Enhancing Aspects Of Time-to-event Analysis Methodology In Randomised Trials
Funder
National Health and Medical Research Council
Funding Amount
$548,446.00
Summary
Time-to-event analysis is a statistical method for examining the occurrence of disease-related events in individuals followed for varying periods of time. The method is widely used in health research. The technicalities of the methods are subtle and by paying careful attention to these this grant will provide extended methods, new software, and apply methods more effectively to gain new insights to disease progress, and to enhance the efficiency of health research.
How Should We Analyse, Synthesize, And Interpret Evidence From Interrupted Time Series Studies? Making The Best Use Of Available Evidence
Funder
National Health and Medical Research Council
Funding Amount
$445,144.00
Summary
Interrupted time series (ITS) studies are frequently used to evaluate whether policy interventions are successful. The findings from these studies are often collated into systematic reviews, which are used to inform healthcare decisions by clinicians, consumers and policy makers. It is not known how the statistical methods, which underpin the findings from ITS studies, perform. This proposal will evaluate the statistical methods and provide guidance on how to analyse and interpret ITS studies.
Complex Statistical Analyses Of Genome-wide Association Studies Related To Breast And Prostate Cancers Using High Performance Supercomputing
Funder
National Health and Medical Research Council
Funding Amount
$656,073.00
Summary
Breast and prostate cancers are the most common cancers in Australian women and men. Simple analyses of genome-wide association (GWAS) studies explain only a fraction of why these cancers run in families. The University of Melbourne now has a supercomputer that can conduct much more complex analyses. We will apply these to the world’s GWAS data for breast and prostate cancers. We hope to learn more about the causes of these cancers, and expand an expert Australian workforce in supercomputing.
New Methods And Guidelines For The Design, Analysis And Reporting Of Cluster-crossover And Stepped Wedge Randomised Trials In Clinical And Public Health Research
Funder
National Health and Medical Research Council
Funding Amount
$391,258.00
Summary
Cluster crossover and stepped wedge designs have emerged recently as study designs used in clinical and public health research settings. However, despite their use there has been very little methodological assessment of the statistical analysis methods used in current practice. The proposed research will assess the correctness of current methods and will produce a toolkit of state-of-the art, defensible trial design, analysis and reporting methods for the clinical/health researcher.
Cost-effective Development And Validation Of Risk Prediction Models: Methods And Applications
Funder
National Health and Medical Research Council
Funding Amount
$157,322.00
Summary
We will develop statistical methods and software that will enable risk calculators being built and validated using study designs that require only a fraction of current cost. The methods and software will be freely-available to researchers worldwide.
Constructing Control Samples For The Australian And Other Populations: Improving Power And False Positive Rates In The Next Generation Of Genetic Association Studies With A Focus On Controlling For Fine-scale Population Structure In DNA Sequence Data
Funder
National Health and Medical Research Council
Funding Amount
$283,447.00
Summary
Individuals who live near each other tend to be more similar genetically than individuals who live in different parts of the world. One reason is that they share more of their genetic ancestry. There can be very subtle differences in patterns of genetic variation even within countries. Accounting for these subtle differences can be important for studies of the genetic basis of diseases. We will develop novel statistical methods to control for these genetic differences in disease studies.
Statistical Issues In The Analysis Of Host-viral Genetic Associations
Funder
National Health and Medical Research Council
Funding Amount
$252,849.00
Summary
Viruses such as HIV or Hepatitis B or C may escape detection, and hence control, by the human host immune system through strategic mutations. This project aims to develop statistical methodologies which will facilitate determination of the locations of these strategic mutations and assess the relative impact of different human-host genetic characteristics. Application of the results should benefit the design of tailored vaccines and studies of drug resistance.
Implementing Multiple Imputation With Sensitivity Analysis In Large-scale Longitudinal Studies
Funder
National Health and Medical Research Council
Funding Amount
$473,507.00
Summary
Missing data arise in most research studies and if not handled appropriately can mean the study results are not correct. With researchers now conducting larger and longer studies the challenges posed by missing data are increasing. In this grant we study a powerful technique for handling missing data, which in its current form often cannot be applied effectively in large studies. By developing this approach we will improve the accuracy of results from large-scale epidemiological studies.