Modeling and Predicting Real World Behavior Using Mobile Sensor Data on Patients With Major Depressive Disorder

Ginger.io

Status: Completed

Conditions: Depressive Disorder

Treatments: Other: Behavioral-data driven support

Study type: Interventional

Funder types: Industry

Identifiers

NCT02499094

GIO-001

Details and patient eligibility

About

The purpose of this study is to validate the effectiveness of using an integrated mobile sensing platform to deliver large-scale data-driven interventions to patients with depression.

Full description

This study is a smartphone-based, randomized, single-blind, controlled parallel-

design study with two intervention arms and one control arm. The two intervention arms

will receive in-app messages and phone-based support, which will be triggered by

participant's self-reported surveys and passive behavioral data gathered through a

smartphone app. The study will include a nationwide sample of adult (18 years or older)

smartphone users, who are currently experiencing depressive symptoms. The primary

outcome will be decrease in depression symptom severity, as measured by the 9-item

Patient Health Questionnaire, over 6 months.

Enrollment: 1,004 patients

Sex: All

Ages: 18 to 65 years old

Volunteers: No Healthy Volunteers

Inclusion criteria

• Currently suffering from depression (as measured by a PHQ-9 score of 10 or more at the time of screening)
• Own an iPhone or Android smartphone with a mobile voice calling plan with a US carrier
• Fluency in English

Exclusion criteria

• Participants with visual or hearing impairment
• Recent history of pregnancy (currently pregnant or those who have given birth within the past four months at the time of screening)
• Recent loss of a loved one (within the past two months at the time of screening)
• Unable or unwilling to accept End User License Agreement, or to provide information regarding their demographic characteristics and mental health history

Trial design

Primary purpose: Supportive Care

Allocation: Randomized

Interventional model: Parallel Assignment

Masking: Double Blind

---

1,004 participants in 3 patient groups

Intervention A- Heuristic based

Experimental group

Description:

Behavioral-data driven support, both mobile phone application and phone-based, informed by their self-reported symptom assessment as well as simple heuristic-based behavioral measures

Treatment:

Other: Behavioral-data driven support

Intervention B- Machine Learning Based

Experimental group

Description:

Behavioral-data driven support, both mobile phone application and phone-based, informed by their self-reported symptom assessment as well as machine learning model-based behavioral measures

Treatment:

Other: Behavioral-data driven support

Control

No Intervention group

Description:

No intervention