Seminars


For further information on any CSM events please e-mail: csm@lshtm.ac.uk.


Centre for Statistical Methodology Seminar
Meta-analysis and Systematic Reviews Theme

Friday 24 November 2017, 12:45-2:00pm
LG9, Keppel Street
Combining diverse information sources with the II-CC-FF paradigm, with applications in meta-analysis and beyond
Celine Cunen (University of Oslo)

Abstract: Combining information across different sources is an important statistical challenge, arising in many different fields. Combination is especially difficult when the sources are very diverse, requiring approaches beyond standard meta-analysis methods. The II-CC-FF paradigm (Independent Inspection, Confidence Conversion, Focused Fusion) is a general three-step method for such problems. The first step, II, uses different techniques to translate the information from each source to confidence distributions. Then these confidence distributions are transformed into confidence log-likelihoods in the CC-step, before being combined in the FF step. In this talk, the II-CC-FF scheme will be presented and illustrated by several non-standard applications. Demonstrating that II-CC-FF can be applied successfully both in traditional meta-analysis settings and in more challenging cases, where we need to combine classical experimental data (hard data) with less systematically collected information (soft data).

Centre for Statistical Methodology Seminar
Design and Analysis for Dependent Data Theme
Friday 1 December 2017, 12:45-2:00pm
LG9, Keppel Street
How to fit generalised linear mixed models and keep smiling!
Prof Patricia Solomon (University of Adelaide)

Abstract: Recently, I have been evaluating the mortality outcomes of Australian and New Zealand intensive care units. This has involved fitting complex hierarchical logistic regression models to large datasets, with mixed success in terms of model convergence and accuracy of parameter estimation. These model fitting issues motivated me to take a closer look at how well widely available statistical software, including R, SAS and Stata, perform. I will present key findings to date on parameter estimation, computing time and model selection using both simulated data, and data from the Australian and New Zealand Intensive Care Society Adult Patient Database. The data analysis involves fitting two and three level hierarchical models to more than 500,000 patient records. I will also recommend some dos and don’ts.

Centre for Statistical Methodology Seminar
Big Data Theme
Friday 26 January 2018, 12:45-2:00pm
Manson Lecture Theatre, Keppel Street
Statistical methods for real-time monitoring of health outcomes
Prof Peter Diggle (University of Lancaster)

Abstract: The terms “e-health” and “health informatics” are increasingly used to describe a body of methodology intended to extract useful information from health-related data that is routinely acquired, often in real-time, during the operation of a health care system, rather than through specific, planned research projects. These data are often temporally and/or spatially referenced and can be linked to other electronically accessible data-sets such as census records, environmental features or social media activity. In this talk, I will argue that current research on statistical modelling of spatially and/or temporally referenced data can make an important contribution to real-time analysis of e-health data, with a view to identifying underlying trends and, often more interestingly, unexpected departures from those trends. I will then describe several applications of these ideas that are in various stages of development.

Centre for Statistical Methodology Seminar
Causal Inference Theme
Thursday 22 February 2018, 12:45-2:00pm
LG9, Keppel Street
Title TBC
Prof Stijn Vansteelandt (University of Ghent & LSHTM)

Centre for Statistical Methodology Seminar
Time Series Regression Analysis Theme
Friday 23 March 2018, 12:45-2:00pm
LG81, Keppel Street
Title TBC
Antonio Gasparrini (LSHTM)