Keywords: Primary care, Death certificate, Fragility fractures, Diagnostic discrepancy, Old women

Background:

Cause of death certificates (CoDCs) are critical for health statistics, resource allocation, and epidemiological research. In Sweden, as in many countries, most CoDCs for older persons are completed by general practitioners (GPs). However, several diagnoses with significant clinical impact—such as fragility fractures, anemia, and dementia—are often underrepresented in official mortality data.

Research questions:

To examine discrepancies between clinical records and CoDCs in deceased women aged 75—103, focusing on conditions commonly associated with frailty and ageing.

Method:

Data from a retrospective cohort of 893 deceased women in a southern rural Swedish setting were analyzed. Clinical diagnoses were retrieved from an electronic health care record system, categorized per ICD-10 and then compared with listed causes of death. Fragility fractures were analyzed in relation to time of death. The proportion of GPs writing CoDCs and autopsy rates were also assessed.

Results:

GPs completed 78% of all 893 CoDCs. Anemia and dementia were frequently recorded in clinical data but often absent in CoDCs. Among individuals with hip or pelvic fractures, 85% of 34 cases were reported in death certificates when the fracture occurred within 30 days prior to death. In contrast 8% of 77 cases were reported when the fracture occurred between 31 and 365 days before death. Autopsy rates were 2%.

Conclusions:

Significant discrepancies existed between clinical diagnoses and reported causes of death, especially for frailty-related conditions. These gaps may compromise the validity of mortality data. Enhancing training and support for GPs in cause of death reporting could improve data quality and public health planning.

Points for discussion:

How can we better support GPs in accurately certifying causes of death, especially for frailty-related conditions like fractures, anemia, and dementia?

Should the time interval between clinical events (e.g., fractures) and death influence whether they are included in death certificates?

What are the implications of diagnostic underreporting in primary care-certified death certificates for health policy, resource allocation, and epidemiological research?