Methodology for adding glycemic index values to 24-hour recalls

Objective: To describe a standardized method to assign glycemic index (GI) values to food items, obtained from 3 x 24-h recalls among Aboriginal and Torres Strait Islander and non-Indigenous Australian children, which can be adapted for use with simple food composition databases.

Methods: Four published GI databases were used as the source of GI values. Changes were made to a previously published methodology for GI value assignment to accommodate the needs of the Many Rivers Diabetes Prevention Project.

Results: There were 1132 food items in the recall database. Two hundred nineteen (19.3%) food items were directly linked to the FoodWorks GI database and 545 (48.1%) items were assigned the GI value of a "closely related" food item in the four GI databases used. Among the top carbohydrate contributors, 113 (35.3%) items have a direct linkage with the FoodWorks GI database. The mean ± SEM dietary GI and glycemic load (GL) of the study population resulting from this methodology are 57.5 ± 0.3 and 143.4 ± 2.6, respectively.

Conclusion: This simple method provides opportunities for countries without food composition database that are comprehensive for GI/GL but which contain accurate information on carbohydrates in foods to assign high-quality GI values to food items in epidemiological studies based on 24-h recalls.