Dataset and Enhanced Model for Eligibility Criteria-to-SQL Semantic Parsing

摘要: Clinical trials often require that patients meet eligibility criteria (e.g., have specific conditions) to ensure the safety and effectiveness of studies. However, retrieving eligible for a trial from electronic health record (EHR) database remains challenging task clinicians since it requires not only medical knowledge about criteria, but also an adequate understanding structured query language (SQL). In this paper, we introduce new dataset includes first-of-its-kind eligibility-criteria corpus corresponding queries criteria-to-sql (Criteria2SQL), translating executable SQL queries. Compared existing datasets, in here are derived clinical include Order-sensitive, Counting-based, Boolean-type cases which seen before. addition dataset, propose novel neural semantic parser as strong baseline model. Extensive experiments show proposed outperforms state-of-the-art general-purpose text-to-sql models while highlighting challenges presented by dataset. The uniqueness diversity leave lot research opportunities future improvement.