A comparative study of engagement in mobile and wearable health monitoring for bipolar disorder

Abstract

Objectives: Self-monitoring is recommended for individuals with bipolar disorder, with numerous technological solutions available. This study aimed to identify basic components of these solutions that increase engagement with self-monitoring.

Methods: Participants with bipolar disorder (n = 47) monitored their symptoms with a Fitbit and a smartphone app and were randomly assigned to either review or not review recorded symptoms weekly. We tested whether individuals would better adhere to and prefer monitoring with passive monitoring with an activity tracker compared to active monitoring with a smartphone app and whether individuals would better adhere to self-monitoring if their recorded symptoms were reviewed with an interviewer.

Results: Monitoring with a smartphone app achieved similar adherence and preference to Fitbit (P > .85). Linear mixed effects modeling found adherence decreased significantly more over the study for the Fitbit (12% more, P < .001) even though more participants reported they would use the Fitbit over a year compared to the app (72.3% vs 46.8%). Reviewing symptoms weekly did not improve adherence, but most participants reported they would prefer to review symptoms with a clinician (74.5%) and on monthly basis (57.5%) compared to alternatives. Participants endorsed sleep as the most important symptom to monitor, forgetfulness as the largest barrier to self-monitoring, and raising self-awareness as the best reason for self-monitoring.

Conclusions: We recommend a combined strategy of wearable and mobile monitoring that includes reminders, targets raising self-awareness, and tracks sleep. A clinician may want to review symptoms on a monthly basis.

Trial registration: ClinicalTrials.gov NCT03358238.

Keywords: Bipolar disorder; Fitness Trackers; Monitoring; Physiologic; Smartphone.

© 2019 John Wiley & Sons A/S. Published by John Wiley & Sons Ltd.

Figures

Figure 1.

Comparison between active monitoring with a smartphone app and passive monitoring with a Fitbit. A) Mean (± SE) overall adherence rates across participants was measured for the app and the Fitbit.B) Mean (± SE) adherence by study day across participants was measured for the app and the Fitbit. C) End of the study preference across participants with 95% confidence intervals was measured between using an app over the Fitbit. D) Participants reported they would use the Fitbit over a year but not the app, and most participants found that the Fitbit and the app were not at all complicated and required little to no effort.

Figure 2.

Comparison between reviewing recorded symptoms weekly (Arm R) and not reviewing recorded symptoms weekly (Arm NR). A) Mean (± SE) overall adherence rates across participants were measured by study arm.B) Most individuals stated they would want to review their symptoms on a monthly basis and preferably with a clinician.

Figure 3.

Most participants stated they would continue to visualize their symptoms for over a year and thought visualization of symptoms was not complicated but required a fair amount or a lot of effort.

Figure 4.

Other insights into participant engagement. Top three A) barriers, B) symptoms, and C) uses for monitoring were collected. Forgetfulness was most commonly endorsed barrier, raising self-awareness was most commonly endorsed use, and sleep was the mostly commonly endorsed symptom. D) General attitudes were also collected, with most participants stating that they learn a fair amount or a lot from monitoring but did not feel their emotional health improved.