[Concurrent validity of self-reported weight and height for diagnosing elderly women's nutritional status]

The present study was aimed at assessing the concurrent validity of selfreported weight and height measurements when diagnosing low-income elderly women's nutritional state.

The sample consisted of 181 women (average age being 70 + or - 7.2) taken from groups running activities for elderly people in the city of Jequié, Bahia, Brazil. Reported weight and height measurements were collected through individual interviews preceding direct measurements being taken. Central tendency and dispersion measurements were used during data analysis for characterising the sample and Student's t-test was used for evaluating the differences between the averages for the reported and real measurements. The intraclass correlation coefficient and the Kappa index were used for checking agreement between measurements (p<0.05).

The results highlighted a tendency to underestimate reported weight and height anthropometric measurements. Satisfactory agreement indexes were noted for the reported weight measurements (varying from r(icc) = 0.74 to r(icc) = 0.97 in eutrophic people) when considering economic class and the absence of cognitive deficit. On the other hand, concordance indexes for the reported height measurements were low (varying from r(icc) = 0.03 to 0.28).

Data analysis results led to concluding that the reported weight measurements could be considered to be valid in elderly, low-income, low scholastic level populations. However, the same cannot be said about the reported height measurements, or about the use of the nutritional state indicator, based on combined reported weight and height measurements.