Abstract
<title>Abstract</title> <p> <bold>Background</bold> Food Balance Sheets (FBS) are used to estimate national energy availability and support food security indicators such as the Prevalence of Undernourishment (PoU). However, FBS provide a single per-capita value that may not represent population energy intake distribution. This study proposes an algorithm to estimate usual intake distributions from FBS data alone. <bold>Methods</bold> A Gaussian-based refinement algorithm was developed using two assumptions: a population coefficient of variation of 25% and the interpretation of the reported FBS dietary energy value as the upper bound of the population intake distribution rather than the population mean. The algorithm was validated using nationally representative dietary survey data from six countries. <bold>Results</bold> The proposed algorithm produced dietary energy intake distributions that closely matched the observed survey distributions across the six study countries. The mean absolute error was 125 kcal/day, the mean relative error was 7%, and no statistically significant difference was observed between the refined and survey estimates (P = 0.456). Compared with the current FAO methodology, the refined distributions showed substantially better agreement with nationally observed dietary energy intake distributions. <bold>Conclusions</bold> The proposed refinement algorithm provides a practical approach for estimating population usual energy intake distributions directly from FBS data, with improved agreement with nationally representative dietary surveys. </p>