by Günther Deuschl
Many political decisions within the health system are critical depending on the number of patients affected by a particular disease group and hence which discipline of medicine is responsible for the care of these patients. This applies for budgets for health care, number of beds for a particular disease group, overall budget for a discipline within a hospital, budgets for out- or inpatients care, but also percentage of research money which goes into particular fields of medical research. Also, the general public counts on these measures because mortality, morbidity and costs of disease are important to determine the importance of a disease group and usually, the higher the numbers the more important is the disease.
In this respect, it is of particular interest that special diseases are correctly grouped into the medical disciplines which take care for these patients. The internationally accepted classification tool for this is the International Code of Diseases which is currently used in its 10th version (ICD-10)[1]. In its chapter G00-G99 neurological diseases are coded, but many neurologic diseases are coded elsewhere in other chapters. This code has been criticised since its early days because it seems to systematically relate neurological disease to other disciplines.
A very recent assessment[2] on behalf of the German Neurological Society led by Richard Dodel from the Department of Neurology, Marburg University, Germany has estimated the difference between comparing ICD-10-coded ‘neurological diseases’ with a more realistic coding of disease which are currently treated by the neurologist. Health care data from different German sources were used to identify indicators (e.g. mortality, morbidity, DALYs, healthcare costs, inability to work, in-patient diagnoses, etc.) for the care in the German healthcare system. The method applied was straight-forward: Two independent neurologists classified the total of 12.333 possible single ICD-10 codes (high number because of the many subcategories, i.e. full four digit codes). 385 four-digit codes can be identified in the ICD-10 system under G00-G99. However, in their study, 734 codes were attributed to code independent neurological diseases (n=734), and if the condition needs neurological co-treatment an additional 280 codes were identified. Depending on the ICD-10 code they could compare ICD-10 with the real life situation. They compared the number of patients which had a subclassification as obvious neurological diseases or the diagnoses with a high likelihood of neurological co-treatment with the official ICD-10 subgroup ‘Diseases of the nervous system (G00-G99)’ regarding several important outcomes.
Their analysis shows that there is a profound underestimation of neurological disease when the ICD-10 codes are applied. The ranking of neurological diseases is reported here. Deaths due to neurological diseases are ranked no. 8 if only the ICD-10 G00-99, neurological diseases are taken but they rank on no. 3 if the realistic classification is used. The number of hospitalisations are classified on rank 10 according to ICD-10, G00-G99 but the realistic classification would classify them on rank 1. Regarding days with sickleave, the number increases from rank 11 to rank 3. In relative numbers deaths increase 4-fold, the days with sickleave increase 3.8 fold and the numbers of hospitalised patients increase 5.6 fold when the group of ICD-10 is replaced by the more realistic classification. The diseases which made the highest contribution to this increase were cerebrovascular diseases, headache and dementia. There are limitations of the methodological approach particularly when generalising the results to the European situation as not all countries have the same spectrum of diseases which are taken care of by neurologists, but most countries meanwhile treat stroke, headache and dementia. As these are the most important diagnoses the differences may be small but still need to be explored.
This underrating of neurology as a discipline has long been identified. The ICD-11[3] is presently updated under the leadership of Raad Shakir, President of the World Federation of Neurology and Chair of the Nervous System Disease Topic Advisory Group for ICD-11 Revision at the WHO[4]. The process is a quite complicated one and changes made today are not yet definite. As far as can be judged from the current beta-version of the ICD-11 stroke and other cerebrovascular diseases, the major part of dementia, headache and other important neurologic conditions will in the future come under the chapter of neurologic diseases[5]. The final version is still lacking and will not be effective for 2 years.
Neurologists in Europe should currently be aware that statistics when based on ICD-10 underrate the importance of neurology in terms of both important patient outcomes and the need for resources by 2-4 times.
References:
[1] ICD-10
[2] Reese JP, Ziemek J, Linnemann A, Oertel WH, Dodel R. The burden of neurological diseases in Germany. Impact of alternative classifications using the ICD-10. Marburg: Philipps-Universität Marburg, 2015 (in preparation).
[3] ICD-11 (betaversion)
[4] Mateen FJ, Dua T, Shen GC, Reed GM, Shakir R, Saxena S. Neurological disorders in the 11th revision of the International Classification of Diseases: now open to public feedback. Lancet Neurol. 2012 11: 484-485.
[5] Rajakulendran S, Dua T, Harper M, Shakir R. The classification of neurological disorders in the 11th revision of the International Classification of Diseases (ICD-11). J Neurol Neurosurg Psychiatry. 2014 85: 952-953.
