Diane Hindmarsh

CSSM, School of Mathematics and Applied Statistics, University of Wollongong

The NSW population health survey and small area estimation and analysis

This is joint work with David Steel and Ray Chambers, both CSSM

The CATI-based population health survey program run by NSW Health publishes annual estimates at the state level and for the eight health areas for a large number of health related measurements. Each of the eight health areas is geographically and socially diverse, and a single estimate for an entire area ignores this fact. Direct estimators at the Local Government Area level, or for age-sex domains at the health area level, have relatively high standard errors, which affects their use in health evaluation and planning. There is a desire to use small area estimation methodologies to provide estimates that have greater precision, but two questions need answering: which of the many methods is best suited to this specific situation, and, how do we educate the end-users about the resultant estimates so that they use them with confidence rather than revert to the direct estimates. This paper will outline the NSW Health Survey program and how we are answering these two questions.