The vital registration system in Mexico relies on information collected from death certificates to generate official mortality figures. A study by researchers at IHME and the National Institute of Public Health in Mexico set out to test the validity of this system.
While Mexico’s vital registration system is rated among the best in terms of quality and completeness, the quality of information provided in medical death certificates varies according to the people responsible for completing them, whether physicians or lay people, as well as if deaths are coded as ill-defined. The objective of the study was to assess the concordance between the cause of death from official statistics from death certificates and a gold standard diagnosis of the same deaths from hospital medical records. The gold standard causes of death used in this study were collected as part of the Population Health Metrics Research Consortium (PHMRC) gold standard verbal autopsy (VA) validation study.
The researchers studied both chance-corrected concordance, which assesses the accuracy of individual cause of death assignments, and cause-specific mortality fraction accuracy, which assesses the fractions of all deaths due to a specific cause in the population. When considering only the underlying cause of death on the medical death certificate, median chance-corrected concordance between the medical death certificate versus the gold standard was 54.3% for neonates, 38.5% for children, and 66.5% for adults. Cause-specific mortality fraction accuracy for the medical death certificate versus the gold standard was 0.756 for neonates, 0.683 for children, and 0.780 for adults. Both median chance-corrected concordance and accuracy increased when all causes of death on the medical death certificate were included, not just the first underlying cause.
Concordance varied depending on the cause of death. For example, diabetes was more often recorded in the medical death certificate than the gold standard, suggesting that physicians are overstating this cause in the medical death certificate. Misclassification of deaths was also seen in diarrhea, pneumonia, burns, lung cancer, falls, and poisonings, while AIDS, cervical cancer, and leukemia/lymphomas had very little misclassification.
The study examined 1,589 deaths that occurred in 34 public hospitals in the Federal District and the state of Morelos, Mexico in 2009. The underlying cause of death obtained from medical death certificates was compared to a gold standard diagnosis derived from a review of medical records developed by the PHMRC. The gold standard causes of death required review by a committee of physicians and documented diagnostic criteria, including laboratory tests, medically observed and documented illness signs, and medical records containing specific terminology of illnesses. As part of the PHMRC study, VA interviews were conducted with relatives of deceased persons whose diagnoses were classified as gold standard.
Since the medical death certificates contain a sequence of causes of death, the authors were able to estimate concordance for the first underlying cause of death, as well as all causes of death recorded in the death certificate. Chance-corrected concordance and cause-specific mortality fraction accuracy were used to evaluate the quality of performance of the death certificate.
Accurate information about mortality is critical for tracking health and implementing policies in populations. Reliable statistical mortality data are therefore needed, which depend on the completeness and accuracy of cause of death diagnoses on the medical death certificate. While the study found a reasonably high concordance and accuracy of the assignment of individual causes of death from the medical death certificates compared to the gold standard, concordance varies depending on cause of death, and low concordance and cause-specific mortality fraction accuracy were found in both child and neonatal deaths.
The results suggest that using all of the causes of death recorded on the medical death certificate provides better population mortality data compared to only using the first underlying cause of death from the medical death certificate. In addition, death certification procedures need to be improved in children and neonates to better target health policies to these age groups.