OBJECTIVE: To describe the methods used in nine provincial surveys carried out as part of the Canadian Heart Health Initiative. DESIGN: Population-based cross-sectional surveys, following a core standardized protocol, implemented by provincial departments of health in collaboration with Health and Welfare Canada. Data were obtained through a home interview and a clinic visit. A standard manual of field operations and standardized training procedures were used in all provinces. SETTING: Nine Canadian provinces during the period 1986 to 1990. PARTICIPANTS: A probability sample of 26,293 men and women aged 18 to 74 years was selected from the health insurance registries in each province. Over 30% of the participants had post-secondary education. About 50% were 18 to 34 years old. OUTCOME MEASURES: Data on sociodemographic characteristics, hypertensive and diabetic status, knowledge and awareness of the causes and consequences of cardiovascular disease and two blood pressure measurements were obtained in a home interview. During a clinic visit, data were collected on height, weight (waist and hip circumferences in four provinces), two blood pressure measures and a blood sample. Total plasma cholesterol, triglycerides and high- and low-density lipoprotein cholesterol were measured in the Lipid Research Laboratory, University of Toronto and St. Michael's Hospital. MAIN RESULTS: Of the subjects invited to participate in the survey, 78% were interviewed, 69% attended the clinic and 64% provided a fasting specimen (8 hours or more). The response rates were slightly lower for men aged 18 to 34, for women aged 65 to 74 and for those with fewer years of education. Data from the provincial surveys (Ontario will complete the survey in 1992) are being compiled in the Canadian Heart Health Database. CONCLUSION: The process followed in the implementation of the provincial heart health surveys is a model of how provincial departments of health may carry out epidemiologic investigations in support of their mandate. The approach illustrates how a country-wide database can be built through partnerships among different levels of government. The use of community health nurses was instrumental in the efficient implementation of the surveys and in the realization of the relatively high rates of response attained.