Hi - I’ve done prior work as a member of Brown University’s Evidence-based Medicine team on the Systematic Review Data Repository (srdr.ahrq.gov) project. I am presently implementing an application on top of an OMOP CDM for our CTSI OneHealth Alliance (COHA) network at Tufts Cummings veterinary school. The COHA OMOP CDM contains over 20,000 PubMed Case Literature Reports - its MeSH and Substance terms parsed and mapped to OMOP’s MeSH and SNOMED vocabularies. Therefor you can run the same query on your Electronic Health Record (EHR) records - what most OMOP implementations contain - as well as on published abstracts to generate clinical trial and EHR evidence.
We run queries through the above application (COHA Research Workbench) via a clinician focused interface for preparatory to research activities - these are iterative exploratory queries to help formulate your hypothesis. As I mention - most OMOP CDM implementations are focused on EHR data only. Any other data like PubMed publications will need to be parsed and loaded into an OMOP CDM before it can be queried - you will need to write an Extract/Transform/Load (ETL) program as well as decide how to represent your data. Once you do that - than any of the OHDSI tools can be used.
Otherwise you would use two applications to generate published evidence first - PICO against PubMed.gov and then use OMOP CDM tools to generate patient evidence. We are at the tail end of our COHA Research Workbench grant so its usefulness and objectives are to be tested this year. If successful - I will look at moving it over to the human side of our Tufts CTSI OMOP CDM and bring in more types of PubMed abstracts. There are 120K Case Literature Reports on the human side. Then randomized clinical trials.
I would get familiar with SQL and the OMOP CDM. I’d be happy to discuss PubMed data in OMOP CDMs further. Best.
