Heterogeneity Matters: Predicting Self-Esteem in Online Interventions Based on Ecological Momentary Assessment Data

Area of Science:

• Psychology
• Digital Health
• Statistical Modeling

Background:

• Self-esteem is vital for mental health, with low levels linked to depression and anxiety.
• Ecological Momentary Assessments (EMA) in internet-based interventions often collect ordinal self-esteem data.
• Analyzing this data requires specialized statistical approaches to account for individual differences.

Purpose of the Study:

• To demonstrate a method for analyzing EMA data to predict self-esteem in online depression treatment.
• To explore the relationship between mood, worries, sleep, activities, social contact, and self-esteem.
• To investigate the impact of patient heterogeneity on prediction performance.

Main Methods:

• Applied ordinal outcome models to predict self-esteem using diary data from 130 patients in an online depression treatment.
• Explored various ordinal models with differing heterogeneity levels, estimated using Bayesian statistics.
• Analyzed relationships between mood, worries, sleep, enjoyed activities, and social contact with self-esteem.

Main Results:

• Models incorporating greater patient heterogeneity demonstrated superior performance in predicting self-esteem.
• Higher mood levels and engagement in enjoyable activities were significantly associated with increased self-esteem.
• Sleep, social contact, and worries showed predictive value for self-esteem in a subset of individuals.

Conclusions:

• Accounting for patient-heterogeneity significantly enhances the prediction of self-esteem in digital mental health interventions.
• Individual-level analysis of psychological factors provides valuable insights for therapists and practitioners.
• This approach offers a pathway for more personalized and effective internet-based mental health care.