amounts of data (Orlob et al. 2022).

Chest compression fraction calculation: A new, automated, robust method to identify periods of chest compressions from defibrillator data - Tested in Zoll X Series

AIM

To introduce and evaluate a new, open-source algorithm to detect chest compression periods automatically by the rhythmic, high amplitude signals from an accelerometer, without processing single chest compression events, and to consecutively calculate the chest compression fraction (CCF).

METHODS

A consecutive sample of defibrillator records from the German Resuscitation Registry was obtained and manually annotated in consensus as ground truth. Chest compression periods were determined by different automatic approaches, including the new algorithm. The diagnostic performance of these approaches was assessed. Further, using the different approaches in conjunction with different granularities of manual annotation, several CCF versions were calculated and compared by intraclass correlation coefficient (ICC).

RESULTS

131 defibrillator recordings with a total duration of 5755 minutes were analysed. The new algorithm had a sensitivity of 99.39 (95% CI 99.38, 99.41)% and specificity of 99.17 (95% CI 99.15; 99.18)% to detect chest compressions at any given timepoint. The ICC compared to ground truth was 0.998 for the new algorithm and 0.999 for manual annotation, while the ICC of the proposed algorithm compared to the proprietary software was 0.978. The time required for manual annotation to calculate CCF was reduced by 70.48 (22.55, [94.35, 14.45])%.

CONCLUSION

The proposed algorithm reliably detects chest compressions in defibrillator recordings. It can markedly reduce the workload for manual annotation, which may facilitate uniform reporting of measured quality of cardiopulmonary resuscitation. The algorithm is made freely available and may be used in big data analysis and machine learning approaches.

Methodology and framework for the analysis of cardiopulmonary resuscitation quality in large and heterogeneous cardiac arrest datasets

BACKGROUND

Out-of-hospital cardiac arrest (OHCA) data debriefing and clinical research often require the retrospective analysis of large datasets containing defibrillator files from different vendors and clinical annotations by the emergency medical services.

AIM

To introduce and evaluate a methodology to automatically extract cardiopulmonary resuscitation (CPR) quality data in a uniform and systematic way from OHCA datasets from multiple heterogeneous sources.

METHODS

A dataset of 2236 OHCA cases from multiple defibrillator models and manufacturers was analyzed. Chest compressions were automatically identified using the thoracic impedance and compression depth signals. Device event time-stamps and clinical annotations were used to set the start and end of the analysis interval, and to identify periods with spontaneous circulation. A manual audit of the automatic annotations was conducted and used as gold standard. Chest compression fraction (CCF), rate (CCR) and interruption ratio were computed as CPR quality variables. The unsigned error between the automated procedure and the gold standard was calculated.

RESULTS

Full-episode median errors below 2% in CCF, 1 min^-1 in CCR, and 1.5% in interruption ratio, were measured for all signals and devices. The proportion of cases with large errors (>10% in CCF and interruption ratio, and >10 min^-1 in CCR) was below 10%. Errors were lower for shorter sub-intervals of interest, like the airway insertion interval.

CONCLUSIONS

An automated methodology was validated to accurately compute CPR metrics in large and heterogeneous OHCA datasets. Automated processing of defibrillator files and the associated clinical annotations enables the aggregation and analysis of CPR data from multiple sources.