Citation

Jacobucci R, Ammerman BA, McClure K. J. Clin. Psychol. (Hoboken) 2024; ePub(ePub): ePub.

Copyright

(Copyright © 2024, John Wiley and Sons)

DOI

10.1002/jclp.23728

PMID

38943339

Abstract

The use of intensive time sampling methods, such as ecological momentary assessment (EMA), has increased in clinical, and specifically suicide, research during the past decade. While EMA can capture dynamic intraindividual processes, repeated assessments increase participant burden, potentially resulting in low compliance. This study aimed to shed light on study-level and psychological variables, including suicidal ideation (SI), that may predict momentary prompt (i.e., prompt-to-prompt) completion. We combined data from three EMA studies examining mental health difficulties (N = 103; 10,656 prompts; 7144 completed), using multilevel models and machine learning to determine how well we can predict prompt-to-prompt completion and which variables are most important. The two most important variables in prompt-to-prompt completion were hours since the last prompt and time in study. Psychological variables added little predictive validity; similarly, trait-level SI demonstrated a small effect on prompt-to-prompt completion. Our study showed how study-level characteristics can be used to explain prompt-to-prompt compliance rates in EMA research, highlighting the potential for developing adaptive assessment schedules to improve compliance.

Language: en

Keywords

assessment; machine learning; ecological momentary assessment; intensive longitudinal