We studied variation in the year of onset of ischemic heart disease mortality decline among regions (n = 39) in the Netherlands. Using loglinear regression methods, a quadratic regression model was fitted to the observed numbers of male deaths in each region in the period 1950-84, controlling for changes in age-structure of populations. The quadratic regression model proved inadequate to describe the mortality experience of females. For the country as a whole, the estimated year of onset of male ischemic heart disease mortality decline is 1973.9. The difference between the earliest and the latest region is almost nine years (1970.0 and 1978.9, respectively). An early onset of decline (less than or equal to 1972) is only found in the urbanized, western part of the country. A later (greater than or equal to 1975) onset of decline is characteristically found in more peripheral regions in the South-West and South-East, as well as in the North. Exploratory correlation and regression analyses show that both average income and percent of population living in larger cities have independent, negative associations with the year of onset of male ischemic heart disease mortality decline. We argue that regional variation in the timing of lifestyle changes is a more plausible explanation of these observations than regional variation in the timing of medical care improvements.