Statistical Analysis of Large Administrative Health Databases: Emerging Challenges and Strategies

> français

Collaborative Research Team Project: 2017-2020

Administrative health data offer us a valuable opportunity to deal with practical situations where planned studies or randomized experiments are infeasible, unethical or too costly to perform. While administrative data offer researchers a rich source of information, they posit many statistical challenges, which have to be adequately addressed to ensure a positive trade-off between quantity versus quality of the data. Typically, health administrative data databases are large in scale and may be of variable quality: they record only those variables that are easily available (or, in the case of claims data, only variables relevant to billing and finances) without consideration of the use of these data in research or clinical practice.

Our primary goals are to address emerging statistical challenges in the analysis of large-scale administrative health data, such as the Clinical Practice Research Datalink (CPRD) and CAYACS databases, and to develop original and innovative methodology in advancing foundational work and facilitating genuine application. We target to provide valuable insights into making best use of available large administrative health data and enhance understanding of health care demand, hence realizing the best public health value by capitalizing on the information carried by abundant large-scale health administrative data.

The team leaders are Grace Y. Yi of the University of Waterloo, Robert Platt of McGill University and X. Joan Hu of Simon Fraser University. The collaborators include Michal Abrahamowicz of McGill University, Wenqing He of the University of Western Ontario, Lawrence McCandless of Simon Fraser University, Rhonda Rosychuk of the University of Alberta, Donna Spiegelman of the Harvard School of Public Health, Samy Suissa of McGill University, and Mireille Schnitzer of the Université de Montréal.

Comments are closed.