Saved in:
| Main Authors: | , , , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2025
|
| Subjects: | |
| Online Access: | https://arxiv.org/abs/2505.21801 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866908551036796928 |
|---|---|
| author | Stoisser, Josefa Lia Martell, Marc Boubnovski Märtens, Kaspar Phillips, Lawrence Town, Stephen Michael Donovan-Maiye, Rory Fauqueur, Julien |
| author_facet | Stoisser, Josefa Lia Martell, Marc Boubnovski Märtens, Kaspar Phillips, Lawrence Town, Stephen Michael Donovan-Maiye, Rory Fauqueur, Julien |
| contents | Electronic health records (EHRs) contain richly structured, longitudinal data essential for predictive modeling, yet stringent privacy regulations (e.g., HIPAA, GDPR) often restrict access to individual-level records. We introduce \textbf{Query, Don't Train} (QDT): a \textbf{structured-data foundation-model interface} enabling \textbf{tabular inference} via LLM-generated SQL over EHRs. Instead of training on or accessing individual-level examples, QDT uses a large language model (LLM) as a schema-aware query planner to generate privacy-compliant SQL queries from a natural language task description and a test-time input. The model then extracts summary-level population statistics through these SQL queries, and the LLM performs chain-of-thought reasoning over the results to make predictions. This inference-time-only approach enables prediction without supervised model training, ensures interpretability through symbolic, auditable queries, naturally handles missing features without imputation or preprocessing, and effectively manages high-dimensional numerical data to enhance analytical capabilities. We validate QDT on the task of 30-day hospital readmission prediction for Type 2 diabetes patients using a MIMIC-style EHR cohort, achieving F1 = 0.70, which outperforms TabPFN (F1 = 0.68). To our knowledge, this is the first demonstration of LLM-driven, privacy-preserving structured prediction using only schema metadata and aggregate statistics -- offering a scalable, interpretable, and regulation-compliant alternative to conventional foundation-model pipelines. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2505_21801 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | Query, Don't Train: Privacy-Preserving Tabular Prediction from EHR Data via SQL Queries Stoisser, Josefa Lia Martell, Marc Boubnovski Märtens, Kaspar Phillips, Lawrence Town, Stephen Michael Donovan-Maiye, Rory Fauqueur, Julien Databases Electronic health records (EHRs) contain richly structured, longitudinal data essential for predictive modeling, yet stringent privacy regulations (e.g., HIPAA, GDPR) often restrict access to individual-level records. We introduce \textbf{Query, Don't Train} (QDT): a \textbf{structured-data foundation-model interface} enabling \textbf{tabular inference} via LLM-generated SQL over EHRs. Instead of training on or accessing individual-level examples, QDT uses a large language model (LLM) as a schema-aware query planner to generate privacy-compliant SQL queries from a natural language task description and a test-time input. The model then extracts summary-level population statistics through these SQL queries, and the LLM performs chain-of-thought reasoning over the results to make predictions. This inference-time-only approach enables prediction without supervised model training, ensures interpretability through symbolic, auditable queries, naturally handles missing features without imputation or preprocessing, and effectively manages high-dimensional numerical data to enhance analytical capabilities. We validate QDT on the task of 30-day hospital readmission prediction for Type 2 diabetes patients using a MIMIC-style EHR cohort, achieving F1 = 0.70, which outperforms TabPFN (F1 = 0.68). To our knowledge, this is the first demonstration of LLM-driven, privacy-preserving structured prediction using only schema metadata and aggregate statistics -- offering a scalable, interpretable, and regulation-compliant alternative to conventional foundation-model pipelines. |
| title | Query, Don't Train: Privacy-Preserving Tabular Prediction from EHR Data via SQL Queries |
| topic | Databases |
| url | https://arxiv.org/abs/2505.21801 |