Some methodological problems with the nutritional assessment of the 1997-98 El Nino drought in PNG

摘要

Assessment of the impact of the 1997-98 drought on the rural population of PNG was primarily based on a sequence of three, rapid, national evaluations of food and water availability in local areas. Direct nutritional assessment played a minor, and late, role. An ambitious sample nutrition survey was conducted in early 1998, but the worst of the crisis was over by April when the survey report was completed. This paper describes three methodological problems associated with the 1998 survey that may have limited its value to emergency administrators. Firstly, comparisons between the children's nutritional status of 1998 with those obtained from the 1982-83 PNG National Nutrition Survey (NNS) were not valid. This is because the age groups measured were different, with very young infants (045 months) not included in the 1998 survey. Since the nutritional status of PNG infants in this age group is relatively better than for older children, exclusion of this age group may have biased the 1998 results to appear worse than the 1982-83 ones. Secondly, comparisons of the results of the 1998 and 1982-83 surveys were invalid because of differences in the population sampling methods. Unless adjustment for geographical location is made, significant bias can occur. Finally, there were problems with making accurate anthropometric measurements of adult women, as indicated by those from some locations showing unusually high or low bodymass indices. Comparison of the results from one area with those of an earlier survey shows that measurement error is likely. These findings emphasise both the need for appropriate sample selection if future anthropometric surveys are to take advantage of the 1982-83 NNS as baseline information, and the value of previous baseline data for evaluating results. It is suggested that the NNS still provides an invaluable baseline for children's nutritional status.