WWhen Data Is Missing — That's the Finding

In clinical research, missing data is rarely random. When key predictors are absent in more than 45% of cases, the gap itself carries meaning — and simply dropping variables or running sensitivity analyses isn't enough.

At the Survival Analysis for Junior Researchers Conference (SAfJR 2026) (external link, opens in a new window), held 25–27 March 2026, we presented "Precision Joint Modeling under Missingness: A Unified Bayesian Approach for Predicted Individual Treatment Effects (PITEs)" — showing how treating missingness as information, rather than noise, unlocks precise predictions with remarkable computational gains.