Methods and tools to evaluate the quality of vital statistics
Classification of unusable ICD codes for mortality coding
There is no standard list of ICD codes to avoid, as it depends very much on the intended use of the data. There is no standard name used for such codes, for example, they are commonly referred to as: ill-defined, R-codes, garbage codes, unusable codes or insufficiently specified. Although all lists are based on the ICD, they differ in the number of codes they include on their lists. Therefore, countries that wish to compare themselves to other countries, or earlier assessments of their country, need to make sure that their comparators are based on the same classification, or incorrect conclusions will be drawn.
What list of unusable codes should be used? The most important characteristic to consider is that the criteria for inclusion are clearly stated, and that the list comes from a respectable source, a peer-reviewed journal or an agency that can be cited. Knowing the basis on which the list is composed will clarify whether it will serve the intended purpose.
ANACONDA uses the same garbage code list as the Global Burden of Disease study and groups the individual causes into two different typologies. The first list is based on ICD concepts and the causes are grouped according to the type they belong to and, in the second one, they are classified by the impact they have on the overall distribution of COD (see table below). In other words, causes that can be totally misleading are considered to have a ‘very high impact’, such as septicaemia, cardiac arrest or heart failure. None of these would allow one to know whether the real underlying cause was a communicable or noncommunicable disease. By contrast, causes such as unspecified pneumonia or stroke would be in the ‘low impact’ category, as one would at least know that they belonged to a specific group of related diseases.
For the specific ICD codes in each category, please see ANACONDA.
Guidance for assessing and interpreting the quality of mortality data using ANACONDA
For users of ANACONDA (statisticians and/or analysts in health and statistics departments, researchers, or other experts working with mortality data).
Authors: Mikkelsen L, Lopez AD
Publication date: October 2017
Resource type: CRVS resources and tools
Related resources: Course prospectus: ANACONDA