The use of cohort vs repeated cross-sectional sample survey data in monitoring changing breast cancer screening practices.

Abstract

The method used to select a study sample is a key element in designing a research protocol. This article explores two of the more common methods used, focusing on the relative advantages and disadvantages of each one. Both cohort and repeated, independent cross-sectional surveys were conducted in each of 3 years (1988-1990) in the Awareness of Breast Cancer Screening Project to follow changing breast cancer screening rates among a population of women 50-75 years of age on Long Island, New York. Both survey methods revealed statistically significant increases in self-reports of mammography use. The cohort and repeated cross-sectional survey sample responses to questions concerning knowledge, attitudes, and behavior regarding breast cancer screening were comparable. An educational effect of the baseline survey itself on the cohort was not seen. Each of the two survey methods has advantages and disadvantages with respect to the other. The cohort method permits examination of changes in the same individual over time and is less costly and less time-consuming to perform. On the other hand, the cross-sectional method does not suffer from cumulative losses in respondents with repeated surveys and better reflects the changing community. The study findings can be used to guide the selection of an appropriate survey methodology for monitoring breast cancer screening practices in other settings.