Overview of the EHRSQL 2024 Shared Task on Reliable Text-to-SQL Modeling on Electronic Health Records

Electronic Health Records (EHRs) are relational databases that store theentire medical histories of patients within hospitals. They record numerousaspects of patients' medical care, from hospital admission and diagnosis totreatment and discharge. While EHRs are vital sources of clinical data,exploring them beyond a predefined set of queries requires skills in querylanguages like SQL. To make information retrieval more accessible, one strategyis to build a question-answering system, possibly leveraging text-to-SQL modelsthat can automatically translate natural language questions into correspondingSQL queries and use these queries to retrieve the answers. The EHRSQL 2024shared task aims to advance and promote research in developing aquestion-answering system for EHRs using text-to-SQL modeling, capable ofreliably providing requested answers to various healthcare professionals toimprove their clinical work processes and satisfy their needs. Among more than100 participants who applied to the shared task, eight teams completed theentire shared task processes and demonstrated a wide range of methods toeffectively solve this task. In this paper, we describe the task of reliabletext-to-SQL modeling, the dataset, and the methods and results of theparticipants. We hope this shared task will spur further research and insightsinto developing reliable question-answering systems for EHRs.