Mortality statistics are employed in comparing populations of health status. They are derived from individual death certificates, yet causes of death are subject to errors. As a case in point, Cardiovascular Disease (CVD) death determination is characterized by sensitivity (SE) and specificity (SP) are below 85 %. Moreover, differential misclassification may be entailed in homogenous target populations. We examine the subjectivity of standardized mortality ratios (SMR) with the aid of realistic data.
SMR values are always subjective because of the diagnostic text character of specifying death. In most of epidemiological studies, bias ought to be directed to the null-hypothesis (non-differential misclassification). Nevertheless, caution is necessary regarding differential misclassification, possibly experienced in studies on homogenous subgroups, and in large prospective cohorts with trained personnel in particular.