The Biostatistics and Genetic Epidemiology groups in the Department of Health Sciences at the University of Leicester has a three-year PhD studentship to start in October 2012.
The studentship will cover full fees for UK/EU students, a conference allowance and a maintenance stipend at RCUK rate (currently £13,590pa).
Project Outline: Estimation of absolute risks in large epidemiological cohort studies.
Aims: To develop efficient statistical methods to estimate both absolute and relative risks in outcome-dependent sampling designs in large cohort studies.
Substantial resources have been invested in large cohort studies, which are used to study the relationship between exposures and the occurrence of disease. Cohort studies are used in both traditional and molecular epidemiology, but it can be costly, inefficient and sometimes unfeasible to obtain exposure information on all individuals in the cohort, particularly in molecular epidemiology studies which collect and store biological samples (e.g. UK Biobank).
Nested case-control and case-cohort studies have been proposed for these situations. The key advantage of these studies is a much reduced sample size is required when compared to analysing the full cohort, with only a minimal loss in statistical efficiency.
This project will build on the group's work on flexible parametric survival models by developing methods to simultaneously estimate absolute as well as relative risks in case-cohort and nested case-control studies. There are a number of advantages to using flexible parametric survival models in standard cohort studies including (i) the ease at which non-proportional effects can be incorporated, (ii) estimation of absolute risks, (iii) extension to multiple time scales and (iv) extending to competing risks. These extensions have the potential to be further extended for the analysis of case-cohort studies and nested case control studies.
More attention will be given to case-cohort studies as we believe that there are a number of advantages when compared to nested case-control studies. For example, they (i) are more efficient for analysing several endpoints of interest, (ii) do not have to specify a single time-scale in advance, (iii) can potentially incorporate multiple time-scales and competing risks and (iv) require less complex information on population weights that are needed to estimate absolute risks.
The exact topic is fairly flexible and would be selected based on the successful applicant's interests after discussion with the supervisors. User friendly software will be developed so that that any methodological developments can be applied in practice.
The successful applicant should be highly motivated, have a strong statistical background (preferably an MSc in Medical Statistics or similar) and an interest in developing and applying statistical methodology. The student should be able to demonstrate their commitment to undertaking doctoral study. The studentship will provide an opportunity to develop both advanced methodological and applied research skills.
The studentship will be under the supervision of Dr Paul Lambert with Professor Paul Burton as a second supervisor and the applicant will join a team with an international reputation for both methodological and applied research. The project will involve collaboration with researchers in and outside Leicester, thus ensuring a rich research environment. There will be opportunities to attend internal and external training courses and to present at national and international conferences.
Applicants should send a statement of interest (no more than 1 page) and a CV with details of two referees to: Dr Paul Lambert ([log in to unmask]) by 9th March 2012. Informal enquiries prior to this date are welcome.
Dr Paul C Lambert
Reader in Medical Statistics
Centre for Biostatistics & Genetic Epidemiology
Department of Health Sciences
University of Leicester
2nd Floor, Adrian Building
University Road
Leicester LE1 7RH
UK
Tel: +44 (0)116 229 7265, Fax: +44 (0)116 229 7250
e-mail: [log in to unmask]
Homepage: http://www2.le.ac.uk/Members/pl4/ ________________________________________
You may leave the list at any time by sending the command
SIGNOFF allstat
to [log in to unmask], leaving the subject line blank.
|