Estimates of apparent food consumption can be computed from various types of household consumption and expenditure surveys (HCES). Such surveys typically collect information on food acquired and/or consumed (i.e. apparent consumption) at the household level, usually during the last 7 or 14 days, without considering food wasted and the cooking method. This contrasts with individual-level quantitative dietary intake surveys (e.g. 24-hr recall or food record), which are purposefully designed to collect detailed information on all food and beverages consumed by individuals over a defined period of time. Furthermore, while individual level dietary surveys are designed to appropriately capture foods prepared and consumed away from home, HCES usually collect information only on the monetary value of a set of foods/dishes which include foods that are not well-described (e.g., lunch, dinner, etc.).

Computing food and nutrient statistics from HCES data

To produce estimates of apparent nutrient intake in a population, food quantities collected in HCES should be first adjusted for non-edible portions and then converted into nutrients. Food matching is the process of linking nutrient content in foods from food composition tables/databases with information on food quantities reported by households. FAO promotes the use of FAO/INFOODS food matching guidelines, and a specific nutrient conversion table (NCT) is produced for each survey food list. These NCTs are available upon request for a set of surveys. For foods that were reported but no quantities were recorded (which is often the case for food prepared away from home, for example), or the nutrient content is unknown, an indirect method is used. The nutrient content is then estimated based on the median nutrient unit cost (for the region, urban-rural group, and income quintile group) computed based on food reported for at-home consumption.