Statalist The Stata Listserver

[Date Prev][Date Next][Thread Prev][Thread Next][Date index][Thread index]

Re: st: ICD9 data

Subject   Re: st: ICD9 data
Date   Fri, 23 Feb 2007 12:14:29 -0600

Fred Wolfe <> asks:

> Does anyone know how to modify the ICD9 database supplied with Stata.
> Stata appears to be a little out of date. We recently ran chceks on
> our hospital data and received errors for the codes below.But they
> are valid in the current ICD9 codebook. They are all decimal
> extensions of previously existing codes, i.e., what used to be 585
> for chronic renal failure is now 585.1 through 585.6 for stage I,
> stage II, stage III, etc and 585.9 for CRF unspecified.
> Or perhaps there is another way to handle this. Suggestions appreciated.

We at StataCorp would like to update the -icd9- and -icd9p-
commands, but have hit into some complications.  We need the
advice or assistance of those of you on statalist familiar with
the ICD-9 codes.

If in Stata you type -icd9 query- and -icd9p query- you will see
a history of where we obtained the ICD-9 data and the updates
that we applied to the data, along with a complete list of the
changes that we applied to the data.

The last substantive update to Stata's -icd9- command was in July
2005.  We obtained a nice ascii text file starting at

and then downloaded the file

which had the files we used.  However, those sites went away over
a year ago (we have copies of all the data we previously
obtained, so no worries about that) and we can not find the
latest ICD-9 data in a similar format to what we used before (a
flat ascii file).

What we do find is at

and this looks like the most official site to use.  We then click
the 2006 link to obtain the latest data and it takes us to

After downloading and unzipping the files, the two files with the
data are Dtab07.RTF and Ptab07.RTF.

Here are the problems we are encountering:

  * It is not obvious to us when viewing Dtab07.RTF and
    Ptab07.RTF in an RTF aware word processor how we would take
    the information and merge it with the data our existing
    -icd9- and -icd9p- command use.

    It is not a file that simply lists a code and the short
    description that goes with the code, and we are not competent
    to write short descriptions ourselves.  There is also extra
    information scattered around in the file: notes, includes,
    excludes, ...

  * There are so many wording changes between our existing data
    and this latest data, we do not know whether to stay with our
    current short wording or adopt new wording, and we are not
    competent to judge.

    Example:  In Stata type -icd9 lookup 001.0-.  It shows

                  001.0    cholera d/t vib cholerae

              The corresponding information in Dtab07.RTF is

                  001.0    Due to Vibrio cholerae

              Which is better?  We can't tell.

    The majority of entries have these issue.

  * There are multiple lines for many descriptions and the length
    of some (if the full thing is used) will be excessive.

    Example:  In Stata type -icd9p lookup 00.01-

                  00.01    ther ult head & neck ves

              The corresponding entry in Ptab07.RTF is

                  00.01    Therapeutic ultrasound of vessels of head and neck
                        Anti-restenotic ultrasound
                        Intravascular non-ablative ultrasound

                  Excludes:diagnostic ultrasound of:

              And then it continues with 8 more lines giving
              details of the "Excludes".

              Is "ther ult head & neck ves" adequate?  We can't

    Example:  In Stata type -icd9p lookup 15.3-

                  15.3     temp detach >1 xtroc mus

              The corresponding entry in Ptab07.RTF is

                  15.3     Operations on two or more extraocular ...

              I did not fill in the "..." above.  The description
              takes 2 lines and says: "Operations on two or more
              extraocular muscles involving temporary detachment
              from globe, one or both eyes".  This is a 104
              character long description.

    Stata's current -icd9- and -icd9p- descriptions are max
    length 25.  How should -icd9- and -icd9p- handle such long

Is there anywhere that has an official data file of codes and
SHORT descriptions for ICD9 like we previously obtained?

If not, what should happen to Stata's -icd9- and -icd9p-

If any of you have insight into this, please contact me.

Ken Higbee
StataCorp     1-800-STATAPC

*   For searches and help try:

© Copyright 1996–2017 StataCorp LLC   |   Terms of use   |   Privacy   |   Contact us   |   What's new   |   Site index