Objective: To review evidence for the efficacy of biofeedback in the treatment of gastrointestinal problems. Data Sources: Studies were identified through a search of MED-LINE, HealthSTAR, EMBASE, Allied and Complementary Medicine, MANTIS, PsycINFO, Social Science Citation Index, Science Citation Index, and CINAHL. Study Selection: Studies were selected if they used biofeedback as the intervention, addressed the treatment of a gastrointestinal condition, and included a control group. Data Extraction. All titles (4397), articles, and/or abstracts (1362) were evaluated by 2 independent reviewers who extracted data on study design and quality, sample size, type of intervention, and outcomes. Data Synthesis: We found 16 controlled trials of biofeedback for gastrointestinal problems. Ten studies had a comparison group that did not receive biofeedback (5 studies on constipation and/or encopresis, 3 on fecal incontinence, 1 on constipation, and 1 on abdominal pain). Of the 10 studies that had a "no biofeedback" control arm, 7provided sufficient data to calculate an effect size. Of these 7 studies, 2 favored biofeedback: I study revealed a reduction in fecal incontinence among adults, and the other study showed an improvement in constipation with fecal incontinence (encopresis) among children. The other 5 had nonsignificant results. Three studies had a "no biofeedback" comparison arm but insufficient data to calculate an effect size. One of these studies in adult fecal incontinence reported statistically significant improvement. Conclusions: The evidence is insufficient to support the efficacy of biofeedback for these gastrointestinal conditions.