Hostname: page-component-848d4c4894-pjpqr Total loading time: 0 Render date: 2024-06-17T07:52:23.316Z Has data issue: false hasContentIssue false
  • English
  • Français

Using smartphone battery data to infer sleep-wake metrics in psychiatric cohorts – an exploratory study

Published online by Cambridge University Press:  01 September 2022

S. Howes ,
G. Gillett ,
N. Palmius ,
A. Bilderbeck ,
G. Goodwin ,
K. Saunders  and
N. Mcgowan
S. Howes*
Affiliation:
John Radcliffe Hospital, Oxford University Clinical School, Oxford, United Kingdom
G. Gillett
Affiliation:
King’s College London, Institute Of Psychiatry, Psychology & Neuroscience, London, United Kingdom
N. Palmius
Affiliation:
University of Oxford, Institute Of Biomedical Engineering, Oxford, United Kingdom
A. Bilderbeck
Affiliation:
Manor House, P1vital Products, Wallingford, United Kingdom
G. Goodwin
Affiliation:
University of Oxford, Department Of Psychiatry, Oxford, United Kingdom
K. Saunders
Affiliation:
University of Oxford, Department Of Psychiatry, Oxford, United Kingdom
N. Mcgowan
Affiliation:
University of Oxford, Department Of Psychiatry, Oxford, United Kingdom
*
*Corresponding author.

Abstract

Core share and HTML view are not available for this content. However, as you have access to this content, a full PDF is available via the ‘Save PDF’ action button.
Introduction

Disturbances to sleep-wake patterns are associated with bipolar disorder (BD) and borderline personality disorder (BPD). Objective assessment typically involves actigraphy monitoring, although it may be possible to derive sleep-wake metrics from other digital data, such as smartphone battery degradation.

Objectives

To assess whether common actigraphy-derived phase markers of the sleep-wake pattern (L5 and M10 onset) are in agreement with measures derived from smartphone battery data and explore if battery metrics differ between people with BD, BPD , and a healthy control group (HC).

Methods

High frequency smartphone battery data was collected from 30 BD, 19 BPD and 33 HC participants enrolled in the Automated Monitoring of Symptom Severity (AMoSS) study, over 28 days. Participants also wore an actigraph during this period. L5 and M10 values were calculated separately based on the rate of smartphone battery degradation and conventional actigraphy methods. Bland-Altman analyses were performed to assess agreement between battery-derived and actigraphy-derived values, and Kruskal-Wallis tests used to compare diagnostic groups.

Results

For L5, battery-derived and actigraphy-derived values had a bias of 0.46 [-0.10, 1.02], upper limit of agreement (LOA): 5.45 [4.49, 6.41], and lower LOA: -4.53 [-3.56, -5.49]. For M10, the bias was 0 [-0.92, 0.92], upper LOA: 8.19 [6.61, 9.76], and lower LOA: -8.19 [-6.61, -9.76]. Between diagnostic groups, there was no difference for battery-derived M10 (p=0.652), or L5 (p=0.122).

Conclusions

Our results suggest battery-derived and actigraphy-derived M10 and L5 show good overall equivalence. However, battery-derived methods exhibit large variability, which limits the clinical utility of smartphone battery data to infer sleep-wake metrics.

Disclosure

No significant relationships.

Keywords

digital phenotyping
Type
Abstract
Information
European Psychiatry , Volume 65 , Special Issue S1: Abstracts of the 30th European Congress of Psychiatry , June 2022 , pp. S292 - S293
Creative Commons
Creative Common License - CCCreative Common License - BY
This is an Open Access article, distributed under the terms of the Creative Commons Attribution licence (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted re-use, distribution, and reproduction in any medium, provided the original work is properly cited.
Copyright
© The Author(s), 2022. Published by Cambridge University Press on behalf of the European Psychiatric Association
Submit a response

Comments

No Comments have been published for this article.