yehster wrote on Thursday, December 15, 2011:
Just brainstorming. Had a crazy idea as a feature for fee sheet/icd 9 search. Was thinking to pass the search string on to google using cURL. Could parse the diagnosis codes and match them up the data in the codes table. Would do a google search and restrict it to site:www.icd9data.com, parse out by regular expressions the anchor tags with “Diagnosis Code *.*” and viola.
Throwing this out there as a cool thing to do.
bradymiller wrote on Thursday, December 15, 2011:
Hi,
Sounds pretty cool. Would rec making it an option, since some users probably use OpenEMR without an internet connection.
-brady
blankev wrote on Thursday, December 15, 2011:
Using something like the Brady translation sheet? Could this be an alternative?
Contra: Many unused coding for the different specialisms
Pro: easy implementation
Somewhere on the Internet there should be a comparison table.
If the cURL comparison is something with only once a year implementation it is a great idea. The ICD codes do not change every day. More once a year or even less frequent.
Just an other opinion of thoughts………
Pimm
mike-h30 wrote on Thursday, December 15, 2011:
That sounds interesting. However the inefficiency with the fee sheet (as reported from my users) is having to enter the diagnosis codes repeatedly for every date of service for the same patient. Most patients are seen monthly and some twice a week that are receiving physical therapy. So there is a lot of redundancy entering the same diagnosis codes. A feature to clone the previous fee sheet might be a solution or having a central location to enter diagnosis codes (like EZ Claim billing software) one time have been requested to me.
Mike
yehster wrote on Thursday, December 15, 2011:
Mike,
My feature idea was inspired by folks being frustrated trying to find obscure/infrequently used codes.
Cloning the fee sheet addresses a different problem, that of frequently used codes being much harder to access than they could be.
We could also present the most commonly used codes from the billing table on the fee sheet screen which would save time for new patients as well.
Neither of those would be too difficult to implement. At the risk of receiving the ire of others for soliciting, if folks would be willing to pay, I would be willing to do the work.
-Kevin Yeh
kevin.y@integralemr.com
mike-h30 wrote on Thursday, December 15, 2011:
My current work around was that I had created a custom form to store the diagnosis codes for the patient. Then I modified the fee sheet to query the table for that form to display the diagnosis codes when filling out a new fee sheet. This is not a long term solution but merely a temporary work around. I had not spent too much time on modifying the fee sheet due to mention of a code re-factor post certification.
If work is to be done on the fee sheet, perhaps a new thread with proposed enhancements by other users could be discussed before moving forward with a permanent solution.
Mike
sunsetsystems wrote on Thursday, December 15, 2011:
At the risk of receiving the ire of others for soliciting, if folks would be willing to pay, I would be willing to do the work.
The way I usually put this is that things happen when someone is willing to do or sponsor the work. Sometimes users of open source software forget that there is no unpaid team of volunteers working on their list of enhancement requests. 
jason0 wrote on Thursday, December 15, 2011:
Howdy Kevin,
Would the lookup operation occur on the server, or client? I ask because I am starting to look at profiling the apache server, and using curl might cause an apache process to block waiting on a response from another site.
-jason
yehster wrote on Thursday, December 15, 2011:
Jason,
Without using a plugin like greasemonkey to enable cross site scripting, the lookup operation would have to occur server side. I don’t think you could merge the google results with the OpenEMR ui with normal client side javascript.
The query could be done as a seperate ajax request though, and while that operation would be affected by the response time from google, the impact to other processes would be negligible.
What are your plans and goals with profiling?
jason0 wrote on Thursday, December 15, 2011:
Hi Kevin,
That’s a good question: one that I am still forming.
My perspective is from the server’s operations: in other words, mysql efficiency, memory usage, apache processes, how php uses adodb connections, etc. Right now, I have > 20 installations of openemr running on one server. Each has one or more “site” on it. Not all are used by customers, but they certainly present a large load on our mysql server.
At the moment, I am working on how to benchmark the mysql server. Eventually I would like to get into the database schema and perhaps how openemr actually does queries.
Actually, it would be better to say “I am learning what questions to ask”. I would like to know your thoughts on this.
-jason