Techniques and Technologies for Generating Synthetic Longitudinal Data
An overview of techniques and technologies for generating longitudinal synthetic data
Synthetic data generation (SDG) is emerging as practical approach to create non-identifiable information. This allows the information to be shared for secondary analysis with minimal obligations according to contemporary privacy statutes. The availability of such information is critical for health research and innovation in the life sciences sector.
SDG techniques work very well for tabular data, with good results available demonstrating high utility and very small privacy risks. However, the application of SDG to longitudinal data requires a different set of approaches and technologies.
Longitudinal data means that individuals have multiple events over time. Most data used in health research and analytics is longitudinal. In this webinar we will discuss the different approaches that have been used to generate synthetic variants for this kind of data and show examples of how this works.
Khaled El Emam and Lucy Mosquera from the Replica Analytics team will present some of their experiences and results on this topic, and illustrate some of what to expect from longitudinal synthetic data.