Prediction of Intrahospital Cardiac Arrest Outcomes

Completed

• Conditions: Cardiac Arrest
• Interventions: Diagnostic Test: CPC

• Registration Number: NCT05466188
• Lead Sponsor: Kepler University Hospital

Brief Summary

Intrahospital cardiovascular arrest is one of the most common causes of death in hospitalized patients. In contrast to extramural cases of cardiovascular arrest, hospitalized patients often have severe medical conditions that can affect the outcome of resuscitation. Nevertheless, survival rates from resuscitation are better in hospitals than outside, because there is often a rapid start of resuscitation measures and predefined resuscitation standards. Regular CPR training and the availability of defibrillators in all bedside units can also positively influence outcome. Despite these many efforts, survival rates, especially of patients with good neurological outcome, remained stable at low levels even within hospitals in recent years and did not improve.

Most outcome parameters are nowadays well known. (e.g., initial rhythm, age, early defibrillation, etc.) Nevertheless, we still do not know today how relevant the corresponding factors actually are, especially in relation to each other. One approach to this might be machine learning methods such as "random forest", which might be able to create a predictive model. However, this has not been attempted to date.

The hypothesis of this work is to find out if it is possible to accurately predict the probability of surviving an in-hospital resuscitation using the machine learning method "random forest" and if particularly relevant outcome parameters can be identified.

Design: retrospective data analysis of all data sets recorded in the resuscitation register of Kepler University Hospital.

Measures and Procedure: Review of the registry for missing data as well as false alarms of the CPR team and, if necessary, exclusion of these data sets; evaluation of the data sets using the machine learning method random forest.

Detailed Description

Not available

Study Timeline

• Study Start: 2022-06-01
• Study First Submitted: 2022-07-12
• Study First Submitted QC: 2022-07-17
• Study First Posted: 2022-07-20
• Primary Completion: 2022-07-31
• Study Completion: 2022-07-31
• Status Verified: 2023-04
• Last Update Submitted: 2023-04-29
• Last Update Posted: 2023-05-03

Recruitment & Eligibility

• Status: COMPLETED
• Sex: All
• Target Recruitment: 668

Inclusion Criteria

• All adults patients suffering cardiac arrest and having been resuscitated by the medical emergency team of the Kepler University Hospital, Linz, Austria in the period of 2006-01-01 to 2018-10-31.

Exclusion Criteria

• None.

Study & Design

• Study Type: OBSERVATIONAL
• Study Design: Not specified

Arm && Interventions

Group	Intervention	Description
Outcome CPC Positive	CPC	Outcome CPC Positive
Outcome CPC Negative	CPC	Outcome CPC Negative

Primary Outcome Measures

Name	Time	Method
AUROC for Classification of Outcome CPC	2006-01-01 to 2018-12-31	AUROC for Classification of Outcome CPC

Secondary Outcome Measures

Name	Time	Method
Confusion Matrix	2006-01-01 to 2018-12-31	Confusion Matrix Results: true positives, true negatives, false positive, false negatives and values calculated from these results.
Descriptive Statistics	2006-01-01 to 2018-12-31	Descriptive Statistics (age in years, delay in seconds, gender as male/female, agonal breathing/initial rhythm/airway management/iv-access/witnessed cardiac arrest/use of AED/chest compressions as binary features); This outcome measure will compare the individual feature (e. g. height in cm) in one group vs. the other. Significant difference will be described by p-value.

Trial Locations

Locations (1)

Kepler University Hospital; 🇦🇹 Linz, Upper Austria, Austria