Elsevier

Journal of Affective Disorders

Volume 191, February 2016, Pages 88-93
Journal of Affective Disorders

Research paper
Daily mood monitoring of symptoms using smartphones in bipolar disorder: A pilot study assessing the feasibility of ecological momentary assessment

https://doi.org/10.1016/j.jad.2015.11.013Get rights and content

Highlights

  • Ecological momentary assessment is feasible to capture mood changes in bipolar disorder.

  • Subjects with bipolar disorder had lower mean mood and energy than healthy controls.

  • BD had significantly higher variability of symptoms than healthy controls.

Abstract

Background

Personal device technology has facilitated gathering data in real-time using Ecological momentary assessment (EMA). We hypothesized that using smartphones to measure symptoms in auto-generated surveys twice a day would be feasible in a group with bipolar disorder (BD). A second exploratory objective of this study was to compare potential differences in core symptoms between BD and healthy control (HC) groups.

Methods

A two-arm, parallel group, observational study was designed to measure completion rates of surveys of symptoms of mood, energy, speed of thought, impulsivity, and social stress in BD (N=10) and HC (N=10) participants. The surveys were auto-generated twice a day for fourteen days, and subjects could also perform self-generated surveys. Completion rates were compared between BD and HC groups. Scores were averaged for each participant over the 14 day period, and group medians were compared.

Results

Median completion rates did not differ between groups: 95% in BD, 88% in HC (p=0.68); the median completion rate of auto-generated surveys in the BD group was 79% and in the HC group was 71% (p=0.22). The BD group had significantly lower median mood score (p=0.043) and energy score (p=0.007) than the HC group. Median scores of speed of thoughts (p=0.739), impulsivity (p=0.123) and social stress (p=0.056) did not significantly differ between BD and HC. The BD group had significantly higher range of variability of group median mood (p=0.043), speed of thoughts (p=0.002) and impulsivity (p=0.005) scores over the course of 14 days than HC, while range of variability of energy (p=0.218) and social stress (p=0.123) scores did not differ. Results were not significantly different between auto-generated and self-generated surveys for BD or HC.

Limitations

This pilot study was conducted for a short time and with a small sample.

Conclusions

This study demonstrates feasibility of using EMA with a smartphone to gather data on BD symptoms.

Introduction

Bipolar disorder (BD) is a recurrent, episodic illness affecting 1–4.4% of the population (Merikangas et al., 2007). While BD is defined by episodes of depression, mania and/or hypomania, studies have shown that as many as half of patients have chronic, subsyndromal symptoms (Judd et al., 2002). Core symptoms of BD, such as disturbances in mood state and energy level are solicited by the healthcare provider through the use of clinical questioning, and – in systems using best-practice standards – patient-rated questionnaires and mood diaries between weekly or monthly appointments (Harding et al., 2011, Sachs et al., 2003). This current model is generally successful in detecting large fluctuations in mood, yet is limited by recall bias and the influence of manic or depressive states on recollection of symptoms (Ghaemi, 2007).

There is growing interest in self-monitoring of mood and behavior on a daily basis, allowing for observation of progression into episodic depressive or manic states (Bauer et al., 2007, Bauer et al., 2008, Aan Het Rot et al., 2012). The capacity for frequent self-monitoring of symptomology associated with BD, including prodromal symptoms, would allow clinicians to tailor treatment approaches in a more timely manner, and to better study the effects of early interventions on daily symptoms (Proudfoot et al., 2014, Judd et al., 2002). Additionally, monitoring of medication changes and mood changes related to medication are important for treatment (Bauer et al., 2013b, Bauer et al., 2013a).

Ecological momentary assessment (EMA) is a data-capturing method related to experience sampling that is designed to assess data as events happen in real-time (Csikszentmihalyi and Larson, 1987, Johnstone et al., 1989, Terracciano et al., 2003, Ebner-Priemer and Trull, 2009, Trull and Ebner-Priemer, 2009, Shiffman et al., 2008). Emotions and behaviors can be recorded in the moment they occur within the patient's natural environment (Aan Het Rot et al., 2012, Moskowitz and Young, 2006). EMA has been used in paper-and-pencil forms to explore emotional reactivity to daily life stress among participants with psychosis and affective disorder (Havermans et al., 2010, Havermans et al., 2011, Havermans et al., 2007, Myin-Germeys et al., 2003), and to gather mood data to improve prediction of suicidal ideation in subjects with BD (Thompson et al., 2014). Several studies in BD have also investigated negative or positive affect on a daily basis (Knowles et al., 2007, Myin-Germeys et al., 2003, Husky et al., 2010). Response to negative events have been identified as more stressful for BD than HC (Havermans et al., 2007), and correlated with higher cortisol levels (Havermans et al., 2011).

Advancing technology, including internet services and smartphones, has enhanced the ability to gather data in the field and has been well-accepted. Using at-home computer-based software, Bauer et al. (2004) demonstrated a greater than 80% completion rate for more than 90% of participants over a 3-month period; reliability between self-report and clinician-report (Bauer et al., 2004); and that the use of the computer does not bias data collection based on comparable demographics (Bauer et al., 2005). As part of a trial of a new intervention, (Miklowitz et al., 2012) developed a text or email-based system to report scores of self-rated scales to the study team, and subjects responded to 71% of the daily prompts and 88% of the weekly prompts. (Faurhopsen et al., 2013) developed a system for detecting prodromal symptoms and reported participation using this system in a randomized controlled trial of 66.1%. (Depp et al., 2012) found higher compliance using paper-and-pencil data collection when compared to using mobile phones in 56 participants with BD, however, mobile phones were better able to capture intra-participant variability. Concordance between self- and clinician-ratings has been found to be greater for depressive symptoms than manic symptoms (Faurholt-Jepsen et al., 2014, Depp et al., 2012).

We hypothesized that using smartphones to capture EMA ratings of mood symptoms would be feasible in those with BD, as measured by completion rates. To measure symptomology of BD, we selected a visual analog scale for mood, energy, speed of thoughts and impulsivity. These four areas correspond to independent factors of mania shown through factor analysis: dysphoric mood (mood), psychomotor pressure (energy), psychosis (not conducive to self-report measure), increased hedonic function (impulsivity) and irritable aggression (Cassidy et al., 1998a, Cassidy et al., 1998b). Additionally, mood, energy and intellect (thoughts) are the three areas noted through careful, longitudinal study of phenomenology by Emil Kraepelin to cycle out of sync, producing mixed states (Kraepelin, 1921, Mackinnon and Pies, 2006). It was hypothesized that the BD group would demonstrate lower mood and energy; and higher speed of thoughts, impulsivity and social stress. We hypothesized that greater variability in all symptoms would be present in the BD group.

Section snippets

Design and sample

The study was approved by the institutional review board of the Hershey Medical Center (PSU COM IRB # 00251, approval 3/28/2014). A two-arm, parallel group, observational study was designed to measure a method of capturing information on symptoms of mood, energy, speed of thought, impulsivity and stress in a BD and a HC group. Personal health information collected for this study was stored in Research Electronic Data Capture System (REDCap), a secure data-management system supported by the Penn

Completion rates

Demographic characteristics of the control and BD groups are presented in Table 1. Groups did not differ by age, sex or employment rate. Eighteen out of twenty participants completed the full 14 days of the study; one BD participant dropped out on day 10 due to development of a manic episode, and one HC participant dropped out on day 12 due to a family emergency. Descriptive data on the completion of surveys are presented in Table 1. Median completion rates of surveys were not significantly

Discussion

Completion rates for this study demonstrated that the use of EMA on a smartphone device is a feasible modality for data collection in BD. The majority of the surveys were completed through the auto-generated survey, reflecting true EMA. While technological methods for capturing data are generally well-accepted, some have shown less compliance with smartphones than pencil-and-paper (Depp et al., 2012). However, capitalizing on technological methods quells concerns for factors that may generate

Contributors

S. Schultz and S. Schwartz were involved with the design of the trial, the collection of data, analysis of data and preparation of all drafts of the manuscript. AR was involved with study design, data collection, and reviewing the manuscript. ES was involved with study design, data collection, data analysis and preparation of all drafts of the manuscript.

Institutional review board

The study was approved by the institutional review board of the Hershey Medical Center (PSU COM IRB # 00251, approval 3/28/2014).

Conflict of interest

None.

Financial disclosures

EFHS has been a consultant for Projects In Knowledge, CME; AR, SS, SS – none.

Acknowledgements and role of the funding source

The authors wish to acknowledge the participants in the study who donated time and effort to research. The project described was supported by the National Center for Research Resources, Grant KL2 RR033180 (EFHS), and is now at the National Center for Advancing Translational Sciences, Grant KL2 TR000126. The content is solely the responsibility of the authors and does not necessarily represent the official views of the NIH. The sponsors of this research did not have direct influence over the

References (37)

  • M. Bauer et al.

    Drug treatment patterns in bipolar disorder: analysis of long-term self-reported data

    Int. J. Bipolar Disord.

    (2013)
  • M. Bauer et al.

    Using technology to improve longitudinal studies: self-reporting with ChronoRecord in bipolar disorder

    Bipolar Disord.

    (2004)
  • M. Bauer et al.

    Does the use of an automated tool for self-reporting mood by patients with bipolar disorder bias the collected data?

    MedGenMed

    (2005)
  • R. Bauer et al.

    Antidepressant dosage taken by patients with bipolar disorder: factors associated with irregularity

    Int. J. Bipolar Disord.

    (2013)
  • F. Cassidy et al.

    A factor analysis of the signs and symptoms of mania

    Arch. Gen. Psychiatry

    (1998)
  • M. Csikszentmihalyi et al.

    Validity and reliability of the experience-sampling method

    J. Nerv. Ment. Dis.

    (1987)
  • C.A. Depp et al.

    A pilot study of mood ratings captured by mobile phone versus paper-and-pencil mood charts in bipolar disorder

    J. Dual Diagn.

    (2012)
  • C.A. Depp et al.

    Mobile interventions for severe mental illness: design and preliminary data from three approaches

    J. Nerv. Ment. Dis.

    (2010)
  • Cited by (0)

    1

    These authors contributed equally to this manuscript

    View full text