ISO/IEC AI foundational standards

Hi Derek,
I think this is great news. I recall going to an AMIA workshop inn 2005 with Richard Shiffman on GEM and GEM Cutter. There was also the GLIDES project. I have tried several times to see if there were any more recent developments in this regard but was not able to find anything.

Ray, I could not agree with you more, old friend!

There has been a lot of terrific progress on this very topic over the last few years -- with Canadians (myself included) playing a leadership role. I presented on this topic at the 2021 eHealth Conference .

Last month I saw demos of a still-in-stealth-mode product suite that includes a guideline authoring tool (useful to folks like Diabetes Canada for translating their "dusty" narrative PDFs into Computable Care Guidelines) plus a SMART-on-FHIR runtime engine that adds CCG processing to any capable EMR, as a façade. There will soon be tools readily available to realize on the vision you describe, Ray.

I believe we will, soon, begin to reap real population health benefits from our digital health investments. A new day will dawn.

Hi Derek,
I agree with your comment "data precision needed to support health insurance transaction processing seems to be what (most) clinicians are looking for, and what EMR vendors target for". I think that this is exactly where the problem is. The payment system is also a strong contributing factor. There has been so little movement towards using computers to improve clinical practice over several decades that it is really frustrating. Published clinical practice guidelines are thought to be a great resource to improve clinical care, yet they are not published in a computer interpretable format. As a result they just collect dust on doctors shelves. I think we should try and engage clinical practice guideline developers to produce CPGs that could easily imported as clinical decision support rules. This would of course require careful editing of the CPGs and involvement of the EMR vendors and the clinicians. How can this be made to happen?? One would think that knowing that only one third of patients with a variety of chronic conditions are getting guideline recommended treatments.

This is a fascinating discussion; thanks, all!

I am not convinced that we can lay all the blame at the feet of the EMR vendors. In my experience, water finds its own level... and an inconvenient truth is that the degree of data precision needed to support health insurance transaction processing seems to be what (most) clinicians are looking for, and what EMR vendors target for. We may want to take a page out of the Dutch playbook. If better-coded, more fulsome data were required before a provider got paid... I'm quite certain the EMR vendor community would quickly heed the baleful calls of clinicians who wanted help to meet the "new water level".

Now... can we all agree to stand behind the brave ADM that will need to sell this idea to the clinician community in their jurisdiction?

Hi Peter,

Yes, it looks like we are on the same page. Sadly, some of the major vendors are not.

Ray

Great points, Ray.

I wrote office automation software in the 1970s and 1980s, and our physicians appreciated the way that our system handled problem and other lists which were presented in a manner that was easy to update. Admittedly, way back then, we also had more clerical staff to help with all that!

Once you have good data, it is easy to use them to help to keep getting good data, I think. That seems to be aligned with what you are saying about SNOMED and the Problem List, if we are on the same page.