Abstract
Much of what I have read about epidemiology of diabetes in the past 20 years has merely refined what my Aunt Nina said about her non-insulin dependent diabetes. We have learnt more about its causes and have more precise estimates of the relative risks of developing it as a function of age, obesity, level of physical activity, and, from Rimm and colleagues' paper in this week's journal, cigarette smoking and alcohol use (p 555).2 But substantial advances in our understanding of its epidemiology and that of other non-communicable diseases have been rare.