# Call-Efficient ICD-10 Code Prediction via Supervised Candidate Generation and Selective LLM Invocation

> Source: <https://www.machinebrief.com/news/call-efficient-icd-10-code-prediction-via-supervised-candida-lyqf>
> Published: 2026-11-18 05:00:00+00:00

**Recommendation:** Route only uncertain diagnoses to the LLM for reranking instead of calling it on every case; the gated pipeline matched full-invocation accuracy while invoking the model for 31.97% of cases.
More in the playbook "Controlling LLM cost": <https://wpnews.pro/playbooks/llm-cost>

arXiv:2611.21891v1 Announce Type: new 
Abstract: Automatic assignment of ICD-10 codes from short clinical diagnosis text is a practical medical text classification and normalization task. The task is difficult because diagnosis phrases are short and noisy, the ICD-10 label space is hierarchical, and the class distribution is strongly imbalanced. This paper presents a call-efficient ICD-10 code prediction pipeline that combines transformer embeddings, tree-based supervised candidate generation, candidate-level reranking, and a gated LLM assistance. The evaluation uses a leakage-safe validation protocol: model and threshold selection are based on out-of-fold cross-validation predictions, while final reporting is performed on a locked test set with no shared source-file groups. The supervised conditional parent-code candidate generator reaches 74.32% Top-1 accuracy on the locked test set and 93.05% Top-10 candidate coverage. A direct zero-shot GPT-5.6-Sol baseline reaches 81.01% Top-1 accuracy, but requires an invocation of this large model for every diagnosis. The final deployable system, Pre-LLM fusion, reaches 81.24% OOF-CV accuracy and 83.19% locked-test accuracy while calling the LLM for only 31.97% of cases. This improves the locked-test error rate from 25.68% for the supervised parent-code baseline to 16.81%, corresponding to a 34.5% relative error reduction. The results show that the most effective role of the LLM is controlled reranking on uncertain examples rather than unrestricted coding or full-dataset invocation.
