Predictors, Outcomes, and Statistical Solutions of Missing Cases in Web-Based Psychotherapy: Methodological...

Area of Science:

• Psychology
• Clinical Research
• Data Science

Background:

• Missing cases in web-based psychotherapy trials pose challenges for evaluating treatment effectiveness.
• Understanding the characteristics and outcomes of participants lost to follow-up is crucial.

Purpose of the Study:

• To explore characteristics of missing cases in internet-delivered cognitive behavioral therapy (iCBT).
• To identify predictors of dropping out and their impact on clinical outcomes.
• To assess statistical models for approximating missing posttreatment data.

Main Methods:

• Analyzed a large sample (n=6701) of iCBT participants with a high rate of missing cases (36.26%).
• Identified predictors of missingness and moderators of clinical outcomes (distress, anxiety, depression).
• Compared various statistical models for imputing missing data using sensitivity and cross-validation.

Main Results:

• Treatment adherence and higher pretreatment symptom scores predicted missingness.
• Low adherence significantly increased odds of being a missing case and attenuated symptom reduction.
• Models ignoring adherence and baseline severity underestimated outcomes by up to 39%.

Conclusions:

• Treatment outcomes for missing cases differ significantly from observed cases.
• Failure to account for missing case characteristics leads to inaccurate treatment effect estimates.