Inquiry Regarding Implementation and Examples of the Generalized Data Model (GDM)

[iamalonso]

Hello Development Team,

I hope you're having a great day. I wanted to discuss a few aspects related to the Generalized Data Model (GDM) and seek your assistance on some specific points.

Firstly, I have reviewed the available manuscript describing the design of the GDM, as well as the current version of the schema. I would like to inquire if you have the necessary scripts to implement this model on a database engine. Additionally, I'm curious to know if you have practical examples of how to load data, such as laboratory data, into the model.

Having reviewed the GDM and comparing it with the option of directly mapping my data sources to OMOP, I have some questions regarding the practical advantages that the GDM offers in terms of simplicity and utility. I would greatly appreciate it if you could provide more examples or information highlighting the practical benefits and effectiveness of the GDM compared to other options.

Thank you for your time, and I sincerely appreciate any guidance you can provide.

Best regards,

[markdanese]

@iamalonso Thanks for contacting us. Can you give us a little background with regard to your use case? What kind of data are you planning to use? What software or code are you planning to use to extract the data from GDM?

By way of background, the majority of our use cases are for using GDM for Medicare and related data, although we have used data from Marketscan, Health Verity, and CPRD.

I am not sure what you are asking about when you ask for "necessary scripts to implement this model on a database engine". Are you asking about ETL code to transform raw data to GDM? Or are you asking for code to take data from GDM and create analysis datasets? Internally, we have ETL code for the former, and our own software package for the latter. Both may be difficult to share depending on your use case, but we may be able to help guide you.

[iamalonso]

Thanks @markdanese for your response. I'm working on a retrospective study of 143 hospitalized patients from clinical campus in Chile between 2016 and 2020, focusing on Code Blue procedures (Rapid Response Team signaling). These alerts involve an early assessment and treatment process for deteriorating patients. The study aims to predict in-hospital mortality risk using OMOP CDM.

As for your query regarding scripts, I was referring to the necessary DDL scripts for implementing the model in a database engine such as PostgreSQL. Additionally, I have successfully mapped 22 laboratory test data instances to LOINC, associated with specific patient hospitalizations. These lab results are linked to patient visits, and I am looking to structure this data within the GDM framework. However, I'm uncertain about the specific tables to populate for this purpose.

Your insights and guidance would be greatly appreciated, especially given the potential advantages GDM offers in comparison to directly mapping data into OMOP. If you could provide more examples or clarify the relevant tables for structuring laboratory data within GDM, it would be of immense help.

Alonso

[markdanese]

Hi Alonso:
We don't have anything handy that would help, as far as I know. Most of our code base is for different data than you are working with. I am sorry we can't help with your project.
Mark