JMIR Publications

ABSTRACT

Background:

Advancements in mobile health (mHealth) technologies and machine learning approaches have expanded the framework of behavioral phenotypes in obesity treatment to explore the dynamics of temporal changes.

Objective:

This study aimed to investigate the dynamics of behavioral changes during obesity intervention and to identify behavioral phenotypes associated with weight change using a hybrid machine learning approach.

Methods:

In total, 88 children and adolescents (ages 8-16 years, 69.8% males) with age- and sex-specific body mass index (BMI) ≥ 85th percentile participated in the study. Behavioral phenotypes were identified using a hybrid two-stage procedure based on the temporal dynamics of adherence to the five behavioral goals during the intervention. Functional principal component analysis (FPCA) was utilized to determine behavioral phenotypes by extracting principal component factors from the functional data of each participant. Elastic net regression was employed to investigate the association between behavioral phenotypes and weight change.

Results:

The FPCA identified two distinctive behavioral phenotypes, which were named the “early behavioral change phenotype” and “delayed behavioral change phenotype.” The first phenotype explained 47-69% of each factor, whereas the second phenotype explained 11-17% of the total behavioral dynamics. The “early behavioral change phenotype” was associated with weight change for adherence to screen time (β = -0.0766, 95% CI [-0.1245, -0.0312]), fruit and vegetable intake (β = 0.1770, 95% CI [0.0642, 0.2561]), exercise (β = -0.0711, 95% CI [-0.0892, -0.0363]), drinking water (β = -0.0203, 95% CI [-0.0218, -0.0123]), and sleep duration. The “delayed behavior change phenotype” was significantly associated with weight loss for changes in screen time (β = 0.0440, 95% CI [0.0186, 0.0550]), fruit and vegetable intake (β = -0.1177, 95% CI [-0.1441, -0.0680]), and sleep duration (β = -0.0991, 95% CI [-0.1254, -0.0597]).

Conclusions:

Earlier compliance to behavioral goals, the “early behavioral change phenotype,” and a gradual improvement in health-related behaviors, or the “delayed behavioral change phenotype,” were differently associated with weight loss for distinctive obesity-related lifestyle behaviors. A large part of health-related behaviors remained stable throughout the intervention, which indicates that healthcare professionals should closely monitor changes made during the early stages of the intervention. Clinical Trial: cris.nih.go.kr (KCT002718)