Abstract Body: Introduction
High quality of post-cardiac arrest care has been linked to improved outcomes, but data is limited on compliance with current best practices. As part of the ICECAP (Influence of Cooling duration on Efficacy in Cardiac Arrest Patients) trial on targeted temperature management, regular physiologic assessments are collected for each patient, to include hourly measurements of deviations from protocolized pulse-oximetry (SpO2) goals.
Hypothesis
We hypothesize that there is a wide distribution of deviations from protocolized care for measured physiologic parameters both between patients and across clinical sites that enroll patients in the ICECAP trial.
Aims
To quantify and compare the incidence and duration of SpO2 transgressions across ICECAP sites and identify patterns of variability within and between clinical centers.
Methods
Clinical trial case report forms for daily physiologic transgressions were reviewed for all patients included in the ICECAP trial through March 20 2025. Durations of deviations from protocolized care were calculated for each patient for hourly measurements of pulse-oximetry (<90% or >98%). The burden of transgression was expressed as the percentage of monitored time spent outside the target range and split by sex and race. We compared transgression burdens across sites, using Levene’s test for within-site variability and Kruskal-Wallis for between-site differences.
Results
1100 patients were included with a median age of 61 years (Q1 51, Q3 70), 421 (38%) female, 674 (61%) male. American Indian or alaska native - 4 (0.36%), Asian - 37 (3%), Black or African American - 421 (38%), Native hawaiian or Pacific Islander - 7 (0.64%), White - 546 (49.6%), multiple/unknown 85 (7.7%).
Hypoxia (SpO2 < 90%) was rare, with a median burden of 0% (mean [SD]: 4.8% [12.2]; Q1-Q3: 0%-4.1%). In contrast, presumed hyperoxia (SpO2 > 98%) was common, with a median reported burden of hyperoxia of 66.7% (mean [SD]: 61.6% [29.4]; Q1-Q3: 39.3%-88%). Transgression burdens differed significantly between sites (p <0.0001), but were consistent within sites (SpO2 < 90%, p = 0.64; SpO2 > 98%, p = 0.26), suggesting variation in institutional practices.
Conclusions
While there may be a reporting bias, adherence to strict oxygen levels is not consistently done, in particular in regards to presumed hyperoxia. A substantial number of patients are spending over half of their reported monitored time above the recommended range.