The session will consist of four talks given by leading researchers on the topic of statistical data integration. All four presenters are junior, female academics from diverse institutions working in biostatistical application areas.

Data integration methods for multi-source data borrow and leverage information from multiple sources for increased statistical efficiency and greater population coverage to yield novel scientific findings. By jointly utilizing data from previously disconnected sources, new scientific hypotheses have been explored that yield a deeper understanding of human biology. Particular challenges that arise in the integration of multi-source data relate to heterogeneity, privacy constraints, and trade-offs between bias and efficiency. The development of methods to address these challenges is key for the long term success of data integration. The primary focus of this session is to present the statistical challenges, opportunities and recent developments in integrative inference with data from multiple sources. Drs. Duan, Gu, Jin and Wang, four outstanding junior women working in this field, will discuss recent cutting-edge methodological contributions in data integration problems. The scheduled talk titles are:
Dr. Rui Duan: Federated and transfer learning for healthcare data integration
Dr. Tian Gu: Communication-efficient transfer learning for multi-site risk prediction
Dr. Jin Jin: A Quasi-Bayesian framework for integrating summary-level information of multiple models with disparate sets of covariates
Dr. Jingshen Wang: Electronic medical record data assisted adaptive experimental design

Organiser: Dr Emily Hector 

Chair: Dr Emily Hector 

Speaker: Jin Jin 

Speaker: Rui Duan 

Speaker: Tian Gu 

Speaker: Dr Emily Hector