Simulated datasets for health data access and research

Simulated datasets are a useful and innovative approach to facilitating safe access to patient data for research.  By aggregating cohorts and sampling fields at random, we can create realistic models of health databases that can be used to develop queries and explore the structure of the data.  These generative algorithms come with privacy guarantees, so the resulting simulated datasets can be used widely.  In this talk I will discuss some of the methods used to generate simulated datasets.   In addition there will be a practical case study of using the simulated datasets as a model for the data held by Public Health England’s National Cancer Registration and Analysis Service to facilitate innovative research in industry.