IMPERIAL COLLEGE OF SCIENCE, TECHNOLOGY AND MEDICINE
DIVISION OF PRIMARY CARE AND POPULATION HEALTH SCIENCES
MRC AND ESRC PHD STUDENTSHIPS & COLLABORATIVE STUDENTSHIPS IN EPIDEMIOLOGY,
BIOSTATISTICS AND SOCIAL STATISTICS
We are seeking enthusiastic and highly motivated individuals to join a
prestigious and stimulating research environment centred on the Departments
of Epidemiology & Public Health and Medical Statistics & Evaluation. The
departments’ focus includes leading-edge work in epidemiology, spatial
statistics and Bayesian statistical methods, medical statistics, statistical
computing, clinical trials and meta-analysis.
Details of possible projects follow, and candidates are encouraged to
present alternative proposals for consideration.
MRC: TWO RESEARCH STUDENTSHIPS, AND ONE COLLABORATIVE STUDENTSHIP
Possible projects include:
· Space-time interactions between cardio-respiratory mortality and air
pollution across London
· Heterogeneity in couple fertility: the use of frailty models
· Modelling the population transmission of gonococcal infection in the
population
· Bayesian methods for modelling variation in health indicators
· Transformation in the analysis of hierarchical medical data, with focus on
fetal monitoring
· Errors-in-variables modelling in nutritional epidemiology
· Development of exposure modelling techniques for epidemiological studies
considering constituents of drinking water
· Background incidence of gastro-intestinal tract disease and water-borne
micro-organisms
ESRC: ONE RESEARCH, AND ONE COLLABORATIVE STUDENTSHIP
Possible projects include:
· investigating the link between deprivation and spatio-temporal variability
in risk
· modelling health demand/need/supply at small-area scale
The studentships will commence in October, 1999.
INFORMAL ENQUIRIES may be made to Professor Paul Elliott, Epidemiology &
Public Health (email: [log in to unmask], telephone 0171 594 3328),
Professor Simon Thompson, Medical Statistics & Evaluation (email:
[log in to unmask], telephone: 0181 383 1572), Professor Patrick Royston,
Medical Statistics & Evaluation ([log in to unmask], telephone: 0181 383
8425), Dr Jon Wakefield, Epidemiology & Public Health
([log in to unmask], telephone: 0171 594 3336).
APPLICATIONS: please send a letter and CV, including the names of two
referees, to Mr Simon Sheffield, Departmental Manager, Department of
Epidemiology & Public Health, Imperial College School of Medicine, Norfolk
Place, London W2 1PG (email: [log in to unmask], fax: 0171 402 2150,
telephone: 0171-594 3317). Closing date: 31 March, 1999.
PROJECT DESCRIPTIONS
1. SPACE-TIME INTERACTIONS BETWEEN CARDIO-RESPIRATORY MORTALITY AND AIR
POLLUTION ACROSS LONDON (P ELLIOTT)
A correlation between daily variations in outdoor air pollution and daily
mortality from cardio-respiratory disease has been consistently reported for
different cities in the US and Europe (including London), although it is
unclear to what extent, if any, these findings are transferable to the
effects of chronic exposure to ambient air pollution. US studies have also
reported higher average mortality for cities with higher average levels of
air pollution, but interpretation is complicated by the broad-scale
ecological nature of the data.
This project will explore both temporal and small-area variations in
cardio-respiratory mortality (from the national postcoded mortality dataset
held by the Department), meteorological data, and monitored levels of air
pollution across London (from the South East Institute of Public Health).
Data are available for around 30 monitoring sites: mainly SOx and NOx, but
also PM10 at eight sites. The temporal and spatial components of variability
in the health and the pollution data will be explored and quantified.
Possible interactions, e.g., higher peaks in daily mortality in high
compared with low pollution areas, and the possible effects of areal
deprivation measures (calculated from census statistics) will be explored.
Appropriate methods, based on Bayesian statistics, will be used to analyse
and combine the health and pollution data, making allowance for the
correlated nature of the errors in both space and time.
2. HETEROGENEITY IN COUPLE FERTILITY: THE USE OF FRAILTY MODELS (M JOFFE)
The fertility of a variety of populations has been characterised, using as a
measure the time taken by a couple to conceive (Time To Pregnancy, TTP), and
a remarkably large degree of between-couple heterogeneity has been found. On
evolutionary grounds, it is difficult to explain why a substantial
proportion of the population, with no obvious disease or other health
problem, have low per-cycle probability of conception. Furthermore, the
semen quality of men is remarkably poor when compared with that of other
mammalian species. It is unclear whether these phenomena are linked, and if
so, whether they are of recent historical origin, and/or whether there are
major differences between different populations e.g. according to genetic or
nutritional differences. The possibility of a fall in the sperm count, which
appears to have occurred in certain places (e.g. Paris, Gent), needs to be
seen in this context.
This project will explore between- and within-couple fertility in a large
survey, the National Child Development Study, which is representative of the
population born in Britain in 1958. Respondents who were interviewed at age
33 provided values of TTP for almost all non-accidental pregnancies (91
percent of female and 84 percent of male respondents, N=3132 and 2576,
respectively).
These data will be analysed using frailty models, which introduce a
couple-specific element (a random effect) to acknowledge between-couple
differences. The distribution of these random effects may then depend on
characteristics of the couple. The effect of these characteristics on
relative fertility will be explored and quantified, and implications for
models of fertility in the population will be assessed.
3. MODELLING THE POPULATION TRANSMISSION OF GONOCOCCAL INFECTION IN THE
POPULATION (H WARD)
Objectives: To develop a model for the transmission of gonorrhoea in the
population based on analysis of biological and social data
Methodology: Neisseria gonorrhoeae, the causative organism for gonococcal
infection, can now be typed using various genetic techniques. This permits
fine classification of the organism which can be used to study transmission
of the organism in a population over time, testing and extending existing
epidemiological models for sexually transmitted infections. Epidemiological
analysis of such data require the development of statistical methods for
validating results, and for comparing potential routes of transmission
suggested by different sources of data (e.g. social compared with genetic).
This project would be based first on the detailed analysis of results from a
historical social and microbiological data set, and then applying
statistical and epidemiological techniques to the results of further
microbiological work, including sequence data, in collaboration with the
Department of Microbiology, Imperial College School of Medicine (led by Dr C
Ison). The data set includes social and epidemiological information and
microbiological strains for over 500 cases, including some large
transmission clusters and networks. We intend to investigate the
relationship between epidemiologically-linked individuals and
microbiological strain type. This may be carried out using loglinear models,
taking into account the sampling methods for collecting the data, and
likelihood of incomplete data (including non-disclosure of some sexual
partners and inability to trace some linked individuals). The demographic
make-up of the sexual network which gives rise to the data will also be
analysed using categorical data methods. The aim of the project is to
understand how gonorrhoeae is transmitted thus providing a means of
controlling it's spread.
4. BAYESIAN METHODS FOR MODELLING VARIATION IN HEALTH INDICATORS (N BEST)
Health ‘indicators’ based on hospital admissions data vary inherently over
space and time. The interpretation of such variation is complex, and may
reflect: (a) associations between the outcome of interest and measurable
explanatory factors/confounders such as age, sex, socio-demographic factors
and hospital effects; the latter may arise due to differences in e.g. number
of beds, speciality, coding practice, data quality and completeness. Such
relationships may involve non-linear, possibly discontinuous associations
and interactions between the variables. (b) variation induced by dependence
of the outcome on unobserved or unmeasured factors, such as an unknown
environmental pollutant. These factors will typically vary smoothly in space
and time, thus inducing spatial/temporal correlation in the observed
outcome. (c) Residual chance variation.
Recent developments in Bayesian computational techniques (specifically
Markov chain Monte Carlo methods) have allowed the realistic modelling of
complex problems. However, there has been limited serious use of these
methods to answer questions of substantive importance. This project thus
aims to develop and extend such Bayesian methods to: a) realistically model
the relationship between health indicators and relevant measured predictors;
b) provide flexible models for capturing spatial/temporal correlation in
such data; c) provide techniques for model criticism/selection which will
enable genuine explanatory associations to be distinguished from spurious
variation arising by chance.
The methodology will be applied to a range of health indicators derived from
relevant post-coded databases held by the department, including the Hospital
Episode Statistics.
5. TRANSFORMATION IN THE ANALYSIS OF HIERARCHICAL MEDICAL DATA, WITH A
FOCUS ON FETAL MONITORING (P ROYSTON)
Datasets with a hierarchical or multilevel structure are increasingly
important in medicine. Examples include growth curves, cluster-randomised
trials, multiperiod clinical studies and observational studies with repeated
measures. The multilevel random-effects model is the analytical tool of
choice. When one or more predictors are continuous, appropriate regression
models are needed. Polynomials are almost invariably chosen, but they are
often inadequate. The project will explore the use of fractional polynomials
in multilevel modelling. These involve transformations of the predictors and
offer greater flexibility and parsimony than ordinary polynomials. Issues
such as how to detect and deal with heterogeneous curve shapes will be
explored. Secondly, for continuous outcome variables, the multilevel model
assumes a Gaussian distribution for relevant parameters. Transformation of
the response variable may be needed to satisfy this condition. However
transformation affects all aspects of the model, including the shapes of the
response curves and their heterogeneity and the distribution of quantities
at all levels of the hierarchy. The project will investigate the effects of
response transformation on different parts of the model. The aim will be to
develop techniques which will help the analyst decide whether and how to
transform the response and understand the effects thereof. A particular
application is the analysis of longitudinal fetal size data to produce
`conditional reference intervals', which are intended to help the clinician
detect fetuses whose growth is faltering. Transformation of predictor and
response variables is needed here. The project will also consider how best
to present the predictions from such models for ease of understanding and
use by the clinician. Several datasets are available to the project.
6. ERRORS-IN-VARIABLES MODELLING IN NUTRITIONAL EPIDEMIOLOGY (J WAKEFIELD)
There is strong evidence that diet is an important factor in the aetiology
of many diseases, including cardiovascular disease and several cancers.
However, the study of such relationships is severely hampered by the
difficulty in obtaining reliable measures of dietary intake in large
population samples, and in deriving models of lifetime exposure. In
particular, the large component of within-person variability and measurement
error in dietary and nutritional variables leads to underestimation of
effect (for example, in relation to blood pressure or coronary heart
disease), the so-called 'regression-dilution' problem.
While life-long intake may be the gold standard, in practice data are only
available via food-frequency questionnaires, dietary records, urinary
excretion, etc., and usually on only one occasion. Each of these sources of
data suffer, to a greater or lesser extent, from problems of measurement
error. In addition they provide only a snapshot of an individual's diet.
There is large within-individual variability with components due for example
to season, day-of-the-week, and there is also, in general, correlation
between measurements obtained on successive days. When an outcome such as
blood pressure is modelled as a function of dietary variables it is vital to
account for each of these factors, as an analysis which does not acknowledge
the errors-in-variables aspect will produce biased estimates.
In this project, Bayesian and classical parametric errors-in-variables
models will be utilized to provide corrected estimates of effect. Such
models contain layers of assumptions. We will carefully assess each of
these, supported by the dietary expertise within the departments. A number
of large databases, that include repeated measures of dietary variables, are
available for study including INTERSALT and the Dietary and Nutritional
Survey of British Adults. The estimates obtained from our models will be
compared with more-recently proposed semi-parametric techniques.
COLLABORATIVE WITH WRC:
7. DEVELOPMENT OF EXPOSURE MODELLING TECHNIQUES FOR EPIDEMIOLOGICAL STUDIES
CONSIDERING CONSTITUENTS OF DRINKING WATER (J WAKEFIELD)
Current approaches to investigate possible linkages between drinking water
constituents and public health at ecological level use Poisson regression
techniques to consider factors such as demographic profile and mean
distribution and uptake levels of constituents from drinking water and other
sources. This approach fails to take account of the variability of
constituents over time and space (i.e., within water zones) and measurement
error, the so-called errors-in-variables problem. It may lead to attenuation
of effects between exposure data and potential health outcomes, which could
result in mis-specification of regulatory requirements for drinking water
quality based on health standards.
This project aims to develop a range of modelling techniques to take account
of variability across time and within different water distribution zones to
allow for errors in estimation of the effects of long-term exposure. This
would be a simulation based approach and would be tested on a number of
case-studies, eg water hardness and cardiovascular disease, to demonstrate
its applicability.
8. BACKGROUND INCIDENCE OF G-I TRACT DISEASE AND WATER-BORNE
MICROORGANISMS (N BEST)
Studies in North America indicate that there may be a proportion of
gastro-intestinal illness which is due to microorganisms present in drinking
water even when the water supplies concerned show no increases in the
indicator organisms. Such disease could be due to chlorine resistant
pathogens or to the release of opportunist pathogens which can multiply in
biofilm in distribution and plumbing. If this is so, the most likely
situation is that the illness caused will be insufficiently severe to
trigger reporting to public health authorities, and the conditions will go
largely unrecognised against background reporting of symptoms, e.g., in
general practice. The increased level of microbiological contamination is
most likely to occur as a consequence of some operational activity which
will destabilise the biofilm or result in an ingress of resistant organisms.
It is proposed to study the incidence of G-I tract complaints over time
presenting to general practices within a defined geographical area, and to
attempt to correlate changes in incidence with water supply operational
activities. The analysis will require surveillance methods to detect acute
changes in symptom reporting rates based on computerised GP databases, with
concurrence of ‘peaks’ of reporting in both time and space. It is proposed
that such an analysis is carried out within a hierarchical Bayesian
framework to deal appropriately with correlated errors occuring in both time
and space.
The National Centre for Environmental Technology through WRc plc will
provide support in data collection from water company records in addition to
access to three water quality models which will provide a more detailed and
sophisticated analysis of potential exposure data, and the impact of changes
in water quality or operations at the water treatment works. The models are
Watnet, WatQual and Weasel.
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
|