Diagnostic validity of fatal cerebral strokes and coronary deaths in mortality statistics: an autopsy study

Area of Science:

• Epidemiology
• Public Health
• Medical Statistics

Background:

• Mortality statistics are crucial for epidemiological studies.
• Accurate cause-of-death data is essential for public health surveillance.
• Previous studies have indicated potential discrepancies in death certification.

Purpose of the Study:

• To assess the diagnostic validity of cerebral stroke and ischemic heart disease as underlying causes of death in Norwegian mortality statistics.
• To compare mortality data with autopsy findings to determine accuracy.
• To identify factors influencing the selection for post-mortem examinations.

Main Methods:

• Utilized mortality data from the Bergen Clinical Blood Pressure Study (1965-2005).
• Compared registered causes of death with autopsy records from the Gade Institute.
• Calculated sensitivity, positive predictive value, and Cohen's Kappa for diagnostic agreement.
• Analyzed predictors for undergoing post-mortem examination.

Main Results:

• Substantial agreement was found between mortality statistics and autopsy findings for both fatal strokes (Kappa=0.78) and coronary deaths (Kappa=0.80).
• Sensitivity for stroke was 0.75 and for coronary deaths was 0.87.
• Positive predictive values were 0.86 for stroke and 0.85 for coronary deaths.
• Cerebral stroke was negatively associated with autopsy selection (OR=0.69), while coronary heart disease was not (OR=1.14).
• Autopsy selection was associated with age, gender, and cause of death.

Conclusions:

• Norwegian mortality statistics demonstrate substantial diagnostic validity for cerebral stroke and ischemic heart disease.
• Autopsy findings largely corroborate the registered causes of death for these conditions.
• Factors such as age, gender, and specific cause of death influence the likelihood of an autopsy being performed.