Abstract
One of the main hypotheses of the World Health Organization (WHO) MONICA Project is that trends in the major coronary disease risk factors are related to trends in rates of fatal and non-fatal coronary disease events. The units of study are populations rather than individuals. The WHO MONICA Project involves continuous monitoring of all coronary disease events in the populations over a 10-year period and periodic risk factor surveys in random samples of the same populations. Estimation of associations between average annual changes in mortality and risk factor levels is illustrated with the use of data from a subset of MONICA centers. Crude estimates of regression coefficients are compared with estimates obtained by weighting for standard errors in both the outcome and explanatory variables. The results show that the strength of association may be either underestimated or overestimated if these errors are not taken into account.