Methodology for assigning appropriate glycaemic index values to an Australian food composition database
We aimed to produce an updated Australian glycaemic index (GI) database based on a systematic method. GI values were assigned to the 3871 unique foods in an Australian food composition database. Following the method, 1124 (29%) foods had less than 2.5 g of available carbohydrates per 100 g and were assigned a GI of 0, and 416 (11%) foods had a direct match in one of the three data tables used. The GI value of a 'closely related' food was assigned to 1793 (46%) foods; 135 foods (3%) had their GI values calculated using the weighted average GI method; 391 (10%) foods were assigned the median GI of their corresponding food subgroup, and 12 (<1%) foods were assigned a GI of 0 because they were not significant sources of carbohydrates in a typical diet. For the 3634 foods which received a GI value in the 2009 assignment, 1954 (53.8%) had an updated GI value, and the mean ± SD difference between the 2009 and current assigned values was +3.0 ± 16.0 units (paired sample t-test p < 0.001). Acknowledging some limitations, this database will enhance the utility of the GI concept in research and clinical settings in Australia (199 words).