Aggregate disease risk-factor data: is the population mean a valid parameter? The example of serum cholesterol in Oppland county, Norway

Abstract

The arithmetic mean is the most commonly used summary measure for continuous variables in regional studies. However, the mean, or changes in the mean, could be misleading, because these measures do not take into account the underlying distribution of the variable in the population. In this study, disease risk-factor data from Oppland county, Norway, are used as an empirical example to discuss this problem. Serum-cholesterol changes are described for regions with different mortality time trends for ischacmic heart disease. The mean turns out to be a relevant comparative measure in this study. However, one should generally be aware of the possibility of distribution changes. Under certain circumstances important changes of the relative size of the tails of the distribution could occur without influencing the mean.