Proxy reporting in the National Population Health Survey.

OBJECTIVES

This article examines the extent of proxy reporting in the National Population Health Survey (NPHS). It also explores associations between proxy reporting status and the prevalence of selected health problems, and investigates the relationship between changes in proxy reporting status and two-year incidence of health problems.

DATA SOURCE

Cross-sectional results are based on the 1996/97 NPHS Health file and General file. Longitudinal results are based on 1994/95 respondents who were still residing in households in 1996/97.

ANALYTICAL TECHNIQUES

The extent of proxy reporting in the various NPHS files was computed. Prevalence estimates of selected health problems from the two 1996/97 cross-sectional files were compared. Multivariate analyses were used to estimate associations between proxy reporting status and health problems.

MAIN RESULTS

For several health conditions, prevalence estimates based on the 1996/97 cross-sectional Health file (where proxy reporting was less common) were significantly higher than estimates derived from the General file. Individuals whose data were proxy-reported in 1994/95 and self-reported in 1996/97 had higher odds of reporting new cases of certain health conditions.