Background:Sleep problems are common in bipolar disorders (BDs). To objectively characterize these problems in BDs, further methodological development is needed to capture subjective insomnia.
Aim:To test psychometric properties of the Athens Insomnia Scale (AIS), and associations with actigraphy‐derived measures, applying modifications in actigraphy data processing to capture features of perturbed sleep in patients with a BD.
Methods:Seventy‐four patients completed the AIS and the Quick Inventory of Depressive Symptomatology, self‐report (QIDS‐SR‐16). Locomotor activity was continuously recorded by wrist actigraphy for ≥10 consecutive days. We computed the sleep onset/offset, the center of daily inactivity (CenDI), as a proxy for chronotype, and the degree of consolidation of daily inactivity (ConDI), as a proxy for sleep‐wake rhythm strength.
Results:AIS showed good psychometric properties (Cronbach's alpha = 0.84; test–retest correlation = 0.84, P<.001). Subjective sleep problems correlated moderately with a later sleep phase (CenDI with AIS rho = 0.34, P = .003), lower consolidation (ConDI with AIS rho = −0.22, P = .05; with QIDS‐SR‐16 rho = −0.27, P = .019), later timing of sleep offset (with AIS rho = 0.49, P = ≤.001, with QIDS‐SR‐16 rho = 0.36, P = .002), and longer total sleep (with AIS rho = 0.29, P = .012, with QIDS‐SR‐16 rho = 0.41, P = ≤.001). While AIS was psychometrically more solid, correlations with objective sleep were more consistent across time for QIDS‐SR‐16.
Conclusions:AIS and QIDS‐SR‐16 are suitable for clinical screening of sleep problems among patients with a BD. Subjective insomnia associated with objective measures. For clinical and research purposes, actigraphy and data visualization on inactograms are useful for accurate longitudinal characterization of sleep patterns.