Accuracy of the death certificate in a population-based study of asthmatic patients

Objective: To quantify the reliability of death certificate data concerning asthma.

Design: The complete medical records of decedents were reviewed by a physician certified in allergy and pulmonology who determined the cause of death without having access to the original death certificate. Disagreements between the death certificate and the reviewer were adjudicated by an expert panel.

Setting: The community of Rochester, Minn.

Patients: The mortality cohort included 339 deaths from a larger cohort of 5241 Rochester residents who received medical treatment for asthma between 1964 and 1983.

Main outcome measures: Kappa coefficients were used to measure agreement beyond that expected by chance between the reviewer and the death certificate. The sensitivity and specificity of the death certificate diagnosis of asthma were calculated against the standard of the reviewer/panel diagnosis.

Results: Death certificates reported asthma as an immediate or underlying cause of death in 22 instances (6%), whereas the reviewer/panel identified asthma in 53 cases (16%). In four cases, the death certificate listed asthma and the panel identified another cause of death. The death certificate had a sensitivity of 42% and a specificity of 99% compared with the reviewer/panel. Agreement between death certificates and the reviewer was not influenced by whether an autopsy was performed.

Conclusions: Death certificate diagnosis of asthma as an underlying cause of death had a low sensitivity but a high specificity. Increases in mortality due to asthma are not likely caused by false-positive diagnoses of asthma as an underlying cause of death. Asthma mortality rates, determined from death certificate data, may indeed underestimate actual asthma-related mortality.