The normed mortality rate - a new measure for more valid comparisons of disease specific mortality rates

• Applicant: Susanne Stolpe, Ph.D.
• Subject Area: Public Health, Healthcare Research, Social and Occupational Medicine ; Epidemiology and Medical Biometry/Statistics
• Term: since 2024
• Project identifier: Deutsche Forschungsgemeinschaft (DFG) - Project number 547150019

Cause of death statistics are the most frequently used sources for comparative health reporting and are also applied to assess the morbidity of a population or the effectiveness of health care interventions. Our own preliminary work has shown that cardiovascular mortality rates in Western Europe are rather independent of morbidity. Cardiovascular diseases are the most common diseases in a population having the highest proportion among all causes of death. In a linear regression model, however, all-cause mortality was only slightly associated with cardiovascular mortality (both age-standardized) in Western Europe in 2018 (R²adj in men =0.30, in women <0). Conclusions from comparisons of cardiovascular mortality on the effectiveness of prevention and healthcare are therefore only of limited significance. The reason for this is that mortality data for cause of death statistics are not obtained in an objectively comparable manner. Coroners have subjective leeway when selecting the underlying cause of death. In the case of multimorbidity, several causal chains of diseases eventually leading to death are conceivable. The choice which disease among the prevalent conditions is selected as the cause of death is affected by its relevance for a healthcare system. Dementia as cause of death accounts for a different proportion of all deaths in European countries and has been increasingly recorded as cause of death for years. Its share in 2019 was 0.7% in Poland, 6% in Germany, and 11% in the Netherlands, particularly affecting the proportion of cardiovascular diseases among all causes of death which were 43% in Poland, 38% in Germany and 24% in the Netherlands. These different proportions do not necessarily rely on actual differences in population health. This interdependence is not taken into account in published comparisons of cardiovascular mortality, and the validity of the conclusions is not discussed. In order to enable more valid conclusions from comparisons of disease-specific mortality rates, the different probability with which a disease is selected as underlying cause of death – represented by its proportion among all causes of death - must be included in the mortality measure. The applicant proposes a norming of the age-standardized mortality rate to describe the mortality per 1% share of the disease of interest among all causes of death. First analyses show that this measure can enable more valid comparisons. Using the normed mortality rate, there was a significantly closer association of cardiovascular mortality with all-cause mortality (R²adj 0.53 for women and 0.81 for men). In the proposed project, the normed mortality rate will be applied to further scenarios of health reporting. These results will be compared with published data using the non-standardized mortality rate. We aim to further discuss the new measure in the scientific community.

• DFG Programme: Research Grants