Variation in Death Certification Practices in Dementia

Mortality statistics, which are derived from death certificates, are frequently used in palliative care research to examine where people die. However, the usefulness of using mortality statistics to examine place of death depends on their accuracy. There is some evidence that death certification practices vary depending on the setting in which the person dies. For example, in the USA “dementia” is more likely to be written on a death certificate if the deceased dies in a care home than a hospital. For analyses of place of death, unknown variation in certification practice by setting will produce misleading results. In addition to variation in death certification practices by setting, it is likely that certification practices have changed over time. In England, certification of “dementia” on death certificates increased from 7% of all deaths among people aged over 60 in 2000 to 12% in 2010. Without knowing how certification practices have changed over time, it is difficult to interpret time trends in where people die. This study aims to use CRIS data linked to Office for National Statistics (ONS) mortality data to investigate the variation in death certification practices for people with dementia by place of death and year of death. The study population will be identified from CRIS as having a diagnosis of dementia, and having died. Linked ONS records will be used to ascertain whether or not “dementia” was given as a cause of death. Analyses will investigate the variation of death certification of dementia by the setting in which the patient died, and the year of death. This study will provide a better understanding of the accuracy and variation in death certification practices in dementia. The results will inform ongoing analyses of where people with dementia die, and allow more robust recommendations for policy and practice.