Under Reporting of Dementia Deaths on Death Certificates: A Systematic Review of Population-based Cohort Studies

Abstract

The purpose of this review is to assess the extent to which dementia is omitted as a cause of death from the death certificates of patients with dementia. A systematic literature search was performed to identify population-based cohort studies in which all participants were examined or screened for symptoms of dementia with a validated instrument followed by confirmation of any suspected cases with a clinical examination (two-phase investigation). Data were extracted in a standardized manner and assessed through the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) initiative. Seven studies met the selection criteria. These were from the Americas (5 articles: 2 from Canada, 2 from the US, and 1 from Brazil) and Europe (2 articles: 1 from the UK and 1 from Spain). Each met at least 83% of the STROBE criteria. The reporting of dementia on death certificates was poor in these 7 studies, ranging from 7.2%-41.8%. Respiratory or circulatory-related problems were the most frequently reported causes of death among people who were demented but who were not reported as demented on death certificates. The use of death certificates for studying dementia grossly underestimates the occurrence of dementia in the population. The poor reporting of dementia on these certificates suggests a lack of awareness of the importance of dementia as a cause of death among medical personnel. There is an urgent need to provide better education on the importance of codification of dementia on death certificates in order to minimize errors in epidemiological studies on dementia.

International Variation in Place of Death of Older People Who Died From Dementia in 14 European and non-European Countries

“…27 Dementia as cause of death is known to be underreported; mild dementia has a lower probability of being listed and dementia might not be recognized as a cause of death when other (immediate) causes of death are present. 28 However, by including only deaths of people whose underlying cause of death was dementia, (ie, those who died from dementia and thus were probably in an advanced stage of their dementia-related condition at the time of death), 28 bias related to death certification may be limited. Nevertheless, it has to be noted that diagnostic or reporting differences might exist between countries or between different health care settings within one country.…”

“…Nevertheless, it has to be noted that diagnostic or reporting differences might exist between countries or between different health care settings within one country. 28 Furthermore, not all variables known to affect place of death can be included in this study, as death certificates do not include important individual-level information, such as preferences and decision-making processes. 26,27 However, the use of death certificate data to study such cross-national variation has proved to be useful in other studies and makes whole country comparisons possible.…”

Cause-Specific Mortality Trends Following Total Hip and Knee Arthroplasty

“…Dementia is typically under-coded in death certificates, and it is possible that the coding of dementia in our community is more complete than the rest of the country. [18] Elevated mortality from dementia was recently reported in the osteoarthritis literature in a cohort of 1163 osteoarthritis patients from United Kingdom and a large study from Taiwan. [19, 20] The reasons for this finding are unknown, but we speculate that this may potentially be driven by the concomitant risk factors for dementia patients with arthritis treated with arthroplasty, such as obesity, low physical activity, or low grade systemic inflammation associated with advanced osteoarthritis.…”

Long sleep duration in elders without dementia increases risk of dementia mortality (NEDICES)

“…In addition, dementia may be omitted as a cause of death from the death certificates of patients with known dementia in life. 34,35 Such omissions, however, are equally likely for all sleep categories. Finally, our evaluation of depressive symptoms was not complete and we may have underestimated depression, resulting in residual confounding.…”