Application of Factor Analysis to Identify Dietary Patterns and Use of Factor Scores to Study their Relationship with Nutritional Status of Adult Rural Populations

Abstract

The prevalence of chronic energy deficiency (CED) among one-third of the Indian population is attributed to inadequacy of consumption of nutrients. However, considering the complexity of diets among Indians, the relationship between a particular dietary pattern and the nutritional status of the population has not been established so far. A community-based cross-sectional study was undertaken to assess estimates, at district level, of diet and nutritional status in Orissa State, India. Factor analysis was used for exploring the existence of consumption pattern of food and nutrients and their relationship with the nutritional status of rural adult population. Data on 2,864 adult men and 3,525 adult women in Orissa state revealed that there exists six patterns among food-groups explaining 59% of the total variation and three patterns among nutrients that explain 73% of the total variation among both adult men and women. The discriminant function analysis revealed that, overall, 53% of the men were correctly classified as either with chronic energy deficiency (CED) or without CED. Similarly, overall, 54% of the women were correctly classified as either with CED or without CED. The sensitivity of the model was 65% for both men and women, and the specificity was 46% and 41% respectively for men and women. In the case of classification of overweight/ obesity, the prediction of the model was about 75% among both men and women, along with high sensitivity. Using factor analysis, the dietary patterns were identified from the food and nutrient intake data. There exists a strong relationship between the dietary patterns and the nutritional status of rural adults. These results will help identify the community people with CED and help planners formulate nutritional interventions accordingly.

Key words: Caloric intake; Chronic energy deficiency; Community-based studies; Cross-sectional studies; Discriminant function; Factor analysis; Nutritional status; Principal component analysis; India

DOI: http://dx.doi.org/10.3329/jhpn.v29i4.8448

JHPN 2011; 29(4): 327-338