arXiv · 2611.21939
A δ-adjusted Method for Tipping Point Sensitivity Analysis for a Binary Outcome with Missing Values
Abstract
Missing data remain a persistent challenge in clinical trials despite preventive efforts. Handling missing data in clinical trials requires not only a primary analysis but also sensitivity analyses, as the underlying missingness assumptions are difficult to verify. The δ-adjusted analysis, which assesses the robustness of primary results by adding a sensitivity parameter δ for one or both treatment groups, has been widely used for continuous outcomes. Then, a set of tipping points which flip the significance of the primary analysis can be identified. However, there is no corresponding δ-adjusted method for binary variables. In this article, we propose a δ-adjusted method for binary variables that addresses this gap. This method assesses robustness by adding a sensitivity parameter to the logit function of the estimated response probability under the primary imputation approach and shares the same property as the δ-adjusted method for continuous variables. We applied this method to a clinical trial evaluating the treatment effect of tirzepatide 15 mg versus 5 mg on a composite binary endpoint. This method provides a principled sensitivity analysis framework for binary endpoints with broad applicability in clinical trials.
Explore related subjects
Keep this discovery
Explore connections, maps & timelines
Mallikarjuna Rettiganti, Yipeng Wang, Yongming Qu. 2026-11-16. A δ-adjusted Method for Tipping Point Sensitivity Analysis for a Binary Outcome with Missing Values. https://searcharxiv.com/abs/2611.21939
Citations retain the original authors and DOI and include this site’s record URL. If this discovery service contributed to your research, you can also cite Searcharxiv by Indiciae: BibTeX · RIS.