Methods for analyzing longitudinal data from randomized pretest-posttest-follow-up trials in behavioral research: a...

Area of Science:

• Behavioral Science
• Clinical Psychology
• Biostatistics

Background:

• Randomized pretest-posttest-follow-up (RPPF) designs are standard for evaluating longitudinal behavioral interventions.
• Assessing treatment efficacy and sustained effects over time is crucial in these designs.
• Traditional analysis methods may not fully capture the nuances of change in RPPF trials.

Purpose of the Study:

• To introduce Latent Change Models (LCMs) as a practical analytical approach for RPPF trials.
• To demonstrate the utility of LCMs using data from the STAR trial for pediatric epilepsy adherence.
• To compare LCMs with traditional methods like ANCOVA and mixed-effects models.

Main Methods:

• Application of Latent Change Models (LCMs) to analyze RPPF data.
• Utilized the STAR (Supporting Treatment Adherence Regimens) trial data, a pediatric behavioral intervention study.
• Contrasted LCM results with ANCOVA, longitudinal linear mixed-effects models, and latent growth curve models.

Main Results:

• LCMs effectively estimate discrete changes between time points and intervention-control group differences.
• The analysis demonstrated LCMs' ability to control for baseline variability and integrate all longitudinal data.
• LCMs provided a more accurate and nuanced understanding of intervention effects compared to other methods.

Conclusions:

• Latent Change Models (LCMs) are a powerful tool for analyzing RPPF trials in behavioral intervention research.
• LCMs offer strengths in estimating specific changes over time and handling complex longitudinal data.
• These models enhance the understanding of intervention efficacy and its temporal dynamics.