# Unknown is not normal: separating language-model extraction from rule-based decision logic for clinical risk scores

arXiv: https://arxiv.org/abs/2609.34112 · HTML: https://mlsanity.com/paper/2609.34112/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 10 of its 30 references that are in the map.
- First posted on arXiv 28 Sep 2026.

## Builds on (most cited first, 10 of 10)

- [On Calibration of Modern Neural Networks](https://mlsanity.com/paper/1706.04599/) (2017)
- [Selective Classification for Deep Neural Networks](https://mlsanity.com/paper/1705.08500/) (2017)
- [AgentClinic: a multimodal agent benchmark to evaluate AI in simulated clinical environments](https://mlsanity.com/paper/2405.07960/) (2024)
- [MediQ: Question-Asking LLMs and a Benchmark for Reliable Interactive Clinical Reasoning](https://mlsanity.com/paper/2406.00922/) (2024)
- [MedCalc-Bench: Evaluating Large Language Models for Medical Calculations](https://mlsanity.com/paper/2406.12036/) (2024)
- [From Scores to Steps: Diagnosing and Improving LLM Performance in Evidence-Based Medical Calculations](https://mlsanity.com/paper/2509.16584/) (2025)
- [MedMCP-Calc: Benchmarking LLMs for Realistic Medical Calculator Scenarios via MCP Integration](https://mlsanity.com/paper/2601.23049/) (2026)
- [ClinDet-Bench: Beyond Abstention, Evaluating Judgment Determinability of LLMs in Clinical Decision-Making](https://mlsanity.com/paper/2602.22771/) (2026)
- [MedCalc-Bench Doesn't Measure What You Think: A Benchmark Audit and the Case for Open-Book Evaluation](https://mlsanity.com/paper/2603.02222/) (2026)
- [Scalable Stewardship of an LLM-Assisted Clinical Benchmark with Physician Oversight](https://mlsanity.com/paper/2512.19691/) (2025)

## Abstract (arXiv)

Large language models (LLMs) are increasingly used to compute clinical risk scores from free-text notes. Notes are often incomplete, and treating undocumented findings as normal can silently misclassify patients. We test whether separating three-state extraction (present, absent or unknown, by an LLM) from decision logic (deterministic code computing score bounds over unknown inputs) lets a system ask only questions that can change the decision. On 1,200 synthetic emergency cases across six calculators (HEART, CURB-65, qSOFA, PERC, Wells, Cockcroft-Gault), with a simulated clinician answering questions, we compared this bounds policy with asking for every missing input, a missing-equals-normal schema, and an end-to-end LLM agent (Claude Opus 5.5). With Claude Haiku 4.5 as extractor, the bounds policy matched ask-all accuracy (99.4% vs 99.4%) with half the questions (0.92 vs 1.78 per case) and no irrelevant ones. Treating missing as normal dropped accuracy to 91.2% and under-triaged 8.5% of patients (95% CI 7.1-10.2), and under-triage persisted under messy notes and a noisy clinician. The agent was equally accurate under ideal conditions (99.6%) but 9.5% of its questions were irrelevant; with a noisy clinician it was less accurate than the bounds policy (83.5% vs 87.0%, p<0.001) and committed prematurely in 2.7% of cases (bounds: 0%). A 9B local model as extractor reached oracle-level accuracy (99.8%). In 584 real case reports from MedCalc-Bench, only 52% contained enough information to determine the category (HEART 13%). Routing decisions through code that reasons explicitly about unknowns avoids premature commitment and irrelevant questions, halves the questions asked, and works with small local models.