A flexible micro-randomized trial design and sample size considerations

Area of Science:

• Digital Health
• Biostatistics
• Behavioral Science

Background:

• Mobile health interventions leverage technology for personalized support.
• Just-in-time adaptive interventions (JITAIs) deliver timely support based on individual needs.
• Micro-randomized trials (MRTs) are used to evaluate JITAI components.

Purpose of the Study:

• To propose a flexible micro-randomized trial (fMRT) design accommodating the addition of intervention components during a study.
• To address limitations of existing MRT frameworks that require fixed components at study onset.
• To facilitate the development and evaluation of adaptive mobile health interventions.

Main Methods:

• Developed a flexible micro-randomized trial design allowing for dynamic component addition.
• Introduced a novel test statistic and sample size calculator for fMRTs.
• Utilized a generalized estimating equation approach for longitudinal data analysis.
• Conducted simulation studies to validate the sample size calculator.

Main Results:

• The proposed fMRT design offers enhanced flexibility for adaptive intervention studies.
• The new test statistic and sample size calculator provide accurate estimations for fMRTs.
• Simulation studies confirmed the reliability of the sample size calculations.
• An R shiny application was developed for practical use of the calculator.

Conclusions:

• The flexible micro-randomized trial design enhances the adaptability of mobile health intervention research.
• This framework supports the rigorous evaluation of dynamic JITAIs.
• The developed statistical tools and application aid researchers in designing and powering future adaptive intervention studies.