Using passive smartphone sensing to assess circadian rhythms in people at-risk of and with severe mental illness: a systematic review

Circadian rhythm disruptions are common in patients at-risk of or with manifest severe mental illness, like affective disorders or schizophrenia spectrum disorders and have been closely linked with clinical outcomes. Consequently, identifying and monitoring circadian rhythm may help detect prodromal symptoms in at-risk cohorts or monitor disease trajectory or treatment response in patients with manifest disorders. The development of smartphone-based passive sensing technologies has enabled unobtrusive, longitudinal data collection offering a granular temporal resolution and the potential of scalability to large cohorts. Recent years have seen an increase in studies using smartphone-based passive sensing data to extract circadian rhythm parameters, yet they show substantial methodological heterogeneity and consensus is lacking. A synthesized framework to guide the translation of these high-resolution temporal data into actionable clinical insights is missing. Therefore, this review aims to provide an overview of the current literature on passively sensed smartphone-based circadian rhythm metrics in people at-risk for developing or with manifest severe mental illness and to present their associations with validated clinical outcomes. This systematic review has been pre-registered with OSF. The search terms will follow a PECO-framework (population = people at risk-for developing or with manifest affective or psychotic disorders; exposure = passive sensing to measure circadian rhythm via smartphone; comparison = comparison of different passive sensing approaches; outcome = circadian rhythm parameters measured by passive sensing, and their association with clinical outcomes). The databases MEDLINE, PsycINFO, IEEE, Xplore, ACM DL, Embase and Web of Science will be searched, as they offer different foci in order to cover the topic as comprehensively as possible. In short, the inclusion criteria will focus on studies assessing circadian rhythmicity with smartphone-based technology in populations at-risk for developing or with manifest affective or psychotic disorders. All study designs will be considered. Studies including various populations (e.g. studies that include healthy participants and participants with psychiatric disorders) will be included if the results are reported separately for the population of interest. We will exclude studies, that focus only on EMA; actimetry or wearables and do not include passive smartphone data. To assess the quality of the included sources, we will adapt the Newcastle-Ottawa Scale (NOS) for observational studies to evaluate selection, comparability, and outcome. Additionally, we will apply a customized checklist for digital phenotyping, focusing on the report of technical details, data missingness and method validation. If possible, meta-analytic synthesis will be performed. However, given the expected heterogeneity of the studies, the results of the qualitative assessment will likely be reported using a narrative approach. This review will provide an up-to-date overview on smartphone based passive sensing as a potentially promising and scalable approach for assessing circadian rhythm in people with affective or psychotic disorders. This work will highlight, which digital circadian rhythm markers might be promising to be linked to clinical outcomes and where validation is still lacking. Therefore, our review will support the development of more standardized and comparable study designs and possible future applications in psychiatric research and care.

References

No references