Towards Reducing Missing Information in Ecological Momentary Assessment of the Use of Non-Combustible Nicotine Products
Towards Reducing Missing Information in Ecological Momentary Assessment of the Use of Non-Combustible Nicotine Products
Just-in-time adaptive interventions (JITAIs) offer promising tools to help young adults reduce the use of non-combustible nicotine products. However, missing data from ecological momentary assessments (EMAs), essential for evaluating these interventions, threaten data validity and bias the estimation of intervention effects. This project aims to address systematic nonadherence in EMAs and enhance data collection methodologies.
This pilot research addresses critical methodological challenges in JITAIs, paving the way for more accurate and representative assessments that enhance the effectiveness of nicotine reduction interventions.
Young adults increasingly use non-combustible nicotine products such as vaping devices and nicotine pouches. JITAIs, informed by microrandomized trials (MRTs), rely on frequent ecological momentary assessments (EMAs) to evaluate real-time intervention impacts. However, EMAs often lead to participant nonadherence, producing missing data that could systematically bias intervention results and limit generalizability.
This pilot project has two primary aims:
Aim 1: Conduct a systematic review to identify patterns of nonadherence at the moment level, exploring how situational factors (e.g., time, location, mental state) affect EMA response rates. The review will pinpoint factors consistently linked to adherence, guiding future strategies to mitigate missingness and enhance the representativeness of collected data.
Aim 2: Analyze existing data to examine if combining evening recall surveys with EMAs improves data completeness and reduces biases. Evening recalls, being less intrusive, allow retrospective reporting, potentially balancing data representation and reducing systematic nonadherence.
The synergistic findings from these aims will inform future interventions by strategically integrating real-time and retrospective measures to ensure comprehensive and unbiased data collection.

Principal Investigator
Shiyu Zhang, PhD
Research Investigator
Institute for Social Research
University of Michigan
Key Collaborators
Lara Coughlin, University of Michigan
Frederick Conrad, University of Michigan
Lindsey Potter, University of Utah
Mashfiqui Rabbi, Optum AI, United Health Group
Funding Source
d3center Pilot Grant Program