Tag
A study shows that anonymized patient records from decades past in AI training datasets can cause models to misdiagnose current patients by recalling historical health states, increasing the risk of errors in diagnoses.
A doctor argues that AI, despite advances in mathematics, has seen minimal implementation in medicine and clinical trials, missing opportunities to streamline processes and improve treatments.
OpenAI integrates ChatGPT Health with Epic's electronic health record system to let clinicians import patient data for AI-assisted queries, and adds a new Healthcare Public Data plugin for synthesizing medical information.
OpenAI introduces EHR integration and a Healthcare Public Data plugin for ChatGPT, enabling healthcare organizations to connect Epic patient records and access authoritative healthcare datasets.
This paper introduces an expert-guided neuro-symbolic pipeline combining LLMs for semantic normalization and fuzzy logic to assess compliance with sepsis treatment protocols, providing graded insights from clinical data.
CoMedBench is a reproducible benchmark evaluating synthetic medical data generators across 37 dataset-task pairs, showing that synthetic training data preserves most downstream signal on tabular tasks but temporal ICU tasks remain generator-sensitive.
This paper presents FederatedRSF, a Python package for federated random survival forests that handles partially overlapping medical data across institutions without sharing raw data, and demonstrates comparable performance to centralized training on breast cancer data.