International comparability of mortality statistics
31/01/2017 12:30/ Author: Masja de Ree/ Photography: Hollandse Hoogte/ Category: International developments
In November 2016, Statistics Netherlands (CBS) became member of the IRIS Core Group. IRIS is an international software system for the automated encoding of causes of death. With this membership, CBS is able to exert direct influence on the quality of the system and achieve better international comparability.
Differences in interpretation
Cause-of-death statistics are the oldest health statistics in the Netherland and have existed since 1901. According to Peter Harteloh, statistical researcher at CBS and a physician, ‘It means these statistics are eminently suitable for trend analysis.’ Up to 2012, the forms on which medical practitioners recorded the cause of death of a patient were processed manually at CBS. As of 2013, processing is automated using the IRIS software system. Harteloh: ‘This saves time and forestalls different interpretations among encoders or among different countries. Data are better comparable internationally. That is interesting as it also becomes easier to relate data to differences in risk factors, lifestyle and health care among the various countries.’ IRIS is also being used in Belgium, Luxembourg, France, Germany, Czech Republic, Sweden and the United States. Since 2011, it is placed with the Deutsches Institut für Medizinische Dokumentation und Information (DIMDI) in Cologne, Germany.
CBS joined the IRIS Core Group in November 2016. As a member, CBS can make an immediate contribution to quality enhancement of the IRIS system. ‘Encoding causes of death in a medically correct way can be very complicated’, says Harteloh. ‘As a member of the Core Group we will be able to raise or respond to discussion points in the software more easily.’ IRIS encodes all medical terminology on death certificates and subsequently selects an underlying cause of death which initiates the ‘causal chain of events’ leading directly to death. ‘The choices made by this system have direct consequences for statistics’, Harteloh explains. ‘Since working with IRIS we have seen dementia become the underlying cause of death in up to 25 percent more cases. We want to be able to fine-tune such information. In some cases it may make more sense to select the pneumonia that developed during dementia as underlying cause of death. We already do this, but within the Core Group we can start working towards more international consistency.’
One of the tasks for CBS in the Core Group will be to test prototypes of IRIS’ annual software updates. CBS will also be looking at decision guides used by IRIS to select causes of death. An additional advantage of CBS’ membership is that it boosts the medical expertise within the Core Group: both CBS representatives, Harteloh and his colleague prof. Jan Kardaun, are physicians. Harteloh: ‘We can act as advisors in medical or encoding-related issues.’ CBS has also designated its programmer Suzan van der Aart. ‘We have already accumulated a host of experience in the comparison of different versions of this system and in evaluating the guides. This is the time when our contribution becomes official. It is an honour and a form of recognition. IRIS is internationally recognised as a standard for encoding and selecting causes of death. This international standard is being developed by only a few countries, now including the Netherlands. It means that the competence, expertise and integrity of CBS enjoy a lot of confidence.’ With this membership, CBS hopes to provide an international boost to automatic encoding of death causes, thus contributing to the quality of the software as well as the quality and international comparability of mortality statistics. Harteloh: ‘Our aim is to have new users of IRIS in as many countries as possible.’