Use this URL to cite or link to this record in EThOS: http://ethos.bl.uk/OrderDetails.do?uin=uk.bl.ethos.597194
Title: Understanding and controlling for biases in evidence synthesis in population health genetics
Author: Butterworth, A. S.
Awarding Body: University of Cambridge
Current Institution: University of Cambridge
Date of Award: 2009
Availability of Full Text:
Full text unavailable from EThOS. Please contact the current institution’s library for further details.
Abstract:
In this project, I have conducted an empirical investigation of the evidence for the impact of a number of potential diversity factors (or biases) in the field of coronary heart disease (CHD). BY collecting data from 47 published meta-analyses containing over 1000 genetic association studies. I was able to use ‘meta-meta-analytic’ techniques (e.g. a ratio of odds ratios[ROR] method) first established in clinical trials to assess the magnitude and direction of any effects. Chapter 2 of this thesis describes the process of data collection as well as the design and construction of the database containing these data. Chapter 3 summarises the results of the meta-analyses in terms of their results in the form of a ‘field synopsis’ of the latest understanding of the role of genetic variants in CHD, whilst Chapter 4 describes the dataset in the context of the structure of genetic epidemiological research in CHD. The empirical research, outlined in Chapter 5, describes the ROR approach used, as well as alternative methods used to assess whether the potential diversity factors systematically influence the results of the association studies or appear to affect between-study heterogeneity. To further investigate the impact of reporting bias on CHD genetic epidemiology, I surveyed investigations from over 250 collections of disease cases to discover which of ten candidate gene variants they had genotyped and reported on in their collection. Chapter 6 describes the rationale and methods for this survey, before considering what the results can tell us about the amount of unreported data. Having established that study size effects appear to be the most influential source of variation in study results. Chapter 7 looks at how to deal with this issue when conducting meta-analyses. A comparison of traditional methods and more recent methods to address bias is described.
Supervisor: Not available Sponsor: Not available
Qualification Name: Thesis (Ph.D.) Qualification Level: Doctoral
EThOS ID: uk.bl.ethos.597194  DOI: Not available
Share: