Women's Health Australia homepage about the Women's Health Australia project Women's Health Australia staff Women's Health Australia current events Women's Health Australia surveys and data Women's Health Australia substudies information for Women's Health Australia participants University of Newcastle contact Women's Health Australia Women's Health Australia publications and presentations University of Queenlsand Women's Health Australia homepage about the Women's Health Australia project Women's Health Australia staff Women's Health Australia current events Women's Health Australia surveys and data Women's Health Australia substudies information for Women's Health Australia participants University of Newcastle contact Women's Health Australia Women's Health Australia publications and presentations Women's Health Australia homepage about the Women's Health Australia project Women's Health Australia staff Women's Health Australia current events Women's Health Australia surveys and data Women's Health Australia substudies information for Women's Health Australia participants University of Newcastle contact Women's Health Australia Women's Health Australia publications and presentations Women's Health Australia homepage about the Women's Health Australia project Women's Health Australia staff Women's Health Australia current events Women's Health Australia surveys and data Women's Health Australia substudies information for Women's Health Australia participants University of Newcastle contact Women's Health Australia Women's Health Australia publications and presentations

Women's Health Australia homepage about the Women's Health Australia project Women's Health Australia staff Women's Health Australia current events Women's Health Australia surveys and data Women's Health Australia substudies information for Women's Health Australia participants University of Newcastle contact Women's Health Australia Women's Health Australia publications and presentations Women's Health Australia homepage about the Women's Health Australia project Women's Health Australia staff Women's Health Australia current events Women's Health Australia surveys and data Women's Health Australia substudies information for Women's Health Australia participants University of Newcastle contact Women's Health Australia Women's Health Australia publications and presentations

 
 


Research project

Steven Bowe
PhD candidate

 

 

Doctoral Thesis: Statistical methods to adjust for death in longitudinal studies
Supervisors: Dr Anne Young (University of Newcastle) & Dr David Sibbritt (University of Newcastle)
University: Centre for Clinical Epidemiology and Biostatistics, University of Newcastle, Australia.

Aims of the research:

  • To investigate the statistical methods used to account for death in longitudinal studies
  • To apply the current statistical methods to ALSWH data for the older cohort and evaluate the advantages and disadvantages of the methods
  • To determine whether there is a need to improve current statistical methods and apply and assess new strategies if applicable
  • To examine the impact of diabetes on quality of life among older women - adjusting for deaths by applying the methods developed.

Progress:

A literature review to examine the statistical methods that are currently used to account for dropout due to death has been conducted. A method proposed by Diehr and colleagues has been applied to ALSWH data. The method transforms the physical component score (PCS) of the SF-36 to a new score which estimates the probability of being healthy at the next time point. A value of zero is assigned to participants at time points when they have missing data due to death.

The methodology was applied to Survey 1 and Survey 2 ALSWH data and validated by applying the derived transformation to Survey 2 and Survey 3 of the Older cohort. The transformation derived from the ALSWH data provides evidence that the methodology for transforming the PCS to account for deaths is sound. The three-year equation provided good estimates of the probability of being healthy in three years and the method allowed deaths to be included in an analysis of changes in health over time. We applied the method when comparing changes in health related quality of life between Survey 1 and Survey 2 for women with and without self-reported diabetes.

Without adjusting for deaths, there was no significant difference in changes over time in health related quality of life (PCS) for women with and without diabetes. However when analysing the transformed PCS which included a value for deaths, women with diabetes had a significantly greater decline in health than women without diabetes. In a further phase of the method, missing values for reasons other than death are being imputed by a variety of methods, to determine whether including a value for death (zero) is overly influential. The methods of imputation for longitudinal data are the current focus of research. A paper describing the application of the method has been accepted to Medical Care.

To contact Steve:
Steven Bowe
Centre for Clinical Epidemiology and Biostatistics
University of Newcastle, 2308


E mail: steven.bowe@newcastle.edu.au

 
Link to Project Aims page Link to Project Progress page Link to Project Methods page Link to Project Progress page 		 


 
 

Last updated: 25 February 2005 by Cath Chojenta © Copyright