PhD STUDENSHIPS
                           
                    MRC BIOSTATISTICS UNIT CAMBRIDGE

The BSU is an internationally recognised research unit specialising in 
statistical modelling with application to medical or public health sciences. 
Further details of the work carried out in the Unit appear on the website 
http://www.mrc-bsu.cam.ac.uk. The Unit has 3 MRC PhD studenships to start from 
October 2003. The awards cover Cambridge university fees and a stipend of at 
least 10,000 pounds for a period of 3 years. Applicants must have or expect to 
get a first or high 2.1 honours degree in mathematics, statistics or a related 
discipline. A masters degree is desirable but not essential. 

Examples of some of the projects available for PhD study are given below. Other 
student-proposed projects will be considered.

Full research council funding can be awarded to UK citizens, fees only are 
payable for citizens of European Community countries. It is not possible to 
award studentships to citizens of countries outside Europe. Non-UK citizens with 
their own source of funding are encouraged to apply.

Interested students should send CV, letter and contact details of 2 referees to 
Linda Sharples at the Biostatistics Unit by 15th March 2003. E-mail applications 
are acceptable.

PROJECTS

Title: Statistical methods for comparative genomics.

Supervisor: Wally Gilks

Bioinformatics is the relatively new science of developing methodology 
for the exploitation of rapidly expanding genomic and related databases.
Comparative genomics is that part of bioinformatics which aims to extract
biological information by comparing genomes of related organisms. This field
is gaining prominence with the increasing availability of completely 
sequenced genomes, in particular of higher organisms. Statistical PhD projects 
within this broad remit will involve working closely with biologists.



Title: General methodological work in medical statistics based on
substantive applications. 

Supervisor: Vern Farewell

Applying standard statistical methods to medical and biological systems is 
problematics. For example, analysis of longitudinal data and  non-standard study 
designs are common. Development of general methodology to address these problems 
is of interest. An area of particular applied interest is rheumatology, but 
other areas can equally be used to motivate the development of new methodology.


Title: The analysis of health economic data in evaluating medical treatments

Supervisor: Simon Thompson

Studies of new medical treatments have traditionally been based on demonstrating 
clinical effectiveness, in terms of improved survival or improved quality of 
life, compared with the standard treatment or a placebo.  Since health care 
budgets are limited and most treatments are only partially effective, it is 
necessary to quantify the trade-off between effectiveness and the costs of 
treatment, in a cost-effectiveness analysis.  Medical care resource-use data are 
therefore increasingly being collected as part of both clinical trials and 
observational studies that evaluate medical treatments.  Data collected for 
cost-effectiveness analyses give rise to a number of statistical difficulties, 
which need to be resolved in order to put health economic assessment onto a firm 
scientific basis.

These statistical issues include:
·	Handling the variable follow-up time of different patients
·	Methods of adjusting costs for case-mix, as represented by initial 
patient characteristics
·	Derivation of net benefit and cost-effectiveness acceptability curves, 
based on censored data with adjustment for case-mix
·	Use of intermittent quality of life assessments in the calculation of 
quality-adjusted survival
·	Use of intermittent data on medications to model their costs over time
·	Approaches to analysis when components of the resource-use data are 
missing, or patients are lost to follow-up

The project will be based primarily around issues arising in the analysis of a 
large observational study.  4000 patients undergoing investigation for coronary 
artery disease have been followed up for an average of 3 years (currently being 
extended to 6 years) for hospitalisations, operations, medications, mortality, 
morbidity, and quality of life.  The purpose is to investigate the relative cost 
and effectiveness of different initial medical policies, such as bypass surgery, 
angioplasty, and medical treatment.  There will also be the opportunity to 
explore other data sets, including a large trial of screening for aortic 
aneurysms and an international study of costs of care and mortality after acute 
stroke.

The project will be supervised by Prof Simon Thompson, and Dr Linda Sharples and 
Dr Richard Nixon will act as additional advisors.  Collaboration with the 
coronary disease study team will be developed, including Dr Harry Hemingway 
(epidemiologist), Dr Julie Barber (statistician) and Dr Mark Sculpher (health 
economist).


Title: Regaining power lost by non-receipt of randomised interventions

Supervisor: Ian White
 
Ideally, all patients randomised to a particular intervention in a clinical 
trial would receive that intervention as described in the trial protocol. In 
practice, however, some patients will not receive the randomised intervention at 
all, will receive it only for a short period, or will receive other 
interventions. In such cases it is conventional to analyse by 
intention-to-treat, that is, by comparing patient groups as randomised and 
ignoring actual treatment. However, there is inevitably a loss of power due to 
the non-receipt of the randomised intervention. Some authors attempt to regain 
some of the lost power by excluding patients who do not receive their randomised 
intervention, but this is subject to selection bias because remaining patients 
are unlikely to be representative.
This project will explore whether it is possible to regain the lost power 
without excluding non-receivers. Specific methods to be considered are
1.	use of prior information about the degree of similarity between 
receivers and non-receivers;
2.	drawing strength from other trials with different intervention effects 
but similar patterns of non-receipt;
3.	constructing a joint model for receipt of intervention and outcome.
The project will be partly based around the analysis of two randomised trials. 
One trial explored the effect of training psychiatric staff in the management of 
psychiatric patients who also misuse drugs or alcohol. Staff turnover was 
substantial, and trained keyworkers could be replaced by untrained ones, or vice 
versa. The other trial evaluated the effect of insertion of ventiliation tubes 
in children with hearing problems. Some children required insertion or 
re-insertion of ventiliation tubes during follow-up regardless of their 
randomised allocation. 
The project will be supervised by Ian White, and David Spiegelhalter will act as 
an advisor.


_______________________________________________________________________________

Dr. Linda D. Sharples
MRC Biostatistics Unit			R&D Unit	
Institute of Public Health		Papworth Hospital NHS Trust	
Robinson Way				Papworth Everard	
Cambridge				Cambridge	
CB2 2SR					CB3 8RE	

Tel: 01223 330389			01480 364445	
FAX: 01223 330388			01480 831450	

e-mail: 				 
[log in to unmask]	[log in to unmask]

________________________________________________________________________________