Aurum at CRF Filling 2026: Modular DSPy Extractors with Qwen3-Max for Multilingual CRF Filling
Year: 2026
Type: conference-paper
Abstract: This paper describes the submission by Team Aurum to the CL4Health @ LREC 2026 Shared Task on Case Report Form (CRF) Filling from dyspnea patient clinical notes.Extracting 134 structured clinical fields using a single Large Language Model (LLM) call often leads to schema-following errors, hallucination, and poor attention over complex instructions.To address this, we propose a modular extraction pipeline built with DSPy, which decomposes the massive CRF schema into 14 specialized, manually designed domain extractors (e.g., Medical History, Lab Values, Acute Diagnoses).Using Chain-of-Thought reasoning and strict Pydantic-typed validation with the Qwen3-Max (Thinking) model, our pipeline achieved an official Codabench Test Macro-F1 score of 0.68 in English and 0.67 in Italian, securing the 1st place ranking overall in the shared task.Through rigorous manual error analysis and iterative prompt optimization, we demonstrate that establishing explicit instruction boundaries between "not mentioned" (None) and "explicitly absent" (False) is critical for clinical information extraction, and that architectural modularity yields higher performance gains than simply scaling model parameters.
Authors: Vinay Babu Ulli, Jyoti Kumari, Anindita Mondal
Institution: Banaras Hindu University
Language: English
Subfield: Molecular Biology
Domain: Life Sciences
Open Access status: hybrid
OpenAlex ID: W7166555572