Logo

Stanford Center for Biomedical Informatics Research Colloquia - Shared screen with speaker view
Panida Piboolnurak, MD
19:17
Hi Dr. Lin. May I connect with you via e-mail? I would love to learn more from you and would like to have your Advice. I am a neurologist, specialized in Parkinson's Disease and other Movement Disorders. My personal e-mail is pbnr2551@live.com.
Pat Langley
23:41
Q. Do you believe there has been comparable growth in the quality of results produced by AI systems in healthcare? As compared to results from 20 years ago.
Panida Piboolnurak, MD
25:27
Another problem is that AI platform are expensive. Are there any AI platform that we can use now, and not too expensive?
VHAPALGOLDSM
28:06
Dr. Lin, hope we can talk separately - apart from this presentation - about VA work in AI- and ML-related areas for primary care
Jonathan Chen
28:07
Thanks. See the questions and hands raised. Will get in when there's a break in thread, or aggregate to discussion at end
Pat Langley
32:50
Not all AI platforms are expensive. Classic expert systems in medicine, and even diagnostic aids learned using classic induction methods, are not.
Panida Piboolnurak, MD
35:58
Could you give me some samples of those systems? I would like to see if I can start using some of them.
Pat Langley
36:37
This all sounds fine, but it would be good to hear your analysis for why the first wave of AI applications in medicine did not succeed and how you plan to avoid the same problems encountered then.
Pat Langley
40:25
Why do you think there was wide adoption of Web search engines and recommender systems in the absence of a national strategy? Why can’t we take lessons from their success to increase adoption of medical AI technology? Why do we need a centralized solution?
Scott GK MacLeod
40:45
Thanks for your excellent talk, Dr. Lin!You mentioned in your beginning conversation about primary care longitudinal medicine. If both aging reversal genetic engineering drug therapies as well as longevity genetics emerge, - and with AI - and even regarding Avatar Bot Electronic Medical Records, how would primary care medicine with regards to your AI proposal best navigate these remarkable developments in long term health care? How would this change your plan and thinking? (eg Harvard HMS and MIT Professor of Genetics' George Church even mentions the word 'immortality' in video here - https://twitter.com/WorldUnivAndSch/status/1482057007691538434?s=20) So, brainstorming-wise, how could AI help regulate the amount and kinds of genes added back into individuals? Thanks, Scott GK MacLeod (sgkmacleod@worlduniversityandschool.org, planning ~200 online CC-4 MIT OCW-centric Medical Schools in ~200 countries and in their main languages)
Pat Langley
41:43
You’re welcome to prioritize and triage the questions / issues that I have asked here.
Jonathan Chen
42:37
Got it. Yes, several good discussion elements. Dr. Lin asked to have all questions at end rather than in middle, but I'll bring you on in Q&A cycle. See if you want to prioritize 1-2 first, then if get through queue, will come back for more
Blackford Middleton
45:09
Hi all — I have a question about the need for large aggregated national databases — we find most models are most useful when built with ‘local’ data. Thoughts?
satya
46:09
Any specific care plan related AI efforts for seniors apart from pallative care
Scott GK MacLeod
47:06
Blackford, with regards to electronic medical records in all 50 US states, would you recommend such large databases be organized around the local of the hospital or the county, the state complemented by the US federal government? Thanks.
Lee Sanders
49:04
Steve - Thanks for taking this on. I’d love to bring our Pediatric Primary Care and AI researchers into your team. A question: To address inequities and biases, how might we think about enabling patient-centered outcomes and AI for families with limited health literacy and language barriers?
Pat Langley
49:42
@Panida: You may be confusing the cost of developing AI applications and the cost of using them. Once a system has been built, then the cost of using it will be small.
Blackford Middleton
49:58
+1 Scott Hi tricky question, and there may be need practically for 3 levels of aggregation: local IDN/CIN (integrated delivery network/ clinically integrated network), State-based or regional public health reporting authorities (possibly building off regional HIEs or new QHINs), then a national level of aggregation… not all the same data at each level of course.
Pat Langley
50:58
@Nigam: Rather than a gold standard, how about a silver nitrate standard? 🙂
Blackford Middleton
51:26
What about feature engineering / definition and reuse?I think a national library of qualified/vetted features might be helpful.
Garrett Kneese
51:55
Woth Primary Care as a focus of AI advancement, how does this focus/Advocacy specifically implicate Healthcare access and reduction of health inequities via AI in PC? Will AI make care more equitable or overome disparate social determinants of healtb?
Walter Sujansky
52:08
A huge facilitating role of EHRs is obviating the need to *re-enter* patient-record data into the decision-support system. If an AI algorithm required the manual re-entry of 20-100 patient data points, it would be a non-starter in day-to-day use. That, alone, was a big impediment to adoption of decision-support technologies in the 1980’s, for example, IMO.
Blackford Middleton
52:11
Also what about a national certification program? e.g. ULabs for algorithmic vigilance… Mayo is considering now such a program.
Scott GK MacLeod
52:11
Great, Blackford … brainstorming-wise, what about a realistic virtual earth for medicine and also for tele-robotic surgery with electronic medical records in all 200 countries (am thinking Google Street View with time slider, Maps, Earth, TensorFlowAI and realistic avatar bots) for data aggregation?
Quang
52:49
What AI-related health tech solutions have you seen or used recently that you believe does improve primary care physician workflows or quality of life?
Panida Piboolnurak, MD
53:09
I am working in Wilmington, NC. How can I collaborate or involve in your AI projects? My patient population will be very similar to primary care.
Walter Sujansky
54:01
Sorry, Jon, I’m sitting on a plane!
Pat Langley
54:56
Be careful. If AI medical apps become widely enough available, then people will stop going to primary care physicians entirely!
Nigam Shah
55:00
Several Stanford faculty participated in an NAM activity that sketches out what needs to be done about AI in Healthcare — https://nam.edu/wp-content/uploads/2019/12/AI-in-Health-Care-PREPUB-FINAL.pdf
Blackford Middleton
55:18
+1 Scott very interesting… reality or dream? Optum is considering a lifetime PHR for the US. We studied this at the CITL: Value of PHR… avail on my LinkedIn profile
Nigam Shah
55:40
Student question in QA panel - How do you think we should evaluate whether or not a primary care AI tool is providing value? I think there is a serious risk of creating tools that don't provide much value
Pat Langley
56:02
Not area under the ROC curve!!!
Pat Langley
56:34
Not area under the ROC curve!!!
Scott GK MacLeod
57:15
+1 Steven and Blackford, what about developing further an AI strategy further with Google Health, with Stanford Medicine and Duke Medical School study emerging from Google’s Project Baseline, with TensorFlowAI ?
Scott GK MacLeod
01:00:28
Steven, you mentioned in your beginning conversation about primary care longitudinal medicine. If both aging reversal genetic engineering drug therapies as well as longevity genetics emerge, - and with AI - and even regarding Avatar Bot Electronic Medical Records, how would primary care medicine with regards to your AI proposal best navigate these remarkable developments in long term health care? Thank you
Panida Piboolnurak, MD
01:08:50
Precision medicine is also very important.
VHAPALGOLDSM
01:11:12
thank you for the really interesting presentation and discussion! need to leave for next meeting. Hope to be able to connect with Dr. Lin later
Justin Norden
01:12:24
On the Health Technology VC side primary care funding is increasing, and is certainly a focus area for many. https://rockhealth.com/insights/2021-year-end-digital-health-funding-seismic-shifts-beneath-the-surface/
Scott GK MacLeod
01:12:28
Thank you!
Blackford Middleton
01:15:14
+1 Steve — please lmk if I can be helpful in any way with the national initiative.
Scott GK MacLeod
01:16:36
Bring primary care physical live physicians into telemedicine as they possibly become AI avatar bot physicians … as first idea, Satya
Scott GK MacLeod
01:17:25
Thank you, Dr. Lin!