OHDSI Home | Forums | Wiki | Github

Can such a service be supported by the OHDSI community and tools?


(Nigam Shah) #1

I am excited to share our (Stanford and Duke authors’) perspective on how we can learn from aggregate patients to improve care at the bedside. Read, https://www.nature.com/articles/s41746-019-0091-3 or watch, http://tinyurl.com/StanfordGB.

Wouldn’t it be awesome if we could use the OHDSI community and tools to offer such a service at all academic medical centers?


(Patrick Ryan) #2

Thanks @nigam for posting this. I believe the ‘informatics consult’ idea that you and others have been advocating is one of the most exciting and most impactful opportunities in front of the entire field of biomedical informatics. The ability to use the collective past experiences of patient populations across different clinician and health system engagements to inform the future care of a patient seems to be what ‘real world evidence’ should be all about, and makes actionable the insights that can come from clinical characterization, population-level effect estimation, and patient-level prediction. For those of us who aren’t MDs, this is probably the closest we’ll ever get to directly improving the lives of the patients we desire to serve.

I would LOVE to see this idea expanded into an OHDSI contribution and capability, and am quite eager to help in any way I can to make this vision a reality.


(Anna Ostropolets) #3

Thanks for bringing it up! I totally side with Patrick: the concept of Green Button is fascinating in that it can bring the evidence we generate to the bedside. I can see how this project, paired with concise and precise summaries of the current evidence, can have an impact on the way clinicians practice medicine. Instead of using their peers as a last resort in complicated cases, they will be able to leverage the existing knowledge and make decisions guided by the experience of hundreds and thousands of MDs. I also can see how this project can benefit from the large-scale solutions that OHDSI can offer. It would be great if we can couple them together, and I hope Columbia University would be among the first medical centers to implement it. At least, I’m happy to contribute my knowledge and skills to make it happen and maybe even to expand the initial ‘patient like mine’ approach to cover a broader spectrum of clinicians’ information needs.


(Vojtech Huser) #4

I loved the idea of green button the first time you introduced them, Nigam. And I am still a great fan. In fact, I even proposed it as audacious goal for NLM Blue Ribbon Panel call recently made (in their call for such goals).

At some point in the past, I talked with Marshfield Clinic lawyers if it is OK if an AI is using prior patient charts to support current case. (or charts of family members). Can a physician simply open a similar patient chart if the goal is just to learn from that prior case (without being the attending doctor (of the patient of the chart in question).

I am on board to contribute!

edit: link to Stanford youtube video for convenience: https://www.youtube.com/watch?v=6oEFK24s8yU


(Rae Woong Park) #5

Congrats on the start of this huge project!. I posted your article and youtube video at my facebook, and one of my friends at LPCH (Lucile Packard Children’s Hospital) asked me if she can use Green button service at LPCH.


(Nigam Shah) #6

Thanks Rae! Yes, a couple of LPCH clinicians already use it. They all have @stanford.edu emails to make the request.


t