A study using a food-frequency questionnaire--simultaneous discrimination of dietary patterns of anemic and hyperlipemic women

Abstract

A survey investigating the intake frequency of various foods and eating habits was made by a special form of structured questionnaire. Those who answered the questionnaire were 184 women aged from 40 to 59 living in a rural district in Kyoto Prefecture. They were divided into three blood groups according to the contents of hemoglobin (Hb) and serum cholesterol (T-CHO) as follows: Group A (Anemia): Hb less than 12 g/dl and T-CHO less than 220 mg/dl Group B (Normal): Hb greater than or equal to 12 g/dl and T-CHO less than 220 mg/dl Group C (Hyperlipidemia): Hb greater than or equal to 12 g/dl and T-CHO greater than or equal to 220 mg/dl As far as the intake frequency of each food and the eating habits of the three blood groups are concerned, there wasn't any notable difference. However, when the statistical method called "quantification theory II" was applied using the blood groups as the objective variable and the 27 items in the food frequency questionnaire, as well as the obesity index, as explanatory variables, it was found that the dietary patterns were different among the three blood groups. The result was that 68% of women in group A were discriminated correctly, as were 55% in group B, and 83% in group C. The analysis of the data from food frequency questionnaires by the "quantification theory II" statistical methods seems to be useful to reveal dietary patterns.