OCTRU Lead Statistician
Sofia joined NDORMS in February 2023 and she is interested in developing sound and rigorous statistical models for estimating parameters of interest from clinical trials data and health research data.
Before joining NDORMS she was a Senior Statistician and Module Lead in Statistics in NDPH where she co-lead projects on large population studies (UKB, CKB, Cuba Prospective Study, Life Line Screening, TILDA among others) on developing risk prediction models for cardiovascular disease. She also investigated the association of infectious respiratory diseases with smoking and adiposity and was involved in further research in frailty and its assessment in the older population. During this period she was also a Statistical Editor for Cochrane Reviews.
During her PhD she worked on theoretical aspects of Statistics introducing the concept of structural meta-analysis as the combination of evidence from studies where the interest is in finding shared conditional independence relations across statistical models with partially overlapping variables.She found conditions under which conflicting evidence can be encoded allowing for different types of statistical combinations of models building on the concept of hyper Markov laws. The estimation of the parameters of the models expanded the use of the EM algorithm for missing data.
She also extended the use of Gaussian graphical models for case-control studies from microarray data where the interest is in testing the equality of means across two or more experimental conditions. The equality of means was augmented with a further test about the equality of the concentration matrices representing the conditional independence relations.
Sofia has also been a Departmental Lecturer in Statistics teaching Statistical Theory and applications at all levels (college tutorials, undergraduate and MSc students and first year DPhil students) and continuining her research in graphical models and their applications.
She is currently supervising four DPhil student and has supervised one DPhil student to completion. She has also supervised more than 25 MSc students across Statistics, Applied Statistics and Population Health. She has also been an assessor for MSc dissertations and exams and internal and external examiner for DPhil Theses.
Systematic review of the utility of the frailty index and frailty phenotype to predict all-cause mortality in older people.
Kim DJ. et al, (2022), Syst rev, 11
Independent effects of adiposity measures on risk of atrial fibrillation in men and women: a study of 0.5 million individuals.
Camm CF. et al, (2022), Int j epidemiol, 51, 984 - 995
Development of a Model to Predict 10-Year Risk of Ischemic and Hemorrhagic Stroke and Ischemic Heart Disease Using the China Kadoorie Biobank.
Yang S. et al, (2022), Neurology, 98, e2307 - e2317
Variability and agreement of frailty measures and risk of falls, hospital admissions and mortality in TILDA.
Kim DJ. et al, (2022), Sci rep, 12
The impact of case and contact characteristics on contact tracing during the West Africa Ebola epidemic.
Longley JL. et al, (2021), J infect, 83, 496 - 522