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 3 MRC PhD studentships to start
from October 2009. Awards cover Cambridge University fees and a stipend of
at least 12,500 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.

Examples of some of the projects available for PhD study  are given below.
We will also consider suitable projects as suggested by prospective PhD
candidates.

Full research council funding can be awarded to UK citizens and applicants
who satisfy UK residency criteria as specified by the Medical Research
Council. Fees only are provided for citizens of EU countries. Non-EU
applicants who do not satisfy residency criteria are encouraged to apply but
will not be eligible for any MRC funds from the Unit and are requested to
provide their own funding to meet the full costs of a University of
Cambridge PhD. Please note that applicants not meeting eligibility criteria
for full funding will be asked to provide evidence that they have either:

    * applied for alternative funding or
    * will be able to meet the full costs of a University of Cambridge PhD.

All applicants should send CV, letter and contact details of 2 academic
referees to the Postgraduate Administrator at the Biostatistics Unit by 31st
January 2009. It is planned that interviews will be held in February/March
2009. E-mail applications are acceptable to [log in to unmask]

*******************
PROJECTS

Title: Model and parameter uncertainty in health economic modelling.
Joint Supervisors: Linda Sharples and Chris Jackson

When making decisions about cost-effectiveness of different treatments, it
is necessary to construct reasonably realistic models and these are likely
to be complex in most cases. In such models there is often uncertainty about
the most appropriate assumptions.
For example, which clinical states or events should be included in the
model, and how should the occurrence of these events vary through time or
between different people. If the model is very complicated then some aspects
will be estimated very imprecisely and that may produce poor predictions of
the overall costs and effects. However failure to include important features
may result in biased estimates.

This project will consider aspects of uncertainty in health economic models
and specific questions may include:

    * How to allow for uncertainty about the model structure when deciding
the most cost-effective treatment and assessing the need for further research,
    * In what circumstances is it appropriate to base the decision on a
model-averaged result obtained by weighting models according to their
plausibility,
    * How should the plausibility of different models be assessed against data,
    * How to elicit expert opinion and to formalize its use when data are
not available
    * How to deal with missing data in this situation

There is increasing confidence amongst statisticians and economists in the
Bayesian approach to modelling in this situation. Thus we will focus on this
approach, although a strong mathematical background and enthusiasm are more
important than previous experience of working in the Bayesian paradigm.

A range of data sets will be available for testing out methods, including
trials of screening for abdominal aortic aneurysms and of alternative
diagnostic strategies for suspected heart disease, and a cohort study of
patients requiring implantable cardioverter defibrillators.

******************
Title: Extrapolation and inclusion of competing risks in long-term survival
modelling
Joint Supervisors: Linda Jackson and Chris Jackson

Survival analysis is central to many medical studies and we have a battery
of statistical tools for handling many situations. However when there is a
need to estimate mean survival the complete survivor function for a
population is required. This is not always available and we are required to
extrapolate survival beyond that observed in the sample data. Assumptions
about the nature of the extrapolation are often not testable and rely either
on assuming a parametric form for the data or on incorporating data from
other sources, either via prior distributions or using methods for more
general evidence synthesis. Within this context there may be competing
causes of death, and in particular, we may be more confident about
assumptions made for specific hazards than for survival overall. For
example, extrapolating treatment effects for the specific target hazard may
be more plausible than extrapolation in the case of overall survival,
especially if the contribution that each hazard makes to the total hazard
changes over time. Competing risks methods in this context will be studied,
using both parametric and non-parametric survival distributions and with
applications to a number of studies in cardiothoracic surgery.

******************
Title: Cognitive change modelling for individuals in the presence of missing
data.
Supervisor: Fiona Matthews

The investigation of cognitive decline has often been conducted by
examination of change in cognitive abilities at the marginal level, that is,
by aggregating data across individuals over time. However, the process has
been identified to be too complex to be described at this level and models
at the person level are needed to gain full understanding of the loss of
cognitive abilities in old age. A problem faced by researchers when
investigating cognitive decline in old age is that attrition plays a
fundamental role in the occurrence of missing observations as individuals
die or dropout of the study. Unless considered in the models in an adequate
way, estimates of change will not reflect this situation. In these potential
projects, data from two population based longitudinal studies of ageing will
be used to motivate examples and test models.

Two large population based studies the Medical Research Council Cognitive
Function and Ageing Study (CFAS) and the Cambridge City over 75 Cohort Study
are rich resources of longitudinal data on the health of the older
population. The MRC Biostatistics Unit is working with the Department of
Public Health and Primary Care at the University of Cambridge on the
statistical analysis of these data, with the CFAS data archive held within
the Unit.

The aims of the potential project are to assist with the modelling of
cognitive change that can provide individual trajectories, either singlurly
or as part of groups with common shapes. Potential investigations include:

    * Investigation of penalised splines with random coefficients models
within a Bayesian framework.
    * The jointly modelling of person specific curves in this context with
models for the description of possible informative missing data processes.
    * Functional data analysis area will be considered to compare modelling
flexibilities across modelling frameworks.
    * Investigation of classes of individuals with similar decline over time
has identified groups of individuals who experience decline at similar rates
over time.
    * Informing the process of class formation may be disproportionally
driven by the missing data. The aims of this project include an initial
extension of previous models for the estimation of random effects models
with a mixture of distribution for the random effects, the fit of missing
data models allowing for non monotone missing data.
    * Identifying individuals who have reached the beginning of the terminal
process may identify more realistic treatment options and potential for
support. Terminal decline is not a simple process and the impact of
individuals who do not die during the time of the study or who have
accidental deaths are likely to be important processes to model.

The work will involve

    * development of suitable statistical models
    * fitting these models in two population based epidemiological studies
    * simulation studies to explore the properties of the methods
    * interpretation of results in an epidemiological context

These projects will be supervised by Fiona Matthews. The student will gain
both a methodological insight and experience of applied studies as the data
is collected from a wider collaboration within the ageing programme of the
University Department of Public Health and Primary Care.

***********************
Title: Benchmarking measurements of healthy life.
Supervisor: Fiona Matthews

In many disease areas the focus of research has been to concentrate on
reducing mortality from the disease. Now as the population ages the focus is
changing to concentrate on living a longer healthier life. Two projects are
proposed within the methodology of measuring healthy life.

Many methods have been proposed to estimate healthy life. The project
proposed here will use large cohort studies from the UK and US to
investigate the different methods of measuring healthy life. Initial
investigation will use the current methodologies to include methodological
work on how the different methods interrelate, study design and sample size
considerations and how missing data impacts on the estimates obtained. These
methodologies will then be extended to generate new methods for better
estimation in the future. The methods proposed have been formulated based on
different types of longitudinal studies. Some formulations may be more
methodologically sound for different study designs. Missing data may well
differentially impact on some types of study designs. Information for
researchers is required for help on analysis, but more importantly
information on the precision using different study designs.

This project will be supervised by Fiona Matthews, with input from Prof
Carol Jagger University of Leicester. The student will gain both a
methodological insight and experience of applied studies as the data is
collected from a wider collaboration within the ageing programme of the
University Department of Public Health and Primary Care, Cambridge and
within the Epidemiology Department of Leicester.

**********************
Title: Cost-effectiveness of dementia treatment
Joint Supervisors: Fiona Matthews and Sheila Bird

Recent guidelines on the use of cholinesterase inhibitors that only patients
who have moderate AD can be treated with cholinesterase inhibitors on the
NHS has raised many issues. This appraisal (with the independent statistical
analysis undertaken by Fiona Matthews) has highlighted the delicate question
of the importance of proving the cost-effectiveness of a drug in a situation
where no other treatment of the disease itself exists and from which there
is no cure. Current methods for cost-effectiveness for dementia treatment
are based primarily on the prevention of the transition from the community
setting to institutionalised care. However, the cost of dementia is much
higher in terms of informal care costs and the impact on relatives’ health
besides the mortality effects of the disease itself.

Our aims for this project are to develop a more realistic base for measuring
the cost-effectiveness for existing treatments, and the potential for new
treatments, that combine epidemiological data and clinical trial data within
a sound framework that introduces realism into policy decisions. Using the
existing methods and drawing on wider experience within the Unit on evidence
synthesis the epidemiological data from the life course of individuals both
with normal ageing and dementia will be combined with cost information,
utilities, and outcomes from clinical trials (undertaking a class
meta-analysis) to provide a robust framework of the true cost of treatment
for dementia within a population setting. This framework whilst initially
being based on information on the cholinesterase inhibitors currently
licensed can be used to develop models whereby cost-effectiveness can be
examined for future drugs.

This project will be supervised by Dr. Fiona Matthews and Prof Sheila Bird.
The student will gain both a methodological insight and experience of
applied studies.

************************
Title: Cost-effectiveness of dementia treatment
Joint Supervisors: Krista Fischer and Ian White

Randomized Controlled Trials (RCTs) are the most common research method used
nowadays to provide evidence about efficiency of treatments. Patients
failing to take their allocated treatment (non-compliance) could bias
results, compared with results given perfect compliance. When there is
compliance data available, it is often of interest to assess the effect of
compliance on the trial outcomes. Simple regression methods would in general
not allow proper adjustment for compliance, since compliance is a
post-randomization variable and may be influenced by the same treatment- or
disease-related factors (possibly unknown confounders) that are influencing
the outcome.

Often the compliance and outcome data is available in the form of repeated
measurements over time. Structural Nested Mean Modeling provides a general
theoretical framework for analyzing such trials properly. The task of this
project include:

    * develop an estimating algorithm for such models that could be easily
implemented in practice
    * careful study of the assumptions involved and their practical implications
    * investigation of the properties of the resulting estimator (possibly
in comparison with alternative methods) and develop of some simple tools for
model-checking
    * possible extensions of the methodology to
                  a)analyze trials that compare two or more active
treatments (such as equivalence trials) 
                  b) analyze trials with binary outcomes
    * apply the methodology in the analysis of randomized trials (data sets
in hypertension, alcohol, mental health and child health are available)

***********************
Title: Characterising patterns of injecting drug use
Joint Supervisors: Daniela De Angelis, Shaun Seaman (MRC BSU), Matthew
Hickman (University of Bristol)

Injecting drug use (IDU) is a chronic relapsing disorder that causes
substantial health and social harm. The risk of fatal overdose is about 1%
per year worldwide; 85% of the prevalent 200,000 hepatitis C infections in
the UK are due to IDU; and the annual cost of drug related crime by heroin
and crack users is over £12 billion in the UK. Knowledge of pattern and
duration of injecting drug use is critical in order to inform effective
public health action and drug policy. It is known that the potential for
drug dependence is high and that after progression to dependence injecting
activity can be long lasting often with multiple periods of relapse and
recovery. However, up to now few attempts have been made to quantify the
injecting duration producing very imprecise estimates and still failing to
capture its characteristic periodicity. Questions of the type: “How long do
injecting drug users continue injecting? How many times do they relapse
before permanently abstaining from injecting? What factors influence
abstinence and relapse?” are still unanswered. One of the reasons is that
data for investigating these questions are not easily available and, if
available, are subject to various biases.

This project will involve developing and applying statistical methodology to
investigate the above questions, taking into account of the biases affecting
longitudinal data collected from cohorts of injecting drug users.

***********************
Title: Bayesian hierarchical modelling and simulation of inter- and
intra-individual variation in insulin-glucose kinetics
Supervisor: David Lunn

Type 1 diabetes is characterised by an absolute insulin deficiency. The gold
standard treatment by multiple daily injections or subcutaneous insulin
infusion is suboptimal. Closed-loop insulin delivery also known as the
artificial pancreas promises to revolutionise the treatment of T1D
(www.jdrf.org/artificialpancreas). A key component of an accelerated
development of the artificial pancreas is an in silico simulation
environment capable of predicting the interaction between a glucose
controller and subjects with T1D in short-term (hours to a day) and
longitudinal studies (weeks to months). Inter- and specifically
intra-subject variability of insulin-glucose kinetics in T1D is poorly
understood. This project will use stochastic modelling within the Bayesian
hierarchical modelling framework to develop a virtual population of subjects
with T1D. Data collected at home will be combined with data collected in
clinical research studies to obtain time invariant and time variant model
parameters. Different models will be developed and rational model-selection
approaches will be used. The virtual population may represent several
subgroups including children, adults, and pregnancies with T1D and will be
used to simulate the interactions with different closed-loop insulin
delivery systems differing in the glucose controller, insulin type (rapid vs
very rapid), insulin delivery characteristics (microneedle, dermal) etc.

***************************
Title: Relating genetics of disease to biological processes
Joint Supervisors: Frank Dudbridge and Lorenz Wernisch

Following the human genome sequencing project there has been a great deal of
interest in finding genetic causes of diseases such as coronary artery
disease, schizophrenia and osteoporosis. This relies heavily on statistical
methods because for these common diseases, genetics affects the risk only
probabilistically. Typically, not one but several genes are needed to
increase disease risk, and they may interact to do so. There is a rapidly
growing body of knowledge on how networks of genes govern basic cellular
processes, and this is just starting to be related to disease processes. In
the project, we will explore methods for combining evidence from genetic
case control studies with models of gene and protein networks. This will
allow us to extract more information from case control studies as well as
improving our knowledge of how disease develops.