A policy perspective on household food security measures (extent, severity, and correlates): A population-based survey in Western Australia

Abstract

Objectives

To compare the extent, severity and correlates with household food insecurity among Western Australian adults using the U.S. and Canadian Household Food Security Survey Module (HFSSM) coding systems and assess their suitability for government policy considerations.

Study design

Continuous cross-sectional population representative computer-assisted telephone interviews.

Methods

Between 2017 and 2021, the United States Department of Agriculture's HFSSM was added to the Western Australian Health and Wellbeing Surveillance System in addition to the usual single item measure. Weighted population estimates of the extent and severity of food insecurity were calculated (n = 8684 households), and demographics, socio-economic and lifestyle-related health risk factors were analysed using bivariable regression models, with multivariable models fitted to assess associations for both the U.S. and Canadian coding systems.

Results

Food insecurity prevalence was 3.0, 3.6 and 5.1 percent using the single item, U.S. and Canadian schemes, respectively. There are significant differences between HFSSM classification schemes (p=0.04), the U.S. reported 1.8 % for both of moderate and severe, compared to 3.4 % moderate and 1.7 % severe with the Canadian. A significantly higher proportion of food insecurity was reported among households with children. Both systems found significant positive associations with inability to save money, low educational attainment, low income, not owning a house, higher Body Mass Index, stressful life events, and psychological distress (p<0.05), with the strength of associations lower using the Canadian system.

Conclusions

Dimensionality and thresholds applied in HFSSM classification influence the extent and severity of food insecurity reported. Robust definition of the policy problem, the first step in policy deliberations, is critical. Findings highlight the importance of instrument and choice of classification scheme in defining the problem and the inadequacy of the single item measure routinely used in Australia.