Notifications
Clear all
Topic starter
When you use the Smart List to select a problem, we are seeing no call to the Centricity database to populate that list. So where does this data come from? We've tried to figure this out for the Data Mapping screen in Administration, and the Reference List lookup as well. Is it an Internet lookup?
Posted : June 8, 2015 4:12 am
I like how over 40 people viewed this but nobody responded.
Posted : June 8, 2015 3:21 pm
Topic starter
GE says:
The CPS problem search utilizes the Solr Search and Lucene search engine.
At a high level, indexes are created from CPS tables and it’s that indexed data that used to search for problems using the Smartlist. That’s why you’re not seeing calls to the database when doing a problem search. These indexes are built when the Jboss service starts.
So it’s not a CPS db table that’s queried when a Smartlist is used. The Solr/Lucene engine utilize the indexes generated to query and return search results.
So now they just need to tell me where the data comes from to populate that index. Personally, I've never heard of the Solr/Lucene engine, and certainly didn't know it was running on our application web server.
Posted : June 9, 2015 2:06 am
On the JBoss server under jboss\server\default\deplay\SID-ml.ear\ there is a Solr directory, I bet the information is in one of the xml files there.
Posted : June 9, 2015 3:10 am
@rfisher, apparently I, along with 39 other people, were also interested in the answer but didn't know it. I viewed the thread to enable notifications. I'm glad I did because I just gained valuable information.
Thanks Ron & David. I'm going to wander through that directory now.
Posted : June 9, 2015 3:27 am
Topic starter
Yeah, there's definitely something going on in there. Multiple subfolders called "index" under directories partially named after our database... yep, that seems like a clue. But I don't see any recognizable file types in there. I doubt there's anything I can work with directly.
Just to take a step back...the reason I'm investigating this is that we have auditors telling us that problem entries in our patients' charts must conform to the standard ICD-9 descriptors, no custom entries. This is news to us, as we've been allowing customization since day one, and GE encourages this. Cleaning up the custom problem lists is a piece of cake (just an update of ICDDEPT from MasterDiagnosis). But we still have all these entries in the Smart List, so those need to go as well.
I've discovered that the Smart List entries appear to all be in the CommonDescription column of the ProblemDescriptionMap table. I emphasize appear because obviously I haven't searched for every entry. But after looking up a few dozen oddball descriptions, I'm feeling pretty confident. Now all I need GE to tell me is...
1. Is this table the source of the indexes?
2. Will cleaning up this table cause the indexes to refresh themselves?
3. Will more unwanted entries appear in this table over time, and if so, how?
4. What other potential impact could there be to tinkering with this relatively unknown table?
Posted : June 9, 2015 3:28 am
I've found Solr is built on top of Lucene. http://wiki.apache.org/solr/FrontPage may have some useful info.
I attacked two .csv files with the list of files/folders that contain the name lucene and solr if that's useful to anyone.
Posted : June 9, 2015 4:09 am
Great thread, very informative - I'd love to know if you can find out any more from GE. Thank you for posting.
Mike
Posted : June 10, 2015 5:23 am
Topic starter
We have some answers from GE, and they're pretty much what I'd already figured out:
Q1: What SQL tables are Solr indexes created from? We suspect ProblemDescriptionMap, but need confirmation.
The SOLR indexes are created use the ProblemDescriptionMap and multiple other tables such as MostUsedProblems , MasterDiagnosis and Diagnosis .
Q2: What is the source of the data in ProblemDescriptionMap, or whatever table populates the indexes?
The source of data for ProbleDescriptionMap and other problem search related table is the Full KB update process
Q3: Will altering the data in this table cause the indexes to refresh themselves?
Altering data in the tables will not cause the indexes to refresh , I would caution against doing this as we generate an automated ID for each row as part of the KB update process. It would break any index relationships with the other tables used by SOLR
Q4: If we do alter the data in this table, will additional entries appear in this table over time, and if so, how?
Altering the table would cause it to get wiped out when the next full KB is applied . The roadmap for our Problem Search Capabilities, includes a feature to allow customers to Add/Edit the SmartSearch, this is currently scheduled for CPS 12.2 and EMR.9.8
Posted : June 11, 2015 7:51 am
I really just want to purge the smart list before moving to ICD10, because we are getting back ICD9 results first in the list when searching. I'd rather the providers start with a blank slate than confuse them with ICD9 codes. Would removing all items from these tables do that? I may try on a test database and see if I get any errors.
Posted : July 24, 2015 9:12 am
Did anyone ever get a better understanding of the ProblemDescriptionMap table and how to get Athena to update?
We have found entries we believe to be wrong, but cannot figure out how to go about getting updates in the monthly KB's. The entries in question we disagree with the wording of the problem description.
Details of the offending entries we see can be seen in this post
https://forum.centricityusers.com/forum/problem-smartlist-search/
Thanks
Davy
Posted : April 29, 2021 9:06 am
