A population-based telephone survey of acute gastroenteritis (AG) was conducted in Hong Kong from August 2006 to July 2007. Study subjects were recruited through random digit-dialling with recruitments evenly distributed weekly over the 1-year period. In total, 3743 completed questionnaires were obtained. An AG episode is defined as diarrhoea ⩾3 times or any vomiting in a 24-h period during the 4 weeks prior to interview, in the absence of known non-infectious causes. The prevalence of AG reporting was 7%. An overall rate of 0·91 (95% CI 0·81–1·01) episodes per person-year was observed with women having a slightly higher rate (0·94, 95% CI 0·79–1·08) than men (0·88, 95% CI 0·73–1·04). The mean duration of illness was 3·6 days (s.d.=5·52). Thirty-nine percent consulted a physician, 1·9% submitted a stool sample for testing, and 2·6% were admitted to hospital. Of the subjects aged ⩾15 years, significantly more of those with AG reported eating raw oysters (OR 2·4, 95% CI 1·3–4·4), buffet meals (OR 1·8, 95% CI 1·3–2·5), and partially cooked beef (OR 1·8, 95% CI 1·2–2·7) in the previous 4 weeks compared to the subjects who did not report AG. AG subjects were also more likely to have had hot pot, salad, partially cooked or raw egg or fish, sushi, sashimi, and ‘snacks bought at roadside’ in the previous 4 weeks. This first population-based study on the disease burden of AG in Asia showed that the prevalence of AG in Hong Kong is comparable to that experienced in the West. The study also revealed some ‘risky’ eating practices that are more prevalent in those affected with AG.