Analysis of Electronic Medical Record-Based Sleep Metric and Outcomes in Critically Ill Patients

Introduction: Sleep disruption is pervasive in the ICU and associated with adverse outcomes. The relationship between sleep opportunity and outcomes is unclear. The purpose of this study was to evaluate the association between a novel electronic medical record-derived sleep opportunity (SLOP) metric, delirium and subjective sleep quality in ICU patients.


Methods: This was a single-center, cohort study at an academic medical center (6/2019-6/2020). Adult patients admitted to the ICU ≥ 24 hours were eligible for inclusion. Exclusion criteria included primary neurologic diagnosis, dementia, or active substance withdrawal. SLOP was developed to quantify the longest uninterrupted time (hours) available for sleep between documented intrusive care events (2200-0600). ICU sleep deficit (8 - SLOP) was calculated for each patient night. Confusion-Assessment Method for the ICU was used to assess delirium. Subjective sleep quality was measured via modified Richards-Campbell Sleep Questionnaire (RCSQ). A Mixed Linear Model Regression was used to examine the relationship between RCSQ and sleep deficit. A Generalized Linear Mixed Model examined the impact of cumulative sleep deficit on delirium, adjusting for covariates, non-normal data, and accounting for repeated measures.


Results: A total of 283 ICU patients were included. Average SLOP per night was 1.35 +/- 0.39 hrs. There was a statically, but not clinically significant, association between daily sleep deficit and total RCSQ [0.003, 95% CI 0.000-0.006, p = 0.039] and mean RCSQ [0.020, 95% CI 0.001-0.038, p = 0.039]. In a subset of 228 patients without delirium in the first 24 hours of ICU stay, cumulative sleep deficit during admission was associated with increased odds of developing delirium [OR = 1.113, 95% CI 0.028-0.186, p = 0.008] in the first 4 days of ICU stay. Sleep deficit on the prior night was not independently associated with next-day ICU delirium [OR = 0.993, 95% CI -0.045-0.032, p = 0.734].


Conclusions: The EMR-based SLOP metric was not associated with next-day delirium or clinically significant changes in subjective sleep quality in ICU patients. Cumulative ICU sleep deficit was a significant predictor of delirium. Cumulative sleep deficit may be a modifiable risk factor and potential target for future research and sleep-promotion interventions.