In-hospital non-arrest (IHNA) cardiopulmonary resuscitation (CPR) alerts, while consuming substantial resources, are poorly characterized. We quantified the burden of IHNA and identified clinical and operational factors that differentiate them from true in-hospital cardiac arrests (IHCA). A retrospective cohort of adult CPR-team alerts at a German tertiary center between January 2011 and February 2015 was conducted. CPR-positive (CPR + ) events were confirmed IHCA requiring chest compressions; others, i.e., IHNA, were CPR-negative (CPR — ). Demographic, clinical and workflow data were abstracted from hospital records. Factors associated with CPR — /IHNA were analyzed using multivariable logistic regression with non-parametric bootstrap validation. Among 288 CPR-team deployments, 154 (53.5%) were CPR — /IHNA and 134 (46.5%) CPR + /IHCA. Patients experiencing CPR — /IHNA were younger (51.5 ± 20.7 vs. 64.1 ± 13.9 years) and predominantly female (57.2% vs. 32.1%). CPR — /IHNA occurred more frequently during morning shifts and in non-clinical on-campus locations, whereas CPR + /IHCA clustered in monitored wards and among inpatients. In multivariable analysis, older age, inpatient status, witnessed events and location in Intermediate Care/Intensive Care Unit (IMC/ICU)/telemetry wards independently reduced the odds of CPR — /IHNA, while morning shift timing increased them; model discrimination was good (AUC 0.84). Over half of CPR-team alerts were CPR — /IHNA and were associated with younger age, unwitnessed collapse, non-inpatient status and non-clinical locations. These findings delineate contexts in which non-arrest emergencies are more frequently encountered and may inform targeted training and early assessment strategies, while underscoring the need for cautious interpretation given the observational design. High-risk contexts for CPR — /IHNA, such as morning shifts, non-clinical campus areas and events involving non-hospitalized individuals, highlight settings in which enhanced staff awareness, early clinical assessment, and education on non-arrest emergencies may be particularly relevant. Potential effects on alarm burden or resource utilization require prospective evaluation.
Nashtar et al. (Thu,) studied this question.