======================================================================= SEMINARS MRC BIOSTATISTICS UNIT, CAMBRIDGE - EASTER TERM 2004 ======================================================================= 20th April Professor Martin Crowder, Department of Mathematics, Imperial College London. Title: Competing Risks: a brief intro. Abstract: If something can fail, it can often fail in one of several ways and sometimes in more than one way at a time. It is puzzling that so much work is done in Survival Analysis without mentioning Competing Risks. There is always some cause of failure, and almost always more than one possible cause. In this sense Survival Analysis is a lost cause. The origins of Competing Risks can be traced back to Daniel Bernoulli's attempt in 1760 to disentangle the risk of dying from smallpox from other risks. Much of the work in the subject since that time has been demographic and actuarial in nature. However, the results of major importance for statistical inference, and applications of the theory in Reliability and Survival Analysis, are quite recent, these fields themselves being relatively recent. The talk will cover some of the basic ideas and application of the subject. ============ 11th May Dr. Nicky Welton, MRC Health Services Research Collaboration, Department of Social Medicine, University of Bristol. Title: Estimation of Markov Chain transition probabilities and rates and the propagation of uncertainty. Abstract: Markov transition models are frequently used to model disease progression. We show how the solution to Kolmogorov's forward equations can be exploited to map between transition rates and probabilities from probability data in multi-state models. We provide a uniform, Bayesian treatment of estimation and propagation of uncertainty of transition rates and probabilities from three commonly encountered data structures. 1) Successive observations of state membership are made on short intervals (only one transition can occur per interval). 2) Observations are available on all transitions and exact time at risk in each state (event history data). 3) We observe the initial state and final state after a fixed interval of time, but not the sequence of transitions. With the third data structure we show how underlying transition rates can be recovered from the data using Markov chain Monte Carlo methods, and suggest diagnostics to investigate inconsistencies between evidence from different starting states. We end by discussing how these methods might be used for various evidence structures in the meta-analysis of randomised trials. ============ 25th May Professor Vern Farewell, MRC Biostatistics Unit, Institute of Public Health, University of Cambridge. Title: Counting Buses: Approaches to the Analysis of SARS Incubation Times. Abstract: The importance of quarantine in the management of the SARS epidemic made understanding of the distribution of incubation times important. In this talk, I consider methodology which might be used in the early days of an epidemic of a new disease to characterise what is known about incubation times ============ 15th June Dr. Graham Medley, Department of Biological Sciences, University of Warwick. Title: Estimating infection and recovery rates from longitudinal data. ======================================================================= Seminars start at 2.30pm in the Large Seminar Room, 1st Floor, Institute of Public Health, University Forvie Site, Robinson Way,Cambridge. All welcome to attend. For details on seminars at the Institute of Public Health see http://www.iph.cam.ac.uk/seminars.shtml =======================================================================