Anybody else had any problems with load_icd_desc.plx not working anymore? It starts to run, gets to inserting code ‘001.9’, then stops, & eventually times out & fails. I’m guessing that they must have changed something on the icd9data.com site that’s causing this, but I just started looking into it.
Perhaps we should consider retiring the load_icd_desc.plx script, since there is now a set of sql scripts to import them:
sql/icd9-codes-insert.sql
or
sql/icd9-codes-replace.sql
Yes, I also had the same issue as you describe with the load_icd_desc.plx script for an installation a few weeks ago. I ended up importing the icd9 codes from the /sql directory.
Brady,
I don’t see a need for the load_icd_desc.plx script now that the icd9 codes are in the /sql directory.
CMS does make annual changes: http://www.cdc.gov/nchs/data/icd9/ICD-9-CMINDEXADDENDAfy12.pdf
So without the perl script working, there isn’t a simple way to keep the .sql up-to-date.
The current sql is almost certainly out of date, although the changes in it probably don’t matter to most people.
Keeping it up to date will sooner or later become something that has to be done, especially as ICD10 phases in.
I ended up just pulling a copy from a server built a couple of months ago when the perl loader was still working.
Yet another little kink to work out.
Doing a pretty serious amount poking around I found this site which may be of help: http://aehrc.com/med_eval/
It has 2011 ICD9 code in comma delimited format separated into diseases.csv and procedures.csv. Even better there is a SNOMED to ICD9 mapping file called icd_snomed_mapping.txt which could be very helpful.
I also found a post on stack overflow of someone that had written a parser in VB or maybe C# (I don’t speak those) that can parse the RTF files available for CMS.gov (DTab12.rtf) being the relevant one. This script cleans the code and produces a CSV file, so it says.
Is there an official free site to download the ICD9 code set in a raw format? (for example, the SNOMED, RXNORM, and ICD10 sets are all available in a raw format from an official site, which then makes it rather straightforward to use these native pcakges to import into OpenEMR and track the date/version of the release). If not, then could consider just keeping the perl script up to date and if anybody has the time, to update the sql packaged codes in sql/ directory.
Define ‘Raw’ format. I have not found ICD9 10 in anything but .rtf document format which is a complex construct. ICD9 is probably available in that same RTF format. I have not found any official free sites that proved delimited data files for either of these two code sets.
-Tony
the definitions table contains the additional descriptions for specific codes where they exist.
the codes table includes the hierarchical structure. The rows with “discr” column= “SP” (for specific) are the “terminal codes” which correspond to what is currently in the codes table. (Only stuff from the bottom of the hierarchy).
If someone wanted to update their codes table from these tables, the insert statement would be pretty straight forward.
The Zip files contain space delimited text files and a couple of xls file. One actually has the codes with both long and short text (CMS29_DESC_LONG_SHORT_DX 101111u021012.xls). This file was updated in Feb of this year.
Hello,
I read this ICD 9 post and thought of just adding it onto the way the new code works. The new code handles 1 or more zip files in a feed (i.e. RxNorm/Snomed have a single file and ICD 10 has multiple files). Either way ICD 9 raw feed will fit into this new model. The new code is somewhat generic but based on the original code. I will look for the raw ICD 9 feed on the CMS site unless someone already knows where it is (in that case please post the location of the file(s). I would prefer to use the CMS file as it is where we are basing all the other feeds on.