Dr. Andrew Atkinson joined the Department of Medicine in the Division of Infectious Diseases as an Assistant Professor in March 2024. Dr. Atkinson is a statistician and methodologist providing ad hoc and project-driven support for interventional and observational studies. His research involves developing methods for handling missing data, the adoption of the estimand framework to improve clarity, and developing novel clinical trial methods. He has a broad range of experience in the design and analysis of studies involving people with HIV, and also for antibiotic resistance, pediatric pharmacology and healthcare associated infections. Recently, he has used machine learning and network graph theoretic methods combined with Bayesian spatial-temporal modelling to investigate nosocomial outbreaks.
The further development of such approaches, also considering missing data, is of key interest and importance. After studies in mathematics and statistics, he completed a PhD in Biostatistics from the London School of Hygiene and Tropical Medicine, University of London, on the development of methods for investigating informative missingness for time to event outcomes. From 2015, he combined a position as Senior Statistician at the Infectious Diseases Division of the University Hospital in Bern, Switzerland with a similar role in the Pediatric Pharmacology and Pharmacometrics Research Center at the University Children’s Hospital in Basel (UKBB), Switzerland.
As part of a government funded research grant, he was a visiting researcher in the Division of Infectious Diseases in 2021-22. Prior to moving to St Louis, he was the Head of the Data Science group in the Pediatric Research Center at UKBB. Dr. Atkinson has a secondary appointment in the Institute for Informatics, Data Science and Biostatistics (I2DB) at Washington University School of Medicine. His key interests are methods for handling missing data in randomized controlled trials and for observational studies, novel designs for optimizing treatment (duration), the adoption of the estimand framework to help simplify and clarify study design and the intersection of machine learning approaches and missing date methods.