Evidence-appraisal glossary
Marginal Structural Model
A model, usually fit with inverse probability weighting, built to estimate the effect of treatments that change over time when time-varying confounding is present. It recovers the effect a sustained treatment strategy would have had.
Also called: MSM.
Marginal structural models are the best known of the g-methods for longitudinal data. By weighting each person by the inverse probability of their entire treatment history, they build a pseudo-population free of the feedback loop between treatment and evolving confounders. This lets analysts compare strategies such as always treat versus never treat without the bias that ordinary regression would introduce. They still require that all relevant confounders were measured.
Read the full Reading the Evidence blog.
This is a plain-language methodology definition for reading research. It is general education, not medical advice.