0418 IMPROVING ACTIGRAPHY-BASED SLEEP EFFICIENCY ESTIMATES

Abstract

Introduction: Sleep efficiency (SE) is the guiding index of cognitive behavioral therapy for insomnia (CBT-I), but it is typically estimated from patient self-reports of questionable validity. Actigraphy (ACT) should enhance CBT-I by providing objective estimates of sleep efficiency; however, only two studies have validated actigraphy- based estimates of sleep efficiency (ACT-SE) in sleep-disordered patients studied at home. Both found ACT-SE to correspond poorly with PSG-based SE (PSG-SE). The current study assessed the validity of ACT-SE in a third sleep-disordered sample studied at home and piloted a simple method of improving ACT-based estimates of sleep efficiency. Methods: Participants with panic disorder, posttraumatic stress disorder, or comorbid posttraumatic stress and panic disorder, and controls without sleep complaints, underwent in-home recording of sleep using concurrent ambulatory PSG and actigraphy. Synchronized PSG and proportional integral mode (PIM) ACT recordings were obtained from 41 participants. Sleep efficiency was scored using conventional methods from both ACT and PSG, and ACT-SE/PSG-SE concordance examined. ACT data were then resampled to 30-second epochs and rescaled on a per-participant basis to yield optimized concordance between PSG- and ACT-based sleep efficiency estimates. Results: The correlation between ACT-SE and PSG-SE across participants was statistically significant (r = 0.35, p < 0.025); though ACT-SE failed to replicate a main effect of diagnosis. Individualized calibration of ACT against a night of PSG yielded a significantly higher correlation between ACT-SE and PSG-SE (r = 0.65, p < 0.001; z = 1.692, p = 0.0452, one-tailed) and a significant main effect of diagnosis that was highly correspondent with the effect on PSG-SE. Conclusion: Actigraphic estimation of sleep efficiency in sleep-disordered patients tested at home can be significantly improved by calibration against a single night of concurrent PSG. This improvement must rely on amplification/attenuation of non-zero epochs of ACT. Modern actigraphs provide the temporal and amplitude resolution needed for re-calibration.