Population Health Data Science Research Seminar Series

Staff and students from Queen Mary and Barts Health NHS Trust are invited to join us, to: 

• See and share examples of research excellence in health data science from across the Wolfson Institute of Population Health (WIPH) and Barts Cancer Institute (BCI), Queen Mary, Barts Health, and more widely
• Develop deeper knowledge of the many health data sources available for research and hear directly from colleagues who are using them
• Meet new collaborators and share expertise in using electronic health records and multimodal data to improve population health

Upcoming

Thursday 16 May, 12.00-12.50

Open science for epidemiologists: Should we all be publishing our code?

Presentation by Anna Schultze (London School of Hygiene and Tropical Medicine). Transparency in epidemiological research is critically important, especially given the increased use of real-world evidence for regulatory decision making. Many open science resources encourage the sharing of all research materials, including data and code, to enable analyses to be independently verified and reproduced. Although data sharing is rarely possible for epidemiologists given stringent data protection, sharing code (in which the programming code used to process and analyse research data is made public) often is. The first half of this talk will introduce the rationale for sharing programming code and present some results from a recent systematic review on code sharing in pharmacoepidemiology. The second half will involve an interactive series of polls and discussion about code sharing, including potential barriers and how to overcome them.  

Thursday 20 June, 10.00-10.50

Designing observational studies to emulate a randomised controlled trial 

Presentation by Dr Kevin Wing from the London School of Hygiene and Tropical Medicine, with Dr Paris Baptiste. This seminar will provide an overview of the “target trial” framework, which adds an additional layer of rigour and transparency to observational studies. We will discuss how emulating an existing randomised controlled trial in observational data can improve confidence in the results and generate evidence for patient groups that are typically underrepresented in trials. We’ll also consider the challenges that remain, including residual or unknown confounders - how can research teams know that their observational study reports a “true” effect?

Past seminars

With the consent of the speakers, recordings of our seminars are available to watch here after the event.

Organising committee

Co-chairs: Rohini Mathur, Jianhua Wu (WIPH) and Claude Chelala (BCI) 
Committee members: Paris Baptiste, Fabiola Eto, Harriet Larvin, Jing Hui Law, Judith Offman, Stuart Rison, Mary Thomas, Nicola Firman (WIPH)