LLM-rephrased synthetic clinical notes preserve core information and utility for coarse prediction tasks but lose fine-grained details such as ICD codes, with chunk-wise rephrasing as a partial mitigation that trades off factual accuracy.
Aligning AI research with the needs of clinical coding workflows: Eight recommendations based on US data analysis and critical review
2 Pith papers cite this work. Polarity classification is still indexing.
2
Pith papers citing it
fields
cs.CL 2years
2026 2verdicts
UNVERDICTED 2representative citing papers
Adding ICD-9 data to ICD-10 training boosts micro F1 by 27% on 18K rare codes and improves macro metrics on frequent codes in a label-wise attention model.
citing papers explorer
-
Systematic Evaluation of the Quality of Synthetic Clinical Notes Rephrased by LLMs at Million-Note Scale
LLM-rephrased synthetic clinical notes preserve core information and utility for coarse prediction tasks but lose fine-grained details such as ICD codes, with chunk-wise rephrasing as a partial mitigation that trades off factual accuracy.
-
Bridging the Version Gap: Multi-version Training Improves ICD Code Prediction, Especially for Rare Codes
Adding ICD-9 data to ICD-10 training boosts micro F1 by 27% on 18K rare codes and improves macro metrics on frequent codes in a label-wise attention model.