Identifying change-dropout patterns during an Internet-based intervention for depression by applying the Muthen-Roy...

Area of Science:

• Psychology
• Digital Health
• Psychiatry

Background:

• Non-ignorable dropout in Internet-based interventions is understudied.
• Internet-based interventions are increasingly used for mental health conditions.

Purpose of the Study:

• To investigate change and dropout patterns in patients with mild-to-moderate depression undergoing an Internet-based intervention.
• To apply a Not Missing At Random (NMAR) model to understand dropout mechanisms.

Main Methods:

• Utilized the Muthen-Roy model to analyze data from 483 patients in a 12-week Internet-based intervention (deprexis).
• Assessed depression symptoms using the PHQ-9 biweekly.
• Identified four distinct change-dropout patterns.

Main Results:

• Four patterns emerged: high impairment/improvement/low dropout (C3, N=134, highest reliable change); high impairment/deterioration/high dropout (C2, N=32); low impairment/improvement/high dropout (C1, N=198); high impairment/no change/low dropout (C4, N=119).
• Rapid improvement, not just deterioration, can lead to dropout due to perceived treatment completion.

Conclusions:

• Dropout patterns are complex and influenced by both symptom deterioration and rapid improvement.
• Understanding these patterns is crucial for optimizing treatment selection and improving adherence in digital mental health interventions.