SafeImpute framework improves reliability of clinical data imputation

Researchers have developed SafeImpute, a new framework designed to improve the reliability of imputing missing clinical data. This method uses a graph neural network to capture patient trajectories and similarities, learning imputations with an adaptive fusion technique. SafeImpute also incorporates conformal p-values and the Benjamini-Hochberg procedure to control the rate of unacceptable errors among the imputed values, ensuring statistical reliability for high-stakes clinical decisions. Experiments on datasets from Mayo Clinic, MIMIC-III, and MIMIC-IV demonstrate that SafeImpute offers strong imputation accuracy while effectively controlling errors. AI

IMPACT Enhances the reliability of AI-driven imputation in critical healthcare applications, enabling more trustworthy downstream clinical decision-making.

RANK_REASON This is a research paper detailing a new method for data imputation. [lever_c_demoted from research: ic=1 ai=1.0]

Read on arXiv cs.LG →

• Benjamini--Hochberg procedure
• graph neural network
• Mayo Clinic
• MIMIC-III
• MIMIC-IV
• SafeImpute

• paper
• model release
• safety

AI-generated summary · Google Gemini · from 1 sources. How we write summaries →