Seminars


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


Centre for Statistical Methodology Seminar
Friday 16 December 2016, 12:45-2:00pm
LG9, Keppel Street
Optimal designs: for richer or poorer?
Oleg Volkov (Xitific, UK)

Abstract: This talk considers design of experiments for estimating the parameter of a nonlinear response model. Traditionally these studies utilise “rich” designs, containing a dense grid of points. An alternative is “optimal” designs, which maximise the expected precision of estimates given an informative prior on the parameter. If such a prior is available experimenters should prefer optimal designs, at least in theory. To see whether they should in practice we consider examples from pre-clinical drug development. Hence we examine the drawbacks — but also the promise — of optimal designs. We also examine novel designs that could mitigate such drawbacks and thus lead to better experiments.

Centre for Statistical Methodology Seminar
Analysis of Clinical Trials Theme
Friday 27 January 2017, 12:45-2:00pm
LG9, Keppel Street
Central statistical monitoring in an academic clinical trials unit
Amy Kirkwood (UCL Cancer Trials Centre)

Abstract: Within our academic clinical trial unit we follow a risk based approach to monitoring, avoiding, when possible, on-site source data verification which is an expensive activity with little evidence that it is worthwhile. Central statistical monitoring (CSM) has been suggested as a cheaper alternative where checks are performed centrally without the need to visit all sites. We developed a suite of R-programs which could perform data checks at either a subject or site level using previously described methods or ones we developed. These aimed to find possible data errors such as outliers, incorrect dates, or anomalous data patterns; digit preference, values too close or too far from the means, unusual correlation structures, extreme variances which may indicate fraud or procedural errors and under-reporting of adverse events. The aim was to produce programs which would be quick and easy to apply and which would produce simple tables or easy-to-read figures. We will summarise the methods and, using examples from trials within our unit, show how they are implemented and that they can be easy to interpret. We found CSM to be a worthwhile alternative to on-site data checking and may be used to limit the number of site visits by targeting only sites which are picked up by the programs. The methods can identify incorrect or unusual data for a trial subject, or centres where the data considered together are too different to other centres and therefore should be reviewed, possibly through an on-site monitoring visit.

Centre for Statistical Methodology Seminar
Friday 24 February 2017, 12:45-2:00pm
LG9, Keppel Street
Linking administrative data for epidemiological research
Katie Harron (LSHTM)

Centre for Statistical Methodology Seminar
Survival Analysis and Statistical Computing Themes
Friday 31 March 2017, 12:45-2:00pm
LG9, Keppel Street
Multistate survival analysis in Stata
Michael Crowther (University of Leicester)

Abstract: Multi-state models are increasingly being used to model complex disease profiles. By modelling transitions between disease states, accounting for competing events at each transition, we can gain a much richer understanding of patient trajectories and how risk factors impact over the entire disease pathway. In this talk, I will introduce some new Stata commands for the analysis of multi-state survival data. This includes msset, a data preparation tool that converts a dataset from wide (one observation per subject, multiple time and status variables) to long (one observation for each transition for which a subject is at risk for). I develop a new estimation command, stms, that allows the user to fit different parametric distributions for different transitions, simultaneously, while allowing for sharing of covariate effects across transitions. Finally, predictms calculates transition probabilities, and many other useful measures of absolute risk, following the fit of any model using streg, stms, or stcox, using either a simulation approach or the Aalen–Johansen estimator. Importantly, predictms also allows different parametric distributions to be specified for different transitions, passed as model objects. I illustrate the software using a dataset of patients with primary breast cancer.

Centre for Statistical Methodology Seminar
Friday 28 April 2017, 12:45-2:00pm
LG9, Keppel Street
Title TBC
Nuno Sepulveda (LSHTM)

Centre for Statistical Methodology Seminar
Friday 26 May 2017, 12:45-2:00pm
LG9, Keppel Street
Title TBC
Suzie Cro (Imperial College London)

Centre for Statistical Methodology Seminar
Friday 30 June 2017, 12:45-2:00pm
LG9, Keppel Street
Title TBC
Rebecca Walwyn (University of Leeds)