aiqbench analysis
SafeImpute introduces a novel framework for clinical data imputation that combines graph neural networks with conformal prediction to ensure statistical reliability in high-stakes medical applications. By controlling the false discovery rate of unacceptable errors, this method addresses critical gaps in existing models where average accuracy often masks dangerous unreliability in sparse longitudinal records. The approach represents significant progress in applying rigorous machine learning safety guarantees to real-world healthcare datasets like MIMIC-IV.
SafeImpute: Reliable Clinical Data Imputation via Conformal Selection
arXiv:2607.05613v2 Announce Type: replace Abstract: Clinical care often relies on key laboratory indicators, yet real-world patient visits are sparse and tests are ordered irregularly, leading to pervasive missingness. While many imputation methods improve average accuracy, they provide limited guidance on which imputed values are reliable enough for high-stakes downstream use. In this work, we study reliable clinical imputation, aiming to produce accurate imputations while selectively releasing the reliable results, with statistical control over clinically unacceptable errors. To achieve this goal, we propose SafeImpute, a reliable imputation framework for irregular and sparse clinical longitudinal records. SafeImpute constructs an event graph that captures both intra-patient temporal trajectories and inter-patient clinical similarity, and learns imputations with a two-relation GNN and adaptive fusion, regularized by an auxiliary masked reconstruction objective. For reliability guarantees, SafeImpute converts a proxy risk score into conformal p-values and applies the Benjamini--Hochberg procedure to control the false discovery rate (FDR) of unacceptable errors among released imputations at a user-specified tolerance. Experiments on our Mayo Clinic data, the public MIMIC-III and MIMIC-IV datasets show that SafeImpute achieves strong imputation accuracy while providing reliable error control, outperforming diverse baselines in both standard imputation evaluation and FDR-controlled selective-release evaluation.
Abstract:Clinical care often relies on key laboratory indicators, yet real-world patient visits are sparse and tests are ordered irregularly, leading to pervasive missingness. While many imputation methods improve average accuracy, they provide limited guidance on which imputed values are reliable enough for high-stakes downstream use. In this work, we study reliable clinical imputation, aiming to produce accurate imputations while selectively releasing the reliable results, with statistical control over clinically unacceptable errors. To achieve this goal, we propose SafeImpute, a reliable imputation framework for irregular and sparse clinical longitudinal records. SafeImpute constructs an event graph that captures both intra-patient temporal trajectories and inter-patient clinical similarity, and learns imputations with a two-relation GNN and adaptive fusion, regularized by an auxiliary masked reconstruction objective. For reliability guarantees, SafeImpute converts a proxy risk score into conformal p-values and applies the Benjamini--Hochberg procedure to control the false discovery rate (FDR) of unacceptable errors among released imputations at a user-specified tolerance. Experiments on our Mayo Clinic data, the public MIMIC-III and MIMIC-IV datasets show that SafeImpute achieves strong imputation accuracy while providing reliable error control, outperforming diverse baselines in both standard imputation evaluation and FDR-controlled selective-release evaluation.
Accepted at KDD 2026. Author accepted manuscript
Machine Learning (cs.LG)
arXiv:2607.05613 [cs.LG]
(or arXiv:2607.05613v2 [cs.LG] for this version)
https://doi.org/10.48550/arXiv.2607.05613 Focus to learn more arXiv-issued DOI via DataCite
https://doi.org/10.1145/3770855.3817967 Focus to learn more DOI(s) linking to related resources
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.
Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.
Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.
Full text under the publisher's open licence · 300 words
Related coverage
A Hybrid Predictive Ensemble of Machine Learning and Deep Neural Networks for Early Cardiovascular Disease Risk Assessment
arXiv · Sep 7, 2026
Picking the Right Image to Classify: Reliable-Input Selection in Teledermatology
arXiv · Aug 18, 2026
CRS-Triage: Confidence- and Reliability-Aware Selective Triage under Incomplete Clinical Evidence
arXiv · Aug 5, 2026
CRC-Router: Risk-Constrained Routing for Medical Agentic AI Systems
arXiv · Sep 28, 2026