University of Washington School of Public Health

UW SPH News: Biostatistics students win at 2018 WNAR student paper competition

Biostatistics students win at 2018 WNAR student paper competition

08/01/2018
Share:

Four students from the Department of Biostatistics at the University of Washington School of Public Health took home top honors last month at the 2018 student paper competition hosted by the Western North American Region (WNAR) of the International Biometric Society.

Anu Mishra and Katherine Wilson shared first place for Most Outstanding Written Paper. In clinical settings, intervention decisions are sometimes made by comparing predicted risks to a predefined risk threshold. Mishra’s winning paper proposed a weighted approach for recalibration methods that account for how the risk model will be used. It also illustrated how such methods produce risk models that are well calibrated in the region that affects clinical decision-making.

Wilson’s paper discussed incorporating data from summary birth histories to analyze child mortality trends in developing regions through a data augmentation scheme within a Bayesian framework. This approach reduces uncertainty when assessing goals to reduce child mortality.

Additionally, Phuong Vu won the award for Most Outstanding Oral Presentation. Her winning presentation proposed a probabilistic framework to a novel Principal component analysis (PCA) algorithm previously developed to handle spatially-misaligned multivariate air pollution data. The new method allows for flexible model-based imputation that can directly handle missing data and account for geographical information. The ultimate goal is to improve the performance of the exposure modeling stage in air pollution studies and, as a result, to gain better understanding of the impacts of multi-pollutant exposures on human health.

Finally, Kelsey Grinde won the Distinguished Oral Presentation award. Her presentation proposed an approach for deriving significance thresholds that control the probability of making one or more false discoveries in genome-wide admixture mapping studies. This finding holds many practical implications for researchers regarding best practices for admixture mapping.