Dear Editor,
With great interest we read the article by Oltean et al,1 "Use of Death Records to Augment Notifiable Conditions Reporting in Washington State."
To extract data about the presence of a notifiable condition from the death certificate appears to be a promising application. Especially when using electronic death certificates, contacting the issuer to verify a rare condition is relatively easy, and these data can be used as an additional data source for public health surveillance. A combination of data from different sources may lead to under- or overestimation of a certain condition. One method to prevent double counting is individual record linkage between public health registry systems and death certificate data. Another reason for overcounting can be the consideration of all notifiable conditions listed on the death certificate. Death certificates frequently contain data of disorders that started many years before death.
We found it remarkable that the authors did not mention the approximate time interval for causes of death. This time interval is the onset to death that has to be indicated at the end of each line (cause of death) of part I of the death certificate. Reliance on the words "acute" or "chronic" in the text on a death certificate needs to be augmented by considering the time interval. This is important since death certificates can contain "old conditions" as well (ie, communicable diseases that are no longer active).
The notifiable diseases that represented relatively large groups in the study of Oltean et al were influenza and the totals of different forms of hepatitis. Data from Dutch Cause of Death Statistics show that for influenza, the time interval was left blank in nearly half of the mentions of this disease on the death certificate. If the time interval was known, the interval until death varied in most cases from several hours to some weeks and exceptionally a few months. So death certificates with a mention of influenza were infections generally recently acquired.
A description of the type of hepatitis, and whether the hepatitis was acute or chronic in nature, is often missing on the Dutch death certificate. To distinguish acute from chronic hepatitis, the time interval could provide some valuable clues. Even if about two-thirds of the hepatitis terms do not carry a valid time interval, more than 10% of acute hepatitis mentions have a time interval of more than 3 years before death. This is an indication that a hepatitis designated as acute may reflect a history of acute hepatitis not necessarily reportable in the year of death.
These examples about infectious diseases show that data from death certificates should be interpreted with caution. When the time interval is known, these data sometimes can clarify whether the cause of death/notifiable condition occurred recently or many years before death.
Sincerely,
-Elisabeth M. Taal, MD, MSc
Department of Public Health
Academic Medical Center
University of Amsterdam
Amsterdam, the Netherlands
-Jan W. P. F. Kardaun, MD, PhD
Department of Public Health
Academic Medical Center
University of Amsterdam
Amsterdam, the Netherlands
Statistics Netherlands (CBS)
The Hague, the Netherlands
REFERENCE