0944 A Comparison Of Medical-grade Actigraphy Devices With Polysomnography During Inpatient Rehabilitation For Traumatic Brain Injury (TBI)

Introduction
Sleep disturbances in TBI population can have negative consequences on physical, psychological and cognitive recovery. While polysomnography (PSG) is the gold standard for evaluating sleep disorders, actigraphy has been widely used to screen for sleep-wake disorders, especially ones that change longitudinally. In a small sample of TBI patients (n=50, single center sample of convenience) during inpatient rehabilitation, we have previously found two models (Philips Actiwatch2/Actiwatch Spectrum) to be feasible and valid for collecting sleep data. Herein, we extend this validation to a prospective study of more than 200 TBI patients and include both the Actigraph Core wGT3X-BT (ACT-Core) and Philips Actiwatch Spectrum Plus (PASP).

Methods
This is a secondary analysis from a six-center, prospective, observational cohort study from the TBI Model System Lifetimes Study where overnight, fully attended polysomnography with concomitant actigraphy were conducted on moderate to severe TBI subjects during acute inpatient rehabilitation as part of an ongoing PCORI-funded clinical trial. Actigraphy variables of total sleep time (TST), sleep efficiency (SE), sleep latency (SL), wake after sleep onset (WASO) were compared between actigraphy devices and with Level 1, fully attended PSG (Phillips Sleepware-G3) to evaluate their concordance (Pearson). PASP-derived sleep was determined with Actiware (sensitivity=Automatic), while ACT-Core-derived sleep was determined with ActiLife (Cole-Kripke algorithm).

Results
Simultaneous collection of PSG and PASP (n=230) and ACT-Core (n=179) were examined. Correlations (p>.01) with PSG were found for both PASP and ACT-Core for TST (r=.70, .69), SE (r=.33, .31), and WASO (r=-.19, -.24) respectively. Comparison of actigraphy devices to one another revealed correlations (p>.01) for TST (r=.80), SE (r=.58), WASO (r=.68), and SL (r=.23).

Conclusion
This larger, prospective study across six-centers supports the use of actigraphy as a proxy measure of sleep quantity (TST, SE) in acute neurorehabilitation admissions with moderate to severe TBI. No relationship was found for SL with mixed findings for WASO. Sensitivity analyses with subsets of TBI survivors with comorbidities (sleep apnea, spasticity, muscle strength) may affect agreement and warrant further study.