Parsing Icd-10 With Owl-Api

Question

[Alexis]

hi everyone,

I would like to parse the ICD-10 by Owl-api because it is an ontology and i dont know how!Could someone give me some code about this because its difficult for me!

Thanks!!!

Answer 1

[Joachim]

Hi!

I hope you are not using the OWL ontology from https://dkm.fbk.eu/index.php/ICD-10_Ontology, because that is really just an ontology of the ICD-10 codes.

ICD itself is only released by the WHO under certain conditions, which means that the classification is not publicly available as such. However, you can get it anyway from http://www.who.int/classifications/icd/en/ by reading through their contract. In a nutshell: as long as you do not monetize it, you just create a user account at WHO, agree to the terms and conditions, and then download the classification data in ClaML or plain text files.

• ClaML: http://apps.who.int/classifications/apps/icd/ClassificationDownloadNR/ClaMLFormat.htm
• plain text: http://apps.who.int/classifications/apps/icd/ClassificationDownloadNR/ICD10MetaDataFormat.htm

Both files are not going to help you very much with using an OWL API though, but you might find that parsing the plain text files will be very easy. For example, a line in the code description file looks like this:

4;T;X;01;A00;A04.0;A04.0;A040;Enteropathogenic Escherichia coli infection;006;4-002;3-003;2-003;1-004

If you want to figure out all available titles, you could simply run the shell command:

cut -d ';' -f 9 codes.txt

For more elaborate operations (That is why you are asking for using an OWL API, right?) you will either have to reformat the files or load them into your own data structures.

Alternatives might be

• http://snob.eggbird.eu/manual/claml.html.
• http://smi-protege.stanford.edu/repos/protege/icd/trunk/src/edu/stanford/bmir/icd/claml/ClamlImport.java

I have not used these myself, but at least they seems to be able to understand ClaML and the latter seems to actually create an OWL ontology.

Hope that helps,

Joachim

Answer 2

[Shimcode]

Here's the regular expression for ICD-10:

http://regexlib.com/REDetails.aspx?regexp_id=2276&AspxAutoDetectCookieSupport=1