| OHDSI Home | Forums | Wiki | Github |
I had an idea of making a website that allows users to initiate SQL queries against the CDM from the browser. I was able to prove the concept using the eunomia sqlite database and was thinking of building upon the project, maybe with exercises/tutorials on how to effectively query the CDM to get what you want.
Would there be any issue if I made this public as a gitlab web page? I’m not sure if there are any terms of service or something similar I would need to worry about.
Willing to assist on this effort.
We have been experimenting with using other FOSS tools to elucidate clinical intelligence opportunities with OMOP.
Please let me know.
Oh, great. I was envisioning an interface like sql zoo but, instead of sql clauses, it goes through the CDM tables and important OMOP concepts. You shouldn’t need anything other than a browser.
The current state is basically just a text box, execute button, and a result table, just enough to prove the concept. It currently lives on a private repository on GitLab. If you are interested, I could probably give you permission to collaborate. I would need to clean up a few things first though.
Is anyone having success with ChatGPT and OMOP?
I was impressed with the quality of code for questions such as:
Write a SQL query against the OHDSI CDM to build a cohort of patients with new onset diabetes
Write a SQL query against the OHDSI CDM to calculate the numerator and denominator for the HEDIS measure for Comprehensive Diabetes Care
In general, the SQL was completely accurate or only slightly off. The codes, on the other hand, were spotty. For the first one, it knew to use a code for HbA1c. For the others, it guessed wrong. This is still an accelerator, because it is easy to use Atlas to find the right set of concept codes and paste them in.
I expect that with custom ChatGPT/LLM training and/or an integration that calls the Phoebe API to get a more appropriate set of codes, it may be possible to prototype some queries via LLM prompts.
I am currently working on this very kind of accelerator for the OHDSI community. A ChatGPT-esque interface with an underlying LLM aware of the 8 year history of OHDSI/OMOP assets and discussions on this forum.
