EXPERIMENTAL DESIGNS

What is an MRT?

The Micro-Randomized Trial is an experimental design used to construct an optimized Just-in-Time Adaptive Intervention.

In an MRT, individuals are randomized hundreds or thousands of times over the course of the study. The data can be used to investigate the effects of just-in-time intervention options on proximal outcomes and whether these effects vary with time or by contextual conditions.

Why use an MRT?

When developing a Just-in-Time Adaptive Intervention, behavioral scientists need to optimize multiple components and the way those components work together. An MRT helps them make important decisions about the construction of a JITAI, including whether or not to include a time-varying component as part of an intervention package, and in what contexts certain intervention options are most effective. During the MRT design process, researchers establish desired outcomes and use existing knowledge to build a custom trial that can answer nuanced scientific questions about what intervention components to deliver, under what conditions, and in what order.

Why use an MRT?

When developing a Just-in-Time Adaptive Intervention, researchers need to optimize multiple components and the way those components work together. The MRT helps researchers make important decisions about the construction of a JITAI, including whether or not to include a time-varying component as part of an intervention package, and in which contexts which intervention component options are most effective. The MRT design process leads researchers to answer key questions, then the MRT answers deeper scientific questions that help them optimize their JITAI’s design.

What types of questions can an MRT answer?

What types of questions can an MRT answer?

Intervention engagement questions.

Does delivering a prompt recommending a self-regulatory activity (compared to no prompt) increase engagement in self-regulatory activities, on average across all individual states and circumstances?

Intervention targeting questions.

Under what conditions delivering a prompt is most beneficial? For example, delivering a prompt to engage in a self-regulatory activity (compared to no prompt) may be especially beneficial when the individual experiences stress or craving.

Intervention option design questions.

Which type of prompt is most beneficial? Prompts recommending a low-effort self-regulatory activity may be (on average) more beneficial in promoting engagement than those recommending a high-effort activity, but high-effort activities may be (on average) more beneficial in building resiliency.

Comparative effectiveness questions.

Under what conditions is one type of prompt better than another? For example, what time of day is best to deliver a prompt recommending a high-effort activity, compared to low-effort? Are smokers attempting to quit more likely to engage following a prompt recommending a low-effort activity (compared to high-effort) while at work?

An optimized JITAI…

Leverages both moments of opportunity and moments of vulnerability

Capitalizes explicitly on moment-to-moment, within-person heterogeneity

Adjusts intervention quickly in response to a person’s behavior and to their ever changing context

Actively balances effectiveness and efficiency

Makes accommodations for a heterogenous target population

Integrates digital and human-delivered interventions seamlessly

Constructing an optimized Just-in-Time Adaptive Intervention.

We think of the questions above as optimization questions because their answers can help researchers construct a high-quality Just-in-Time Adaptive Intervention that has the potential to improve health outcomes in the long-term for a greater number of individuals.

Importantly, MRTs are not confirmatory studies designed to evaluate an intervention, rather they are focused on selecting and optimizing Just-in-Time Adaptive Intervention (JITAI) intervention components to be delivered as part of an intervention package.

CASE STUDY

The Substance Abuse Research Assistant (SARA)

This MRT is designed to optimize the delivery of just-in-time prompts that capitalize on principles of behavioral economics to promote daily mobile-based self reporting. Each of the optimization questions below corresponds to a component of the MRT design schematic.

Optimization Questions

GOAL
To determine the effect of an engagement intervention on survey completion.

QUESTION
Does offering a youth-targeted inspirational message increase a person’s likelihood to complete a survey?

GOAL
To determine the effect of an engagement intervention on survey completion.

QUESTION
Does offering a reward message for completing a survey increase the likelihood that someone will complete a survey again the next day?

MRT Resources

Learning Modules

Software

View Projects Using MRTs

References

Boruvka, A., Almirall, D., Witkiewitz, K., & Murphy, S. A. (2018). Assessing time-varying causal effect moderation in mobile health. Journal of the American Statistical Association, 113(523), 1112-1121.

Carpenter, S. M., Menictas, M., Nahum-Shani, I., Wetter, D. W., & Murphy, S. A. (2020). Developments in
Mobile Health Just-in-Time Adaptive Interventions for Addiction Science. Current Addiction Reports, 1-11.

Dempsey, W., Liao, P., Kumar, S., & Murphy, S. A. (2020). The stratified micro-randomized trial design: sample size considerations for testing nested causal effects of time-varying treatments. Annals of Applied Statistics, 14(2), 661-684.

Liao, P., Klasnja, P., Tewari, A., & Murphy, S. A. (2016). Sample size calculations for micro‐randomized trials in mHealth. Statistics in medicine, 35(12), 1944-1971.

LET’S STAY IN TOUCH

Join the d3center Mailing List

Keep up to date with the latest news, events, software releases, learning modules, and resources from the d3center.