CEHR-BERT: Incorporating temporal information from structured EHR data to improve prediction tasks

[ICML'25] MedTok: Multimodal Medical Code Tokenizer

Code to generate realistic synthetic healthcare data with diffusion models

[arXiv 2025] Pre-training script for Clinical ModernBERT

COVID-19 EHR data analysis pipeline

KDD2020 paper; Identifying Sepsis Subphenotypes via Time-Aware Multi-Modal Auto-Encoder

[npj Digital Medicine 2025] Multiple Embedding Model for EHR (MEME) used for strong prediction on Emergency Department tasks

Built a regression model that predicts the expected days of hospitalization time and an uncertainty range estimation.

attribute-based access control implementation for EHRs

Pipeline of Data Extraction, Preprocessing, Representation, and Training for MIMIC-III

R package for cleaning anthropometric (height and weight) data from electronic health record systems.

omopcept : an R package to access OMOP conCEPTs (no cons!) and flexible tidyverse compatible R functions for querying and visualising.

LLM graph-RAG SQL generator for large databases with poor documentation

A software toolkit for the interconversion of standard data models for phenotypic data

Official implementation of TACCO (Task-guided Co-clustering).

Nature Medicine paper. A Multidimensional Precision Medicine Approach for Autism Subtype Identification.

HACSurv: A Hierarchical Copula-based Approach for Survival Analysis with Dependent Competing Risks

This is a set of useful tools for using, creating, validating and generally working with Codelists in Health Research. The tools are in Rust with Python and R bindings so they can be used in any of the 3 languages.

Pipeline for building Machine Learning Classifiers for the diagnosis of EHR text-data. We used this pipeline for our study, published here: https://doi.org/10.2196/23930.

Introduction to CPRD using synthetic datasets