Linking to Clinicalcodes.org

[evank23]

How cool could it be if you were looking for say 'diabetes' and the software could look up the code lists available on the website and return their, say, ids along with a short description. Then the user could input the code list id to define the cohort on the fly using the info on the website

[evank23]

Have you regretted introducing me to this contraption yet?

[DASpringate]

Would be good. May have this as a feature of the website i.e. a search API which I can then link to.

[DASpringate]

I've written this R code to automatically download from the site and call into R: https://github.com/rOpenHealth/rClinicalCodes

[DASpringate]

And yes...

[rosap]

An idea would be to write a tool (or R code) where there are predefined keywords at least for common diseases (or for diseases we already have keywords, e.g. diabetes, psoriasis, MI, stroke, etc. etc). The user could have the choice to modify, add or exclude some of them.

[evank23]

I have something like that in Stata

On 25 February 2014 13:20, Rosa [email protected] wrote:

An idea would be to write a tool (or R code) where there are predefined
keywords at least for common diseases (or for diseases we already have
keywords, e.g. diabetes, psoriasis, MI, stroke, etc. etc). The user could
have the choice to modify, add or exclude some of them.

Reply to this email directly or view it on GitHubhttps://github.com//issues/14#issuecomment-36006122
.

[rosap]

Good! I can give you all the keywords I have used for several "common" diseases

[rosap]

Also, while the list of clinical codes on the website will be outdated in n years time, the keywords search will always be valid and it can be used for different types of coding system (e.g. Read codes, ICD10, etc)

[rosap]

Another idea or alternative would be a text-driven approach:

• Let's say that you can link and download a list of medical codes from the Clinicalcode website. However, you need the keywords rather than the list.
• You could download the list you need, e.g. stroke,;
• create a variable for the codes description;
• parse the description;
• remove unwanted words, e.g. prepositions (on, of, with, due, etc.)
• get the keywords.

Example with two readcodes for stroke (taken from one of the list on the Clinicalcode website):

1. "Cerebral infarct due to thrombosis of precerebral arteries"
2. "Cerebral infarction due to embolism of precerebral arteries"

Keywords will be:
Cerebral, infarct, thrombosis, precerebral, arteries, embolism

• Again the user could modify the list obtained.

[rosap]

Maybe too fancy?? :)

[DASpringate]

No, I like this idea! It wouldn't be difficult to implement either. Parsing the strings into words is a simple regex task and there is a package called tm that would do the stopword filtering to get the keywords.

[evank23]

witchcraft

On 27 February 2014 10:51, David A Springate [email protected]:

No, I like this idea! It wouldn't be difficult to implement either.
Parsing the strings into words is a simple regex task and there is a
package called tm that would do the stopword filtering to get the keywords.

Reply to this email directly or view it on GitHubhttps://github.com//issues/14#issuecomment-36230338
.

[DASpringate]

Closing this issue here and relocating to rOpenHealth/rClinicalCodes#1

This is where I am developing the R software to integrate with clinicalcodes...

[DASpringate]

Lots of things look like witchcraft to a Stata hacker... :)

[rosap]

Great!! I don't think it's very difficult to implement. Even better, if you say that there is a package in R which could filter words.
More than great it would be if we could use additional sources for lists of medical codes other that the Clinicalcode website. For example, it would be great if we could use the headings and subheadings of the MeSh terms in Ovid. However, I am not sure these are available as files.

[DASpringate]

I've done a basic implementation of this here:
https://github.com/rOpenHealth/rClinicalCodes
code:
https://github.com/rOpenHealth/rClinicalCodes/blob/master/R/keywords.R

[rosap]

Great!!! I may try it later! :)

[DASpringate]

Thanks for the idea https://github.com/rosap It will definitely need some tightening up but the basics are there. See the instructions in the https://github.com/rOpenHealth/rClinicalCodes/blob/master/README.md

[rosap]

No problem. I will have a look later on...

On Thu, Feb 27, 2014 at 11:50 AM, David A Springate <
[email protected]> wrote:

Thanks for the idea https://github.com/rosap It will definitely need some
tightening up but the basics are there. See the instructions in the
https://github.com/rOpenHealth/rClinicalCodes/blob/master/README.md

Reply to this email directly or view it on GitHubhttps://github.com//issues/14#issuecomment-36234304
.