Aims: To estimate the prevalence and correlates of Pittsburgh Sleep Quality Index (PSQI)-defined poor sleep in geriatric rehabilitation inpatients, and to compare 72-hour actigraphy between PSQI groups.
Methods: Prospective observational cohort study in a metropolitan geriatric rehabilitation unit (November 2024 to November 2025), embedded within a larger program assessing the ward environment (light/ sound), and co-designing and implementing a multicomponent non-pharmacological sleep intervention. Participants aged 60-95 years completed the PSQI, Insomnia Severity Index (ISI), Depression-Anxiety-Stress Scales-21 (DASS-21), Short Form-36 (SF-36), and wore a wrist ActiGraph for 72 hours. Poor sleep was defined PSQI >5. Logistic regression produced odds ratios (ORs) with 95% confidence intervals (CIs). Actigraphy comparisons were conducted in the valid recording subset (n=45).
Results: Mean global PSQI was 9.87. Poor sleep occurred in 52/62 participants (83.9%), yet documented prior sleep disorder diagnoses were uncommon (13/62, 21%). Odds of poor sleep increased with ISI (OR 1.32 per point, 95% CI 1.09–1.69; p=0.0018) and DASS-21 Anxiety (OR 1.17 per point, 95% CI 1.01–1.45; p=0.0398), and higher SF-36 scores were associated with lower odds (OR 0.92 per point, 95% CI 0.84–0.99; p=0.023). In the actigraphy subset (n=45), total sleep time was 88.4 minutes lower in poor sleepers than good sleepers (95% CI -169.2 to -7.6; p=0.033); sleep onset latency, wake after sleep onset, and sleep efficiency did not differ significantly.
Conclusions: Self-reported sleep disturbance was highly prevalent yet under-recognised in this geriatric rehabilitation cohort. Findings support routine sleep assessment, proactive medication review, and tailored non-pharmacological strategies, with selective use of actigraphy.