Bayesian Analysis of Longitudinal Studies

Estimated cost: R5001 – R10 000
Field Of Study: Statistics
Department Name: SACEMA The SA DST NRF Centre of Excellence in Epidemiologica


¿ To provide theoretical underpinnings, illustrations and practise in the Bayesian approach for the analysis of longitudinal studies. (These constitute an important class of studies in clinical research and arise when a response is repeatedly measured over time. This is in contrast to a survival study where the time to an event is recorded. In the first case the statistical analysis involves techniques for repeated measures, while in the second case one essentially uses survival analysis. In the last decade a combination of the two statistical techniques is becoming popular and involves the joint modelling of survival times and longitudinal measurements. Longitudinal data can be analysed in the classical, frequentist framework, but the Bayesian approach offers more flexible modelling options which are useful for complex data structures, multivariate outcomes, multiple levels in the data, some missing data patterns, joint modelling, etc.) ¿ The course will cover the Bayesian approach as expounded in the book: Bayesian Biostatistics, by Lesaffre and Lawson, and also the analysis of follow-up studies as in the book: Missing Data in Longitudinal Studies: Strategies for Bayesian Modelling and Sensitivity Analysis, by Daniels and Hogan. ¿ To provide a great variety of examples and analyse them using various relevant software packages, including WinBUGS/OpenBUGS/JAGS and R-versions of them.

Entry Requirement: Professional 4-year Bachelor’s degree /Bachelor Honours degree or Equivalent Qualification
Certificate Type: Attendance
Duration: 5 days
Presentation Language: English
Latest Year Offered: 2016