[Influence of methodological changes on unicausal cause-of-death statistics and potentials of a multicausal data...

As a complete survey, cause-of-death statistics are often used for research purposes, but they react sensitively to methodological changes.

To show how sensitive the signing of the underlying cause reacts to methodological changes and what potential there is in the use of multicausal analyses.

The methodological examples are based on a sample of the 2016 annual material from Bavaria (age: 65 years or older). It includes n = 24,752 cases of death with information on underlying cause and multiple causes of death. The standardized ratio of multiple to underlying cause (SRMU) value was used to investigate the extent to which dementia and Parkinson's disease are underestimated in the mortality process and which other diseases are noted on the death certificate besides these causes of death.

Changes in the set of rules, in the decision tables, and in the confidentiality concept can have an important influence on the signing of the underlying cause. This can lead to changes in the ranking of causes of death or to underestimation or overestimation of certain diseases. Dementia and Parkinson's disease are underestimated as a cause of death in the dying process if only the underlying disease is used for analysis.

Temporal and regional comparisons must always be interpreted against the background of changing methodological guidelines and procedures. Multicausal analyses in this respect offer the chance for the future to mitigate the difficulties of a unicausal approach.