Beyond the Food Label: Nutrient Profiling the Indian Nutrient Databank Comparing WHO South-East Asia and NRF 9.3 Scores to Inform Public Nutrition Policy

Abstract

This study applied different nutrient profiling models to the Indian Nutrient Databank (INB), to classify and describe Indian foods. The INB is a database of home-cooked recipes and dishes (rather than packaged, branded products) including 1014 foods, 940 of which met inclusion criteria for this analysis. Foods were organized using a novel “What We Eat in India” food categorization scheme (broad food groups, each with more specific subgroups) comprising of 11 food groups and 28 subgroups[PC1.1]. Two nutrient profiling (NP) methods were applied: the dichotomous (pass/fail[PC2.1]), meaning a food is classified as passing or failing category-specific nutrient thresholds, World Health Organization for South-East Asia Region (SEARO) NP model and the continuous Nutrient Rich Food (NRF) 9.3 nutrient-density score. The WHO model uses 100 g as the base of calculation; the NRF 9.3 system uses 100 kcal. Under the WHO SEARO model, only 27.1% of foods passed; total fat was the leading driver of failure (52.2% of foods), followed by total sugar (30.5%). Under NRF 9.3, most foods scored negative (mean -39.0; 20% positive), indicating most foods had more factors to limit (Fat, sugar and sodium) then nutrient to consume. Vegetables were the only food group with a positive mean. The two models showed agreement on 80.1% of foods. Some foods eaten in small portions were penalized in the WHO SEARO model because of the per 100 g basis. These findings show that NP models can be applied to recipes for home-cooked foods and not just to packaged processed foods as home preparation does not inherently confer a more favorable nutrient composition[PC3.1]. Further, the base of calculation (100 g vs 100 kcal) can alter nutrient density scores, as can the choice of index nutrients; nutrient profiling models should credit beneficial (encourage) nutrients as well as flag nutrients to limit, rather than accounting for only one side of a food's nutrient content. NP models that include typical recipes can provide a more complete basis for nutrition policy in India.