Data Tech 2026

Teams tasked with real world evidence generation often spend significant effort defining patient populations, clinical events, and procedures - only to manage those definitions in spreadsheets, ad hoc SQL, and scattered documentation. These approaches can negatively impact the quality of evidence generated via introduction of errors, no formal validation, and lack of proper governance.

In this session, I will introduce the Computable Phenotype Library (CPL): A lightweight framework for managing clinical definitions as reusable, validated, governed assets. Drawing from a practical case study in real-world evidence operations of a medical device company (Philips), I will cover the framework's core components: The master code set, an internal user interface, reusable phenotype artifacts, automated validation, centralized cloud storage, and an external user interface. Attendees will leave with practical design principals for improving phenotype quality, maintainability, reproducibility, and trust - along with lessons learned about why adoption is often the hardest part of building better data infrastructure.