Prediction on the Recurrence of Manic and Depressive Episodes in Bipolar Disorder

Status

Recruiting

Phases

Unknown

Study type

Observational

Source

ClinicalTrials.gov

Registry ID

NCT05828056

Enrollment

100

Registered

2023-04-25

Start date

2020-03-02

Completion date

2027-12-31

Last updated

2026-07-31

For informational purposes only — not medical advice. Sourced from public registries and may not reflect the latest updates. Terms

Conditions

Bipolar Disorder, Depressive Disorder

Keywords

biomarker, Bipolar disorder

Brief summary

Mood disorders (including bipolar disorder and major depressive disorder) are chronic mental disorders with high recurrent rate. The more the number of recurrence is, the worse long-term prognosis is. This study aims to establish a prediction model of recurrence of manic and depressive episodes in mood disorders, with a hope to detect recurrence relapse as early as possible for timely clinical intervention. We will adopt wearable smart watch to collect heart rate, sleep pattern, activity level, as well as emotional status for one year long in 100 patients with bipolar disorder, and annotated their mood status (i.e., manic episode, depressive episode, and euthymic state). We expect to establish prediction models to predict the recurrence of mood episodes.

Interventions

DEVICEWearable activity tracker

Garmin smartwatch will record features, such as activities, heart rate, sleep, through smartphone App

Study design

Observational model

CASE_ONLY

Time perspective

PROSPECTIVE

Eligibility

Sex/Gender

ALL

Age

20 Years to 60 Years

Healthy volunteers

Inclusion criteria

* DSM-5 Bipolar disorder or depressive disorder * 20\~60 years old * Willing to carry smartwatch and smartphone most of the time

Exclusion criteria

* Comorbid with substance use disorder * Unable to use smartwatch and smartphone

Design outcomes

Primary

Measure	Time frame	Description
Development and verification of mood episode prediction algorithm	1 year	Collected data will apply to learning algorithm, random forest, which constructs a multitude of decision trees at training time and outputting a class that is the mode of the classes of the individual trees. Performance of the trained prediction model was evaluated by assessing the model's accuracy, sensitivity, specificity, and the area under the curve. In a machine learning evaluation process, a part of data is used for model training, and the other portion is used for model testing.

Countries

Taiwan

Contacts

CONTACTYi-Ling Chien

chienyiling@gmail.com+886223123456#66013

Outcome results

None listed

Source: ClinicalTrials.gov · Data processed: Aug 1, 2026