Published Papers - Abstract 83

Lee C. When are longitudinal survey participants too old? Preferences of members of the Women’s Health Australia older cohort. American Journal of Epidemiology, ; :

Objective: Large-scale longitudinal surveys of older cohorts raise ethical and methodological questions about the appropriateness of conducting research with elderly people. The Women’s Health Australia (WHA) project includes a cohort of over 10,000 women who were aged between 70 and 75 when first surveyed by mail in 1996, and who will be between 76 and 81 when surveyed for the third time in 2002. To explore the appropriateness of sending lengthy mailed surveys to women in this age group, a telephone study was used to ask members of the cohort their opinions about survey length, mode of administration, and content.Method: A total of 110 members of the older cohort of WHA, who had responded to a written survey in 1999, were selected randomly, stratified according to their self-rated health in 1999. After a contact letter, they were telephoned and administered a brief structured interview.Results: Response rate was 79%. All indicated willingness to complete a survey in 2002, and over 90% preferred mail rather than telephone. The main reasons were being able to complete the survey at one’s own pace and provide accurate data. The majority were willing to complete the same 24-page survey as in 1999. Most women expressed enthusiasm for the survey. When asked to identify important health issues, most mentioned quality of life, independence and functional capacity.Conclusions: Older women participating in WHA are interested in health issues, willing to complete lengthy written surveys, and positive about public health research. Their preference for mail rather than phone administration is based on a concern that their responses should be accurate.Implications: These older women find survey research acceptable and not at all burdensome. Potential ethical concerns regarding working with this population appear largely unfounded. Mail surveys are the preferred mode of administration, suggesting that data quality will be enhanced if this method is used.