Objective and subjective sleep characteristics in hospitalized older adults and their associations to hospital outcomes

Abstract

Sleep in the hospital setting is understudied, with limited literature describing measurement of sleep quality. This study among older inpatients in an acute-care hospital describes sleep characteristics both objectively and subjectively, and explores the associations of sleep with hospital outcomes.Older patients (n = 112) at Sutter Tracy Community Hospital were enrolled from January 2016 to November 2017. Sleep prior to admission was measured subjectively [Pittsburgh Sleep Quality Index (PSQI)], while sleep during hospitalization was measured subjectively (sleep diaries) and objectively (actigraphy, averaged over all nights). Outcomes measured included change in cognition during the hospital stay (i.e., Montreal Cognitive Assessment), length of stay (LOS), discharge to a skilled nursing facility (SNF), and re-admittance to a hospital within 30 days of discharge.The participants were on average 68.7 ± 6.5 years old, predominately white (77%) and 55% women. Average PSQI was high (9.1 ± 4.2) indicating poor sleep quality prior to admission. Actigraphy was well-tolerated, with most (89%) having complete data. Sleep during the hospital stay was disturbed, with low levels of total sleep time (5.6 ± 2.0 h) and high levels of fragmentation (sleep efficiency 68.4 ± 15.0%). Sleep interruption was reported on 71% of sleep diaries, with the most common reasons being due to medical care [measurement of vitals (23%), staff interruptions (22%), blood draws (21%)]. Those with lower sleep efficiency had more cognitive decline upon discharge. Although underpowered, there was a suggestion of an association with poor sleep and the likelihood of being discharged to a SNF. Those with worse self-reported sleep quality (PSQI) prior to admission had a slightly longer LOS. No associations were seen with sleep quality and likelihood of readmission.Collection of objective and subjective sleep measures was feasible among hospitalized older adults. Disrupted sleep was common, and was potentially related to poor hospital outcomes. Our next steps will be to leverage these results to design and implement an intervention to improve sleep in hospitalized adults.