Is the use of a questionnaire on non-cereal foods consumption is reliable and valid to measure households’ food insecurity: experience and understanding from chronically poverty prone region in Bangladesh

The objective of this study is to assess the reproducibility and validity of Food Frequency Questionnaire (FFQ) from non cereal dietary diversity for the identification of rural households’ food insecurity from an in-depth understanding at Tista and Jamuna basin poverty prone region in Bangladesh. Data in different season from the Bangladesh Nutritional Surveillance Project, 2005 of Helen Keller International was used in this study. Pearson co-relation, regression analysis and Kruskal-wallis test were carried to understand about the reproducibility of the FFQ. Sensitivity and specificity and chi-square test were performed for the validity study. The resulted Pearson correlation between total meal consumption in different seasons and total energy consumption were found 0.83 and 0.81 respectively. During monsoon rice harvesting period (December-January) the mean calorie consumption from non-cereal food were remarkably decreases compare to all other seasons but the energy consumption form meat other than chicken are significantly increased (p 0.05). Sensitivity and specificity were found 97.70% and 12.50% respectively as well as the chi-square test resulted as 74.577 (p 0.001).This FFQ measured can be best lead to suitable experience-based food insecurity measures.