Estimation of dynamics of risk factors by the dynamic regression method

A problem is considered for the estimation of dynamics of risk factors and other indicators of health by the data of a number of population studies performed in various years on the same age category of the population. In the case of the nonlinear time dependence of mean values of the quantities under study, the simple interpolation of the indices by all groups leads to incorrect estimates in view of the neglect of the cohort dynamics. A methods is described for the development and identification of the dynamic regression model of the population health, which is based on the estimation of the cohort dynamics of indices. This makes it possible to prognose the expected lavels of risk factors and to clarify causal relations. The efficiency of the model is demonstrated by an example of the processing of results of the examinations performed in the years 1982, 1987, and 1992 in North Karelia (Finland).