Objective: To examine whether characteristics of very small living areas can be used to predict disease incidence and to use these characteristics to assess socioeconomic differences in stroke incidence in the Netherlands. Design: Characteristics of postcode areas of stroke patients are compared with characteristics of postcode areas of all individual people in the study region, using Poisson regression analysis. Setting: Six provinces of the Netherlands, covering about half of the country. Patients: 760 patients who in 1991 or 1992 were consecutively admitted because of stroke to 23 Dutch hospitals. Main results: Stroke incidence is significantly higher among people living in postcode areas with below average socioeconomic status (relative risk=1.27; 95% confidence intervals 1.08 to 1.51) and among people living in postcode areas with predominantly older inhabitants (RR=3.17; 95% CI=2.29 to 4.39). It is also significantly increased in more urbanised areas compared with the countryside, the highest incidence being found in the large cities (RR=1.78; 95% CI=1.31 to 2.44). Conclusions: A clear socioeconomic gradient in stroke incidence in the Netherlands is observed, with people living in detailed postcode areas with below average socioeconomic status experiencing a significantly higher risk of stroke. The analysis also confirms that characteristics of detailed postcode areas can effectively be used to differentiate between areas with and areas without stroke patients.