Validity of Food insecurity experience scale (FIES) for use in rural Bangladesh and prevalence and determinants of household food insecurity: An analysis of data from Bangladesh integrated household survey (BIHS) 2018-2019

"Access" dimension of Food insecurity (FI) is directly measured by the Food Insecurity Experience Scale (FIES). The current study assessed the appropriateness of the FIES for measuring FI in rural Bangladesh, followed by an assessment of FI prevalence and its correlates utilizing Bangladesh Integrated Household Survey (BIHS) data. The internal validity of the FIES and the prevalence of FI were investigated using the Rasch modeling approach. We utilized equating procedure to calibrate the study's result to the global FIES reference scale and determined FI prevalence rates that were comparable across countries. The external validity of the FIES was evaluated by examining its association with other FI measures using Spearman's rho correlation analysis. With an overall Rasch reliability of 0.84, the FIES met the Rasch model assumptions of conditional independence and equal discrimination, and as well as the fit statistics standards for all eight items. Infit statistics were within the allowed limit for all FIES items indicating good internal validity. However, we noted a high outfit (>2) for the "unable to eat healthy and nutritious food" item indicating the presence of some unusual response patterns. Our analysis found no significant (>0.4) correlation between FIES items. We also found a significant correlation between FIES and other FI proxies, e.g., the Household hunger scale (HHS), Food consumption score (FCS), and Household dietary diversity score (HDDS). Overall, the prevalence of moderate or severe FI was 18.92% in rural Bangladesh. Geographic areas, access to electricity, household ownership, access to sanitation, livestock ownership, family size, education level, and monthly per capita food expenditure significantly explained the variation in FI. Our analyses suggest that the FIES is internally and externally valid for FI measurement in rural Bangladesh. However, FIES questions may need to be reordered to more accurately evaluate lower levels of FI, and the item "unable to eat healthy and nutritious food" may need cognitive testing.