Description

Contributions to the community.

OMOPCDMPathways.jl is a new package in the JuliaHealth ecosystem that standardizes the analysis of patient care progression using OMOP CDM–formatted data. The OMOP Common Data Model, established by OHDSI, enables consistent analysis of real-world health data (e.g., medical claims, electronic health records). OHDSI is a collaborative dedicated to observational health research. This package supports studies in health economics, pharmacovigilance, and patient monitoring by tracking treatment trajectories over time, thereby revealing gaps or abrupt changes in care plans that may signal emerging safety concerns. It is already utilized by researchers at institutions and health systems across the globe.

Topic diversity.

Patient pathways trace a patient's journey from initial contact through subsequent provider interactions. Our package builds on functionality from other JuliaHealth projects (e.g., OMOPCDMCohortCreator.jl) to construct detailed treatment histories that can be visualized via interactive sunburst plots. These visualizations can represent the hierarchical structure of patient treatment pathways, which is particularly useful in observational health research, where the sequence and timing of events—such as overlapping treatments—are key to understanding patient outcomes.

Applicability to the Julia community

This package is highly valuable for Julia users, particularly researchers in health informatics research. Using Julia's performance and the OMOP CDM standard, it uses Julia's performance to construct patient pathways. These pathways can then be used as rich feature sets for machine learning models to predict outcomes such as readmission, adherence to treatment, or adverse drug reactions, advancing personalized medicine. This integration is particularly powerful in personalised medicine, where understanding individual treatment trajectories is key to tailoring future care.

Significance to the community

This package is highly significant to the Julia community because it demonstrates how open source tools can be used to build and analyze patient treatment pathways from scratch. In our talk, we will walk through a hands-on example using synthetic patient data, showing attendees exactly how to transform raw clinical data (present in OMOPCDM format) into meaningful patient pathways. This will not only highlight the power of this software but also it's seamless integration using OMOPCDMCohortCreator.jl.

Clarity

Attendees will learn step by step how to select treatments of interest, filter and combine treatment eras, and build detailed patient treatment histories. They will also see how to visualize these pathways using sunburst plots and how to extract features for machine learning predictions, thereby gaining practical skills that can be directly applied to real-world clinical data analysis.