Bias identification and reduction techniques for large electronic medical record datasets

We study how differences in registration and healthcare use can skew research findings and how we can correct for them in a simple way. We work with pseudonymised Intego data; individuals cannot be identified.

This study examines how to better leverage electronic medical records (EMRs) in general practice for research. Variations in recording and care‑seeking can bias results. We aim to understand these sources of bias and explore simple statistical strategies to mitigate them, using pseudonymised data from the Intego database. No direct access to personal medical records; privacy is strictly protected.