Cohort studies are the preferred design in observational epidemiology, but few involving the general population have been performed in Asia, and most concern affluent urban populations. The Khon Kaen study has recruited about 25,000 subjects, aged mainly 35-64, from villages in the relatively underdeveloped north-east of Thailand. All subjects underwent simple physical examination, completed an interviewer-administered questionnaire (including sections on lifestyle, habits, and diet) and donated specimens of blood, which were processed and stored in a biological bank at -20 degrees C. Female subjects (about 16,500) were offered screening by Pap smear, and specimens of cells from the cervix were stored at -20 degrees C. This paper describes the methodology of the study, and the characteristics of the participants. Almost all subjects are peasant farmers, with low annual income and body mass, although 14.6% of women had a BMI in the obese range (>30 kg/m(2)). Smoking was common among men (78% regular smokers, most of whom used home-produced cigarettes), but rare among women. Fertility levels were relatively high, with a more than half the women having four or more live births. 23.4% of subjects were infected with the liver fluke Opisthorchis viverrini, known to be highly endemic in this region. Follow-up of the cohort is by record-linkage to the provincial cancer registry. By 2003, 762 cancer cases had occurred, the most common being cancers of the liver (363 cases) and cervix uteri (44 cases). The antecedents of these cancers are being investigated using a nested case-control approach. The cohort will yield increasing numbers of cancers for study in the next decade, giving important information on the relative importance of dietary and lifestyle factors in a rural population, undergoing gradual transition to a more westernised lifestyle.