Please find attached an email advertising the PhD projects list
available from October 2010 at the MRC Biostatistics Unit in Cambridge.
I'd be very grateful if you could forward this on to those students who
are finishing their studies and may be interested.
Many thanks and Best wishes,
Rosemary Camperos
MRC BSU PhD Programme Administrator
*
PhD STUDENTSHIPS, MRC BIOSTATISTICS UNIT, CAMBRIDGE *
The BSU is an internationally recognised research unit specialising in
statistical modelling with application to medical, biological or public
health sciences. Details of the work carried out in the Unit appear on
the website www.mrc-bsu.cam.ac.uk . The Unit has 4 MRC PhD studentships
to start from October 2011. Applicants interested in commencing in
January (Lent term ) or April (Easter term ) are more than welcome to
send an email asking for further details to [log in to unmask]
<mailto:[log in to unmask]>
Awards cover Cambridge University fees and a stipend of at least 13,500
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.
The projects list available for PhD study are given below. We will also
consider suitable projects as suggested by prospective candidates.
Full research council funding can be awarded to UK citizens and
applicants who satisfy UK residency criteria, fees only are payable for
citizens of EU 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.
Applicants should send CV, letter and contact details of 2 academic
referees to the Postgraduate Administrator at the Biostatistics Unit by
January 2011. E-mail applications are acceptable to
[log in to unmask]
*Application deadlines*
Our list of possible projects is usually advertised in October and the
closing date for applications to the Unit is in January of the following
year. Interviews are usually held in early February.
Applicants interested in commencing in January (Lent term ) or April
2011 (Easter term ) are more than welcome to send an email asking for
further details to [log in to unmask]
Students wishing to apply for funding from the University should check
the University's website
<http://www.admin.cam.ac.uk/offices/gradstud/fees/funding/index.html>
for details of the closing date for these.
Projects List
* *Incorporating external evidence on heterogeneity and bias in
meta-analysis
<http://www.mrc-bsu.cam.ac.uk/Education/PhD.html#project1>* -
Rebecca Turner <http://www.mrc-bsu.cam.ac.uk/People/RTurner.html>
and Julian Higgins <http://www.mrc-bsu.cam.ac.uk/People/JHiggins.html>
* *Assessment of the impact of learning curves, multiple operators
and non-proportion hazards in clinical trials of surgical
procedures and devices
<http://www.mrc-bsu.cam.ac.uk/Education/PhD.html#Project2>* -
Linda Sharples
<http://www.mrc-bsu.cam.ac.uk/People/LSharples.html> and Adrian
Mander <http://www.mrc-bsu.cam.ac.uk/People/AMander.html>
* *Multi-state survival modelling with time-dependent covariates:
estimation and prediction
<http://www.mrc-bsu.cam.ac.uk/Education/PhD.html#Project3>*- Ardo
van den Hout
<http://www.mrc-bsu.cam.ac.uk/People/AVandenHout.html> and Fiona
Matthews <http://www.mrc-bsu.cam.ac.uk/People/FMatthews.html>
* *Longitudinal modelling of HIV markers to estimate time of
infection and HIV incidence in the general population
<http://www.mrc-bsu.cam.ac.uk/Education/PhD.html#Project4>*- Brian
Tom <http://www.mrc-bsu.cam.ac.uk/People/BTom.html> and Daniela de
Angelis <http://www.mrc-bsu.cam.ac.uk/People/DDeAngelis.html>
* *Misspecification in multi-state models and its impact in the
longitudinal analysis of quality of life measures
<http://www.mrc-bsu.cam.ac.uk/Education/PhD.html#Project5>* - Vern
Farewell <http://www.mrc-bsu.cam.ac.uk/People/VFarewell.html>
* *Statistical modelling of the dynamics of gene regulation
<http://www.mrc-bsu.cam.ac.uk/Education/PhD.html#Project6>*-
Lorenz Wernisch <http://www.mrc-bsu.cam.ac.uk/People/LWernisch.html>
Projects Description
**
*Incorporating external evidence on heterogeneity and bias in
meta-analysis*
Supervisors: Rebecca Turner
<http://www.mrc-bsu.cam.ac.uk/People/RTurner.html> and Julian
Higgins <http://www.mrc-bsu.cam.ac.uk/People/JHiggins.html>
In a meta-analysis, the results from a set of similar studies are
combined, in order that the evidence available on a particular research
topic can be summarised. Meta-analyses are increasingly widely used in
all areas of public health research, and policy decisions taken by
organisations such as NICE rely on the results reported in relevant
meta-analyses. The combined results are reported with much greater
precision than the original studies' results, and are consequently more
influential. It is therefore essential that meta-analyses are carried
out to a high standard.
Many meta-analyses combine the results from a small number of studies
(e.g. less than 5). This is problematic in that the amount of variation
among studies, known as heterogeneity, is imprecisely estimated. An
additional problem is that the original studies are often affected by
varying amounts of internal bias caused by methodological flaws.
Standard methods for meta-analysis do not acknowledge biases in the
studies, and do not allow for imprecision in the estimated between-study
variance. When standard methods are applied in the presence of
heterogeneity and bias, there is a danger that the meta-analysis results
will be seriously flawed.
In recent work carried out within the *MRC Biostatistics Unit*, all
meta-analyses within the Cochrane Database of Systematic Reviews (CDSR)
<http://www.cochrane.org/cochrane-reviews>(CDSR) have been classified
according to type of interventions evaluated (for example,
pharmacological, surgical etc.) and type of outcome (for example,
mortality, mental health etc.). The classified database contains over
22000 meta-analyses and provides a rich resource of information on the
relationship between degree of between-study heterogeneity and
meta-analysis characteristics. These data can be used to inform future
meta-analyses carried out in a wide variety of research settings. In
addition, the Unit has access to recently acquired information on the
biases expected within particular types of studies.
The aims of this project are to develop statistical methods for
meta-analysis in the presence of heterogeneity and bias. Possible issues
to be tackled include:
* Investigate how external information on heterogeneity should be
incorporated in meta-analysis. This may require development of
further informative prior distributions for use in Bayesian
meta-analysis.
* Investigate how external information on bias should be
incorporated in meta-analysis, and explore the merits of using a
combination of empirical evidence and expert opinion on bias.
* Assess the potential gains from incorporating external information
and explore how these would vary across different settings.
* Consider the design implications for future meta-analyses which
could make use of newly available information on heterogeneity and
bias. How should researchers plan their meta-analyses when
heterogeneity and bias are expected?
There will be potential for collaboration with Jonathan Sterne's
research group at the University of Bristol
<http://www.epi.bris.ac.uk/staff/jsterne.htm>.
*NOTE: *Applicants interested in commencing in January (Lent term ) or
April 2011 (Easter term ) are more than welcome to send an email asking
for further details to [log in to unmask]
********************************************************
*Assessment of the impact of learning curves, multiple operators
and non-proportion hazards in clinical trials of surgical
procedures and devices***
*Supervisors: Linda Sharples
<http://www.mrc-bsu.cam.ac.uk/People/LSharples.html> and Adrian
Mander <http://www.mrc-bsu.cam.ac.uk/People/AMander.html>*
Trials of surgical procedures and devices are more complicated than
trials of medicines since these involve multiple features and require
surgeons to reach a minimum standard of competence before they can
contribute to a trial. In addition they trade off a short term risk for
longer term improvements in survival and quality of life, so that
survival hazards are non-proportional. This has impact on both stopping
rules at interim analysis and final trial analysis. During this applied
statistics PhD we plan to investigate:
*
how to incorporate training cases into design and, if there is
some residual concern about training, the trial analysis
*
how to design efficiently, including sample size calculations, if
survival is the endpoint and there are non-proportion hazards.
*
how to monitor trials and create stopping rules if survival is the
endpoint and there are non-proportion hazards.
*
incorporation of surgeon and centre effects in multicentre
surgical trials and how important are they in some case studies
and in general.
A range of trials in cardiac surgery will be available for this study.
*NOTE:* Applicants interested in commencing in January (Lent term ) or
April (Easter term ) 2011 are more than welcome to send an email asking
for further details to [log in to unmask]
********************************************************
*Multi-state survival modelling with time-dependent covariates:
estimation and prediction*
Supervisors: Ardo van den Hout
<http://www.mrc-bsu.cam.ac.uk/People/AVandenHout.html> and Fiona
Matthews <http://www.mrc-bsu.cam.ac.uk/People/FMatthews.html>
Multi-state survival models are used in medical research to investigate
and predict health-related processes over time. Applications are, for
instance, stages of recovery after an operation, stages in the
development of AIDS, and stages of disability in older age. In a
continuous-time multi-state model, the risk of a transition to a next
stage can be linked to covariates (risk factors) such as age, sex, and
smoking behaviour.
In a three-state model for disability in older age (healthy state,
disability state, and death state), a typical research question is
"Which risk factors play a role with regard to a transition from the
healthy state to the disability state?". A question that has to do with
prediction is "Given no disability at age 65, what is the total life
expectancy, and which part of that is expected to be disability-free?".
Statistical methodology for multi-state models is available and can be
seen as an extension of standard survival analysis. An often used
assumption is that the multi-state stochastic process is first-order
Markovian. Such an assumption makes the models relatively easy to
estimate. Extensions to Semi-Markov models can also be found in the
literature. When estimating the model, it is possible to take into
account covariates that change over time. However, if the change of such
a time-dependent covariate is a stochastic process (for example,
changing smoking behaviour), then the prediction of the multi-state
process is not straightforward. The multi-state process is dependent on
the stochastic process of the covariate and joint modelling of the
processes is required.
The project will start with methodology for the estimation of
continuous-time multi-state models. Time-dependent covariates will be
taken into account and the parallels with survival analysis will be
explored. Next, prediction and the joint modelling will be undertaken.
Possible extensions of the methodology may include semi-Markov models
and/or multi-level models that take unobserved heterogeneity into
account. The models will be applied to data from the Cognitive Function
and Ageing Studies <http://www.cfas.ac.uk>to predict changes associated
with ageing.
Population ageing is of growing interest, and this work contributes by
improving methodology for predictions for the future.
********************************************************
*Longitudinal modelling of HIV markers to estimate time of
infection and HIV incidence in the general population*
Supervisors: Brian Tom
<http://www.mrc-bsu.cam.ac.uk/People/BTom.html> and Daniela de
Angelis <http://www.mrc-bsu.cam.ac.uk/People/DDeAngelis.html>
In the late 1990's the idea of identifying new HIV infections using
characteristics of the antibody response following infection was
introduced. Since then a number of antibody biomarkers have been
developed to distinguish between recent and established HIV infection.
Typically a specific threshold/cut-off of the biomarker is chosen,
values below which are indicative of recent infections.
Such biomarkers have attracted considerable interest as the basis for
incidence estimation using a cross-sectional sample. An estimate of HIV
incidence can be obtained from the prevalence of recent infection, as
measured in the sample, and knowledge of the time spent in the recent
infection state, known as the window period. This idea has been used to
estimate incidence rates in both counselling and testing sites and in
sexually transmitted disease clinics.
More recently a biomarker has been proposed based on the principle that
antibodies produced early after infection bind less strongly to the
antigen than those produced in established infection. The avidity of the
antibodies to bind to the antigen can be measured using the Avidity
Index (AI). Conditionally on the choice of a specific threshold,
individuals with measured AI below the threshold are classified as
recently infected and the window period is now the time spent below the
chosen threshold.
However, a number of challenges presently exist when adopting this
strategy to HIV incidence. Firstly, the initial estimation method
suggested by Janssen et al. (1988) assumes that the testing process,
which leads to an individual presenting for an HIV test, is not
associated with the risk of infection. Secondly, issues exist regarding
how estimates of HIV rates derived from a particular sub-population can
be extended to the more general population of interest. Thirdly, it is
still unclear how to perform the estimation when only data from
diagnosed individuals, /who have chosen to test/, are available.
Research into addressing these challenges is being actively pursued.
However the current approaches taken currently have been rather
informal. For this project we propose to tackle this problem of HIV
incidence using HIV markers in a more rigorous manner. We envisage using
a Bayesian framework to tackle this problem, whereby longitudinal
modelling techniques and statistical methods that account for sampling
biases are employed.
*Collaborators:* Health Protection Agency HIV/STI Department
<http://www.hpa.org.uk/Topics/InfectiousDiseases/InfectionsAZ/HIVAndSTIs/>
*
*
********************************************************
*Misspecification in multi-state models and its impact in the
longitudinal analysis of quality of life measures*
Supervisor: Vern Farewell
<http://www.mrc-bsu.cam.ac.uk/People/VFarewell.html>
Multi-state models have proved very useful for the analysis of
longitudinal data that arise in medical studies. In this project we
would like to assess broadly the potential effect of misspecification of
such models, initially restricted to Markov models. The work is
motivated by use of the models in various rheumatological applications,
and will be applicable quite generally, but the practical application of
the results will focus on the longitudinal analysis of quality of life
measures in rheumatology, particularly in patients with psoriatic arthritis.
This project is envisaged to encompass, in broadly equal measure, both
technical methodologic investigations and substantive medical applications.
********************************************************
*Statistical modelling of the dynamics of gene regulation*
Supervisor: Lorenz Wernisch
<http://www.mrc-bsu.cam.ac.uk/People/LWernisch.html>
Biological regulatory processes are very complex and can rarely be
completely represented in a mathematical dynamical model. If
simplifications are unavoidable, how can different models be developed
and compared? Bayesian inference for parameters and models provides a
sophisticated framework for addressing such questions. Time series data
from gene expression experiments and flow cytometry experiments are the
basis for inference of dynamical models. If standard functions and
distributions are too limited to represent complex dynamics,
nonparametric methods, which have gained in popularity in statistics
recently, might be able to provide a well controlled and smooth set of
possible models.
****************************************************************
*For further information please visit our website*
*http://www.mrc-bsu.cam.ac.uk/Education/PhD.html*
--
Rosemary Camperos
MRC Biostatistics Unit
Institute of Public Health
University Forvie Site
Robinson Way
Cambridge CB2 0SR
Tel No.: +44 1223 330376
Fax: +44 1223 330388
Email:[log in to unmask]
Website:http://www.mrc-bsu.cam.ac.uk
You may leave the list at any time by sending the command
SIGNOFF allstat
to [log in to unmask], leaving the subject line blank.
|