The aim of the present study was to validate an in vitro assay for quantifying resistant starch (RS) in foods against an in vivo model. The amount of starch escaping digestion in the small intestine of ileostomates was compared with that observed by using the in vitro assay. Subjects with ileostomies were fed five foods containing different types and amounts of RS (baked beans, pearl barley, cornflakes, and whole and ground rice). The total amount of starch escaping digestion and recovered in the effluent (ranging from 5.7% in baked beans to 0.7% in ground rice) was compared with results obtained by using the in vitro assay with an incubation time of 6 h. The assay was found to be a good qualitative predictor (r = 0.90, P < 0.05), but a poor quantitative predictor of RS amounts. Increasing the duration of incubation with alpha-amylase and amyloglucosidase to 15 h resulted in a very close agreement between results from the in vitro technique and the in vivo model. These data indicate that for a variety of foods the in vitro assay produced an excellent estimate of starch escaping digestion in the small intestine of humans.