Optimizing an existing prediction model for quality of life one-year post-intensive care unit: An exploratory analysis.

BACKGROUND: This study aimed to improve the PREPARE model, an existing linear regression prediction model for long-term quality of life (QoL) of intensive care unit (ICU) survivors by incorporating additional ICU data from patients' electronic health record (EHR) and bedside monitors. METHODS: The 1308 adult ICU patients, aged ≥16, admitted between July 2016 and January 2019 were included. Several regression-based machine learning models were fitted on a combination of patient-reported data and expert-selected EHR variables and bedside monitor data to predict change in QoL 1 year after ICU admission. Predictive performance was compared to a five-feature linear regression prediction model using only 24-hour data (R2  = 0.54, mean square error (MSE) = 0.031, mean absolute error (MAE) = 0.128). RESULTS: The 67.9% of the included ICU survivors was male and the median age was 65.0 [IQR: 57.0-71.0]. Median length of stay (LOS) was 1 day [IQR 1.0-2.0]. The incorporation of the additional data pertaining to the entire ICU stay did not improve the predictive performance of the original linear regression model. The best performing machine learning model used seven features (R2  = 0.52, MSE = 0.032, MAE = 0.125). Pre-ICU QoL, the presence of a cerebro vascular accident (CVA) upon admission and the highest temperature measured during the ICU stay were the most important contributors to predictive performance. Pre-ICU QoL's contribution to predictive performance far exceeded that of the other predictors. CONCLUSION: Pre-ICU QoL was by far the most important predictor for change in QoL 1 year after ICU admission. The incorporation of the numerous additional features pertaining to the entire ICU stay did not improve predictive performance although the patients' LOS was relatively short.

Electroencephalographic reactivity as predictor of neurological outcome in postanoxic coma: A multicenter prospective cohort study.

OBJECTIVE: Outcome prediction in patients after cardiac arrest (CA) is challenging. Electroencephalographic reactivity (EEG-R) might be a reliable predictor. We aimed to determine the prognostic value of EEG-R using a standardized assessment. METHODS: In a prospective cohort study, a strictly defined EEG-R assessment protocol was executed twice per day in adult patients after CA. EEG-R was classified as present or absent by 3 EEG readers, blinded to patient characteristics. Uncertain reactivity was classified as present. Primary outcome was best Cerebral Performance Category score (CPC) in 6 months after CA, dichotomized as good (CPC = 1-2) or poor (CPC = 3-5). EEG-R was considered reliable for predicting poor outcome if specificity was ≥95%. For good outcome prediction, a specificity of ≥80% was used. Added value of EEG-R was the increase in specificity when combined with EEG background, neurological examination, and somatosensory evoked potentials (SSEPs). RESULTS: Of 160 patients enrolled, 149 were available for analyses. Absence of EEG-R for poor outcome prediction had a specificity of 82% and a sensitivity of 73%. For good outcome prediction, specificity was 73% and sensitivity 82%. Specificity for poor outcome prediction increased from 98% to 99% when EEG-R was added to a multimodal model. For good outcome prediction, specificity increased from 70% to 89%. INTERPRETATION: EEG-R testing in itself is not sufficiently reliable for outcome prediction in patients after CA. For poor outcome prediction, it has no substantial added value to EEG background, neurological examination, and SSEPs. For prediction of good outcome, EEG-R seems to have added value. ANN NEUROL 2019.