Case-Specific Rubrics for Clinical AI Evaluation: Methodology, Validation, and LLM-Clinician Agreement Across 823 Encounters

2026-04-27

AI Summary

This study addresses the lack of evaluation frameworks for clinical AI documentation systems that simultaneously account for clinical validity, cost-effectiveness, and adaptability to iterative improvements. Traditional expert review is prohibitively expensive and inefficient. The authors propose a novel paradigm based on case-specific scoring rules: clinicians author rules, and—for the first time—it is demonstrated that large language models (LLMs) can automatically generate rules whose quality matches or exceeds inter-clinician agreement. Experiments show that clinician-authored rules effectively discriminate between high- and low-quality outputs (median score gap: 82.9%), improving system-level accuracy from 84% to 95%. Moreover, LLM-generated rankings exhibit significantly higher concordance with clinician judgments (Kendall’s τ: 0.42–0.46) than inter-clinician agreement itself (τ: 0.38–0.43), while reducing evaluation costs by approximately three orders of magnitude—thus combining expert-level precision with scalable automation.

Technology Category

• Machine Learning: Large Multimodal Models (LMMs)
• Natural Language Processing: Generation
• Humans and AI: Human-in-the-loop Machine Learning

Application Category

• Economics, Online Markets and Human Computation: Cost models of using LLMs in production systems
• Search and Retrieval-Augmented AI: Web evaluation methodologies and metrics
• Semantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactions

Abstract

Objective. Clinical AI documentation systems require evaluation methodologies that are clinically valid, economically viable, and sensitive to iterative changes. Methods requiring expert review per scoring instance are too slow and expensive for safe, iterative deployment. We present a case-specific, clinician-authored rubric methodology for clinical AI evaluation and examine whether LLM-generated rubrics can approximate clinician agreement. Materials and Methods. Twenty clinicians authored 1,646 rubrics for 823 clinical cases (736 real-world, 87 synthetic) across primary care, psychiatry, oncology, and behavioral health. Each rubric was validated by confirming that an LLM-based scoring agent consistently scored clinician-preferred outputs higher than rejected ones. Seven versions of an EHR-embedded AI agent for clinicians were evaluated across all cases. Results. Clinician-authored rubrics discriminated effectively between high- and low-quality outputs (median score gap: 82.9%) with high scoring stability (median range: 0.00%). Median scores improved from 84% to 95%. In later experiments, clinician-LLM ranking agreement (tau: 0.42-0.46) matched or exceeded clinician-clinician agreement (tau: 0.38-0.43), attributable to both ceiling compression and LLM rubric improvement. Discussion. This convergence supports incorporating LLM rubrics alongside clinician-authored ones. At roughly 1,000 times lower cost, LLM rubrics enable substantially greater evaluation coverage, while continued clinical authorship grounds evaluation in expert judgment. Ceiling compression poses a methodological challenge for future inter-rater agreement studies. Conclusion. Case-specific rubrics offer a path for clinical AI evaluation that preserves expert judgment while enabling automation at three orders lower cost. Clinician-authored rubrics establish the baseline against which LLM rubrics are validated.

Problem

• clinical AI evaluation
• case-specific rubrics
• expert judgment
• iterative deployment
• evaluation methodology

Innovation

• case-specific rubrics
• clinical AI evaluation
• LLM-clinician agreement
• automated scoring
• EHR-embedded AI