Defining the Hospital Night: Deriving and Validating Electronic Health Record-Based Measures of the Inpatient Sleep Experience

Study Objectives To identify electronic health record (EHR)–derived sleep measures associated with the inpatient sleep experience during hospitalization. Methods We recruited 247 inpatients who completed the Karolinska Sleep Diary and the Potential Hospital Sleep Disruptions and Noise Questionnaire. Using EHR data, we derived three measures: longest uninterrupted sleep opportunity (LUSO), total uninterrupted sleep opportunity (TUSO), and patient-staff interactions (PSIs) calculated within two timeframes: a fixed window (23:00–06:00) and a patient-defined window (lights out to final awakening). We examined the associations between EHR-derived measures calculated within each window and sleep experience measures. Results Sleep duration correlated more strongly with patient-defined TUSO (r = 0.63, p < 0.001) than with fixed TUSO (r = 0.27, p < 0.001). Ratings of sleep continuity, refreshing sleep, and sound sleep were also associated with patient-defined TUSO (TUSO r = 0.15–0.23; all p < 0.05), supporting its construct validity as an indicator of sleep quality. Sleep disruption ratings were associated with EHR-derived measures of overnight activity, including PSI counts (r = 0.35 within fixed windows vs. r = 0.29 within patient-defined windows; both p < 0.001). Conclusions EHR-derived measures of sleep opportunity and interruption provide scalable tools for characterizing the inpatient sleep experience using routinely collected clinical data. Measures calculated within patient-defined windows show stronger associations with sleep duration and quality, while PSIs calculated within both window types show similar associations with perceived sleep disruption from overnight clinical workflows. These complementary approaches may inform strategies to reduce unnecessary nighttime disturbances and improve sleep in hospitalized patients.