Four Great PhD Opportunities in Medical Statistics at Lancaster University
The Department of Mathematics and Statistics at Lancaster University is pleased to announce four PhD Studentships in Medical Statistics, available for October 2009.
The topics are:
* Methods for extreme values in laboratory safety data
* Joint modelling of repeated measurements and failure times in degenerative medical conditions
* Markov chain Monte Carlo methods for interpreting zoonosis data
* Selection of treatments in phase II clinical trials based on measures of safety and efficacy
Methods for extreme values in laboratory safety data
Sponsor: AstraZeneca
Supervisors: Professor J Tawn, Dr E Eastoe and Dr H Southworth (AZ)
Duration: 3 years
Starting Stipend: £16,000 pa Laboratory data are routinely collected on all participants in pharmaceutical clinical trials. Such data are usually interpreted using summary statistics such as means, medians, standard deviations and counts of observations above certain thresholds. Extreme values of these laboratory variables may indicate potential safety issues. This project would explore the application of multivariate, semi-parametric extreme value methodology (Heffernan and Tawn, JRSS, Series B 66, 497-546, 2004) in their interpretation. Outstanding inference problems that need to be resolved before this can be done include investigation of the identifiability of model parameters from small data sets, the development of hierarchical models for parameters, the use of Bayesian inference, the inclusion of patient specific covariates and allowance for repeated measurements on each patient. Data from studies conducted by AstraZeneca will be available to guide the development of the methodology.
Joint modelling of repeated measurements and failure times in degenerative medical conditions
Sponsor: Medical Research Council
Supervisors: Dr D Costain and Professor J Whitehead
Duration: 4 years
Starting Stipend: £15,290 pa
In medical follow-up studies, observations on the progress of subjects will be collected repeatedly at scheduled visits to the clinic. In addition, the times occurrence of major events will be recorded. For example, in studies of recovery following a stroke, a patient's ability to perform everyday tasks will be regularly assessed, while the times of events such as subsequent strokes or death will also be recorded. Data collected in this way are useful in forming prognoses and identifying high risk groups. Methods for combining normally distributed repeated measures with failure times have already been developed (Diggle et al. Statistics in Medicine 27, 2981-2998, 2008). This project will consider binary and ordinal responses, which are more common in therapeutic areas such as stroke. Clinical data relating to the progression of patients following stroke will be available for analysis while developing these methods.
Markov chain Monte Carlo methods for interpreting zoonosis data
Sponsor: EPSRC
Supervisors: Dr C Sherlock and Professor P Fearnhead
Duration: 3½ years
Stipend: to be determined
A zoonosis is a disease which is transmissible from animals to humans: examples include bird-flu and various diseases that are endemic in rodent populations. This project will involve the analysis of data from a 5 year study of the prevalence of cowpox, anaplasma, babesia, and bartonella in field voles in their natural habitat. Ecologists are particularly interested in the relationships between the diseases, for example does presence of one disease increase the probability of contracting another. Also of interest is the robustness of individual field voles, for example if an animal is resistant to one disease, will it also be resistant to another? The student will develop new methodologies based on Markov chain Monte Carlo (MCMC), and use them to answer such questions about zoonoses. The project will equip the student with expertise in computationally intensive methods which has a wide range of potential application, while being subject to the constructive discipline of finding practical answers to specific questions. Data and biological advice will be available from the National Zoonosis Centre near Liverpool.
Selection of treatments in phase II clinical trials based on measures of safety and efficacy
Sponsor: EPSRC
Supervisors: Dr Thomas Jaki and Professor J Whitehead
Duration: 3½ years
Stipend: to be determined
Phase II clinical trials are conducted in patients to determine whether to take a candidate drug forward to larger, definitive clinical studies that could lead to the granting of a licence. Often several doses of the candidate drug are available for study. Whitehead and Jaki (Statistics in Medicine 28: 828-847, 2009) describe an approach for the selection the best dose when the primary efficacy response is an ordered categorical score. This project would generalise that approach to allow the simultaneous consideration of both ordinal efficacy responses and binary safety indicators such as the incidence of adverse events. Other forms of efficacy and safety measures, such as those with continuous distributions would then be considered. The proposed methodology would be evaluated using simulation. motivating examples and datasets will be sought from the pharmaceutical industry.
Further information
EPSRC and MRC studentships cover fees and stipends (living expenses) for UK students, but only fees for students from elsewhere in the EU. The AstraZeneca studentship will cover fees and a stipend for students from any EU country. The last of these four studentships will be awarded competitively from amongst students applying for a number of eligible research projects within mathematics and statistics.
For further details of these and other PhD opportunities in Statistics at Lancaster University, visit http://www.maths.lancs.ac.uk/department/research/PhD
To apply, visit http://www.maths.lancs.ac.uk/department/research/PhD/applying mentioning the title and sponsor of the project or projects that you are interested in - additionally, please send an email to [log in to unmask] to say that you have applied.
Closing date: 5 May 2009
Professor John Whitehead
Department of Mathematics and Statistics
Fylde College
Lancaster University
Lancaster LA1 4YF, UK
Tel: 01524 592350
Fax: 01524 592681
[log in to unmask]
|