SCARE: A Benchmark for SQL Correction and Question Answerability Classification for Reliable EHR Question Answering

TL;DR

SCARE is a comprehensive benchmark designed to evaluate post-hoc verification methods for SQL queries and question answerability in EHR question answering systems, addressing safety concerns in clinical AI deployment.

Contribution

The paper introduces SCARE, the first unified benchmark for assessing safety verification mechanisms in EHR QA systems, including question classification and SQL correction.

Findings

Identified a trade-off between question classification accuracy and SQL error correction.

Benchmark includes 4,200 question-SQL-output triples from multiple models and datasets.

Experimental results highlight key challenges and future research directions.

Abstract

Recent advances in Large Language Models (LLMs) have enabled the development of text-to-SQL models that allow clinicians to query structured data stored in Electronic Health Records (EHRs) using natural language. However, deploying these models for EHR question answering (QA) systems in safety-critical clinical environments remains challenging: incorrect SQL queries-whether caused by model errors or problematic user inputs-can undermine clinical decision-making and jeopardize patient care. While prior work has mainly focused on improving SQL generation accuracy or filtering questions before execution, there is a lack of a unified benchmark for evaluating independent post-hoc verification mechanisms (i.e., a component that inspects and validates the generated SQL before execution), which is crucial for safe deployment. To fill this gap, we introduce SCARE, a benchmark for evaluating methods that function as a post-hoc safety layer in EHR QA systems. SCARE evaluates the joint task of (1) classifying question answerability (i.e., determining whether a question is answerable, ambiguous, or unanswerable) and (2) verifying or correcting candidate SQL queries. The benchmark comprises 4,200 triples of questions, candidate SQL queries, and expected model outputs, grounded in the MIMIC-III, MIMIC-IV, and eICU databases. It covers a diverse set of questions and corresponding candidate SQL queries generated by seven different text-to-SQL models, ensuring a realistic and challenging evaluation. Using SCARE, we benchmark a range of approaches-from two-stage methods to agentic frameworks. Our experiments reveal a critical trade-off between question classification and SQL error correction, highlighting key challenges and outlining directions for future research.