Current wrist-actigraphy parameters poorly identify the cancer patients, who subjectively report poor sleep. However, most studies average multiple nights of data, not accounting for the changes that can occur while on treatment. Hence, our aim was to understand, on a night-by-night basis, the objective measures which best correlated with subjective sleep complaints. We investigated associations between sleep duration, timing, efficiency, latency and fragmentation, objectively measured by wrist-actigraphy, and the severity of self-rated sleep disturbance reported on the following day using a 0–10 scale questionnaire (M.D. Anderson Symptom Inventory) completed on a touch-screen computer at home. We used a dataset of 595 nights in 31 patients with advanced cancer participating in a pilot domomedicine study (inCASA project). First, we evaluated associations between actigraphy parameters and subjective sleep rating with Spearman’s correlations. Then, we categorised the objective sleep data into two groups depending on whether sleep disturbance was rated lower or higher than 2, which was previously found as corresponding to a level of intensity comfortable for the patient. Actigraphy parameters were compared in both groups using Mann-Whitney U-test. Multivariate regressions and classification and regression tree analysis were performed to identify independent predictors of subjective sleep. Sleep complaint severity was negatively correlated with timing of awakening (r=-0.22, p<0.001), total sleep time (r=-0.19, p<0.001) and sleep efficiency (r=-0.19, p<0.001). These associations were confirmed (p<0.001), with a median 40-min advance in timing of awakening, a median total sleep time shorter by 36 minutes, and a median sleep efficiency less by 2% for the nights rated as associated to sleep disruption as compared to comfortable sleep. The night-by-night approach within a multidimensional home tele-monitoring framework identified early awakening, short total sleep time and low sleep efficiency as significant objective correlates of poor sleep complaint in cancer patients on chemotherapy. These findings, if confirmed, will inform the development of both effective interventions and real time assessment in this frail patient population. European Commission through the ICT Policy Support Programme project inCASA (Contract CIP 250505, FP7), and the Coordinating Action Systems Medicine (CASyM) through research exchange grants in Systems Medicine.
Read full abstract