Dear all
I'd be grateful if you could please bring the following PhD studentship opportunity to the attention of anyone you think would be suitable.
Many thanks, Andrew
Dr Andrew Copas
Statistical Methodology Group
MRC Clinical Trials Unit
[log in to unmask]
+44 20 7670 4888
PhD Studentship: Medical Statistics
An MRC-funded PhD studentship is available from the MRC Clinical Trials Unit for study on the following project:
Efficacy Analysis for Trials Subject to Non-Compliance and Drop-Out
Funding:
The award covers fees and a tax-free maintenance stipend of £14,350
Eligibility:
Following MRC restrictions, full studentships are available only to UK applicants. Other EU applicants may be eligible for a "fees-only" award.
The minimum academic requirement is a 2(i) or 1st class BSc degree, or an MSc, in mathematics or statistics, which should either already be completed or be expected before this studentship begins in late September 2007.
Deadline for applications: 26 April 2007
Academic registration:
The student will register for full-time MPhil/PhD study at the Department of Statistical Science, University College London
Supervisors:
Dr Andrew Copas (CTU), Dr Sarah Walker (CTU), & Dr James Carpenter (LSHTM)
Project summary:
In a longitudinal randomised controlled trial (RCT) to compare drugs analysis is often complicated by non-compliance with randomised treatment, and/or drop-out. In a RCT, intention-to-treat (ITT) provides unbiased inference about the effectiveness of the treatment policy following the randomisation: it does not directly estimate efficacy. Efficacy is of interest to both to regulators and patients themselves, who may want to know what is the likely outcome if they are able to take a drug long-term (efficacy), as well as understanding the likely effect of starting treatment with a drug (effectiveness).
For an analysis of efficacy three major approaches have been taken:
1. Per-protocol analysis restricted to patients who stayed on randomised treatment
2. Analysis of a redefined endpoint, in which change from randomised treatment is treated as a poor outcome
3. Causal analysis approaches
A causal analysis is likely to provide the most accurate inference regarding efficacy, but requires complex statistical methods and may make a number of strong assumptions. The alternative methods are simple to implement, but their interpretation and relationship to each other and to underlying causal effects are difficult to judge, and have not been systematically evaluated.
This project is designed to investigate the various approaches proposed for efficacy analysis, and provide practical recommendations for statisticians. The work will be guided and illustrated through application to large RCTs based at the MRC Clinical Trials Unit.
Applications:
Please contact Dr Andrew Copas, e-mail: [log in to unmask] <mailto:[log in to unmask]> , with any enquiries, information on how to apply, and for further details of the project.
|