A statistical decision support system incorporating personalised adaptive reference ranges for longitudinal monitoring in prostate cancer
Roshan Sangachin, Davood
MetadataShow full item record
This item's downloads: 0 (view details)
The overall aims of this thesis are to use modern approaches in Biostatistics to help clinicians diagnose Prostate Cancer (PCa) early, and treat effectively, and to help patients choose between treatment options. Biostatistics is both a primary and an enabling discipline, a fundamental requirement of all quantitative research, upon which the validity and integrity of research findings are dependent. This thesis encompasses both aspects. The primary (methodological) component involved the development of new and novel methods for generating adaptive ranges in longitudinal monitoring of biomarkers. The enabling component was to deliver on the requirements outlined by my funders (Prostate Cancer Institute (PCI) in the National University of Ireland, Galway), namely to build a decision support system that i) displays useful summary information of PCI data from several sources and ii) presents the results of several statistical analyses relating to treatment comparisons and outcomes in a manner that was informative to clinicians and patients. The thesis is comprised of three Work Packages (WP). In the first WP, methods to generate personalised adaptive reference ranges (i.e. ranges that adapt and account for an individual's previous data) will be developed that allow clinicians to identify meaningful changes in an individual's blood test results more quickly compared to decision making using conventional normal ranges. Application of biomarker monitoring in elite sports will be presented also. Current techniques involve implementation of a Bayesian approach when the variability within individuals is assumed to be fixed. This thesis will further extend the current literature by accommodating different within individual variability structure that is more realistic, and will result in wider applicability. Additionally, the use of an approximate EM algorithm to produce computationally efficient adaptive ranges for large streaming datasets will also be proposed. A comprehensive simulation study will be undertaken to assess the performance of the methods proposed. The second WP relates to identifying and assessing the main health outcomes following PCa treatment. In particular, the PCa treatment outcomes under different treatment options and based on different risk factors will be compared using suitable statistical methods. Finally, in the third WP a modern statistical decision support system will be developed to enable patients make more informed and reliable decisions about their treatment choices. To conclude, areas of further work across the three WPs will be outlined.
This item is available under the Attribution-NonCommercial-NoDerivs 3.0 Ireland. No item may be reproduced for commercial purposes. Please refer to the publisher's URL where this is made available, or to notes contained in the item itself. Other terms may apply.
The following license files are associated with this item: