Error sources in data handling process of the national nutrition survey, Japan, and possible measures for lessening the errors

The National Nutrition Survey estimates the average dietary intake per capita on the basis of household food consumption data. Dietitians at local public health centers manually execute the coding of food names, the total amount consumed, and the proportion shared by family members to compute the nu...

Full description

Saved in:
Bibliographic Details
Published inNihon Eiyō, Shokuryō Gakkai shi Vol. 51; no. 2; pp. 57 - 65
Main Authors Yoshiike, N. (National Inst. of Nutrition, Tokyo (Japan)), Iwaya, M, Otani, Y, Inoue, K, Kawano, M, Seino, F, Matsumura, Y, Yamaguchi, M
Format Journal Article
LanguageJapanese
Published Japan Society of Nutrition and Food Science 1998
Subjects
data management
DATA PROCESSING
dietary survey
ENCUESTAS NUTRICIONALES
ENQUETE NUTRITIONNELLE
error sources
JAPAN
JAPON
National Nutrition Survey
NUTRITION SURVEYS
PROCESAMIENTO DE DATOS
quality control
TRAITEMENT DES DONNEES
Online AccessGet full text

Cover

More Information
Summary:The National Nutrition Survey estimates the average dietary intake per capita on the basis of household food consumption data. Dietitians at local public health centers manually execute the coding of food names, the total amount consumed, and the proportion shared by family members to compute the nutrient intake of individual subjects. The coded data are then checked and corrected by other dietitians at prefectural departments. We examined sources of error in both procedures, and described the frequency of errors and their effect on the calculation of energy and nutrient intake by suing a sub-sample of 132 households, including 445 family members, surveyed in 1995. The errors in coding the foods and the total amount consumed, including no-coding for a specified food, occurred at 1.61 per household, and those in coding the proportion shared by family members were 0.19 at the public health center level. The frequencies for prefectural departments wee 1.57 and 0.16 per household, respectively. This suggests that the present data-handling procedures are not satisfactory. In conclusion, we recommend that a new system for coding and data entry using personal computer software should be adapted by public health centers. Such a system should have automatic error checking procedures by incorporating a food code list and establishing upper limits for the acceptable values for each food
Bibliography:S30
1998005984
ISSN:0287-3516
1883-2849
DOI:10.4327/jsnfs.51.57