The statistician on our research team has explained that missing data is best handled using a “multiple imputation” approach. Why wasn’t multiple imputation, or other modern methods for addressing missing data, part of your answer to the previous question about a missing tailoring variable?


Multiple imputation is used by researchers to address missing data in a research study (e.g., a SMART). However, an adaptive intervention is not a research study. The solutions used to address missing data in an adaptive intervention–such as missing data on a tailoring variable–are not the same as the solutions to missing data in a research study.

The apparent contradiction here is because two different kinds of missing data are being mixed-up unintentionally. The first kind is missing research data. The second kind is missing data on a tailoring variable which is missing clinical data. Your team’s statistician is correct: Often, multiple imputation is a great choice for researchers who are dealing with missing research data, such as missing the primary outcome variable in a randomized trial. However, a tailoring variable is not a research assessment; rather, a tailoring variable is one of the components in an adaptive intervention. Therefore, the missingness that was the topic of the previous FAQ had to do with missing clinical data, not missing research data.

More Frequently Asked Questions: Adaptive Interventions

Our team develops guidelines, principles and methodologies for optimizing new interventions that can be used in practice to address the unique and changing needs of individuals over time.


Join the d3center Mailing List

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