| Abstract | Cognitive Behavioural Therapy for insomnia (CBT-I) is an effective treatment, significantly reducing symptom severity for both primary and comorbid insomnia for years following treatment discontinuation. However, access to therapists, time commitment and journaling, as well as symptom monitoring and adherence remain barriers to this treatment. To address these limitations, web-based and mobile technologies are increasingly being used to provide health-related assessments or interventions; previous research has shown that digital versions of CBT-I (dCBT-I) are an adequate alternative to the traditional approach. Sensors could potentially eliminate tedious and manual sleep data collection required by dCBT-I, and offer objective and additional data pertaining to sleep (e.g. sleep fragmentation) which could be used to monitor and personalise the intervention, ultimately impacting adherence and outcome. This work compares sleep metrics collected using sensors to those obtained using the standard sleep journal, in order to evaluate their potential to inform a dCBT-I. Twenty older adults (65+) with insomnia and five without insomnia were recruited for the study. For 14 days, sleep data was collected using a journal and various sensors. Total Sleep Time (TST), Wake-After-Sleep Onset (WASO), Sleep Efficiency (SE), and Time-In-Bed (TiB) were compared between the journal and two of the sensors: the Withings Sleep Mat and Fitbit Charge 5, using Pearson’s r and Spearman’s rank correlations. Statistically significant positive correlations were found between the sensors and journal for TST and TiB (rTST=0.33, rTiB=0.24, p>0.0001). A strong, positive, statistically significant correlation was found between the two sensors for TST (r=0.89, p>0.0001). For WASO, no significant correlation was found between the sensors and the journal, but a positive correlation was found between the two sensors (r=0.38, p>0.0001). These results suggest that sensors could provide consistent sleep metrics that may be used to inform a dCBT-I. A pilot study is underway comparing intervention outcomes between journal-based and sensor-based dCBT-I. |
|---|
