Any tool to browse patient information

Implementers
1

Hi All,

Is there any tool which can be used to browser/explorer patient informations?

For example, after getting certain cohort via CIRCE, we might want to check lab info and procedure info. Is there any tool I can use to bowser certain patient’s information, one patient by one patient?

Appreciate all kind of helps.

Thanks

Chris Leng

2

Hi Chris: There is not currently an OHDSI open-source tool for browsing
patient-level data, but I know several within the community have built
their own tools for this. A couple weeks ago, David Sontag presented a tool
that he and his team have recently built, which looks very impressive and
hopefully they’ll be able to share with the community in the near future.
I have used Spotfire as a visualization tool to make patient profiles in my
research at Janssen, but that’s not an open-source solution. There may be
interest within the community to build an open-source patient profile
viewer, in which case, I’d encourage those who are interested in
contributing to form a working group and get started.

Cheers,

Patrick

3

Thanks @Patrick_Ryan. We are looking for a solution of OHDSI patient profile viewer.

To everyone:

Is anyone having interesting to develop a open source patient profile viewer for the community?
We have a project that need that to review patient and their lab data, medicine use, diagnosis, even clinical notes. If you already have one, we can help test. Or, we can corporate different teams together to build one for the OHDSI community.

Does any have interesting on this?

4

Developers. We need a software program that can identify a population (cohort), identify patient characteristics (restriction) that include Lab, procedures, outcomes and information in the text note. We also want to retrieve specific text notes on these individuals for annotation and clinical review.

What existing software should we branch off and develop text searching functionality for?

Your feedback will be greatly appreciated.
Brian

5

CIRCES may do what you want. It is the official OHDSI tool for cohort building. Chris has done a really nice job on it and it was demonstrated at AMIA a few weeks ago. I don’t know whether it will pull notes.

We are developing something called Jigsaw for our internal purposes (also demonstrated at AMIA). We would be happy to talk to you about whether it would work and possibly beta-testing it. It is under heavy development, but if it works for you, then we would be happy to let you (and others in the community) use it.

6

Except for looking for the data in the note, identifying patients based on observed events found in the CDM v5 model is what CIRCE is designed to do. I say ‘except for the note’ because the note is unstructured data, and CIRCE looks in the structured elements of the CDM. if you could somehow process notes and generate CDM observations (the general term for an observation: ie: a condition_occurrence, or procedure_occurrence or drug_exposure) from the note data, then you could use CIRCE to query for it. But CIRCE by itself does not do text mining or NLP. Rather it would be geared for finding people that have characteristics based on data that has been extracted from notes.

-Chris

7

I wrote my own “local” R package (called METIS) that has some functions for quickly querying a CDM dataset and seeing quick outputs.

I wrote some rudimentary functions that display patient data. (not a fancy format, though).

8

Thanks everyone, we indexed text notes and built a system to retrieve notes and view them. We want to integrate with CIRCE but we are not sure if this is the best software to integrate with or whether Nigam or other have something else in mind. We can certainly make CIRCE work for us, but we want to contribute back to the community too.

9

Chris, one way to think of it could be that CIRCE identifies a population based on a set of criteria. We could build in a note viewer for the notes associated with specific outcome. We could retrieve note based on the structured query. Ideally, we would also support searching of indexed terms.