Hi Everyone,
It has been a little while since our last publication. Lots has happened and I am going to slowly update you as we continue the publication of this newsletter. Going forward there will be a lot more content relative to our innovation program as this has become a significant portion of the readers of this newsletter. Long time readers can still expect to get a lot of interesting random readings as I come across them. Thank you again for reading this newsletter as it evolves into its new format.
Hamid
Innovation Roundup
The FCVC Innovation Program partnered with the Ann Arbor Black Nurses Association this Saturday, February 11, 2023, to host a Design thinking workshop/African American Healthcare Innovators presentation.
On the day of the event we had 29 attendees present! Bev Willis, an Administrator for the Washtenaw County Historical Society at the Museum on Main Street in Ann Arbor gave the presentation on African American Healthcare Innovators).
During the workshop, attendees broke into three groups to discuss issues and innovative solutions at the University of Michigan Health System. Some topics the groups picked to discuss are listed below:
Group 1: Access
Issue: Providers lack knowledge of some patients’ cultures.
Solution: Education on how medicines affect different ethnicities, diets of different cultures, and holistic treatments.
Group 2: Access and mental health
Issue: Elderly and low-income populations are not able to travel to Ann Arbor for care.
Solution: Mobile clinics to low-income areas, education to inner cities.
Issue: Mental health is a concern for many patients.
Solution: Include a face-to-face option for the mental health hotline for more patient engagement and connection.
Group 3: Discharge
Issue: Long discharge times and unclear communication with patients.
Solutions: Use iPads more with patients (examples - video visits for provider sign-offs and follow-up video calls with patients after discharge)
I am excited to see the progress of these ideas as they develop into solutions for our patients.
University of Michigan Frankel Cardiovascular Center Innovation Challenge is off to another great start this year. Lots of high quality submissions. It is hard to believe that we are now in our seventh year!
Science Roundup
The rise and fall of peer review and Things could be better
I came across the these two pieces recently and I could not recommend them enough:
Scientists often say they take peer review very seriously. But people say lots of things they don’t mean, like “It’s great to e-meet you” and “I’ll never leave you, Adam.” If you look at what scientists actually do, it’s clear they don’t think peer review really matters.
First: if scientists cared a lot about peer review, when their papers got reviewed and rejected, they would listen to the feedback, do more experiments, rewrite the paper, etc. Instead, they usually just submit the same paper to another journal.
Second: once a paper gets published, we shred the reviews. A few journals publish reviews; most don't.
And third: scientists take unreviewed work seriously without thinking twice. We read “preprints” and working papers and blog posts, none of which have been published in peer-reviewed journals.
Instead, scientists tacitly agree that peer review adds nothing, and they make up their minds about scientific work by looking at the methods and results.
Large Language Models Encode Clinical Knowledge
MultiMedQA is a benchmark combining six existing open question answering datasets spanning professional medical exams, research, and consumer queries; and HealthSearchQA, a new free-response dataset of medical questions searched online. They also propose a framework for human evaluation of model answers along multiple axes including factuality, precision, possible harm, and bias. In addition, we evaluate PaLM (a 540-billion parameter LLM) and its instruction-tuned variant, FlanPaLM, on MultiMedQA. Using a combination of prompting strategies, Flan-PaLM achieves state-of-the-art accuracy on every MultiMedQA multiple-choice dataset (MedQA, MedMCQA, PubMedQA, MMLU clinical topics), including 67.6% accuracy on MedQA (US Medical License Exam questions), surpassing prior state-of-the-art by over 17%.
They show that comprehension, recall of knowledge, and medical reasoning improve with model scale and instruction prompt tuning, suggesting the potential utility of LLMs in medicine. Human evaluations revealed important limitations of the models, reinforcing the importance of both evaluation frameworks and method development in creating safe, helpful LLM models for clinical applications.
Doximity launched Docs GPT last week. it's a workflow tool that can generate clinical correspondence, denial appeals letters, procedure notes, progress notes, referrals, patient education, and more -- supported by a robust prompt library. Is it just me or the rate of progress on this is just mind boggling.
Gerotechnology for Older Adults With Cardiovascular Diseases: JACC State-of-the-Art Review
Regulatory Roundup
Unsettled Liability Issues for “Prediagnostic” Wearables and Health-Related Products
Pre-diagnostic products and other health-related applications are bringing exciting technologies directly to consumers.…But these products also present a context that is rife with legal uncertainty for all.
The.authors suggest potential solutions such as legislative changes including carving out safe harbors for physicians whose patients provide them with data from pre-diagnostic products, or adopting language specifying that, for liability purposes, physicians who receive data from a pre-diagnostic product should treat that information like any other patient self-reported symptom in the scope of their diagnostic work. Additionally, public regulators such as the Federal Trade Commission and state attorneys general could vigorously enforce existing laws against manufacturers of pre-diagnostic products that advertise in a misleading manner or otherwise violate state unfair and deceptive trade practices laws.
I think eventually the best practices developed by physicians could help shape liability determinations when they inevitably arise in the courts.
Podcast
This was a really fun conversation I had with the folks at FDA discussing the regulatory frameworks for digital health and software as a medical device.
Industry Roundup
Carbon Health announces $100M in funding days from CVS Health Ventures via Mobihealthnews
Telesair, a medical device company developing hospital-to-home products for respiratory care, announced it had raised $22 million in Series A funding led by Pasaca Capital with participation from Honeywell Ventures, ZhenCheng Capital, Shangbay Capital, Device of Tomorrow Capital, Berkeley Catalyst Fund and Ultrastar Ventures LLC. via Mobihealthnews
Censinet, developer of healthcare cybersecurity software, scored $9 million in a funding round led by MemorialCare Innovation Fund and by other participants in the raise include Rex Health Ventures, Ballad Ventures, LRVHealth, HLM Venture Partners, Schooner Capital, Excelerate Health Ventures and Cedars Sinai. Censinet said the round brings its total funding pot to more than $22 million. via via Mobihealthnews
Interesting Stuff Roundup
OpenAI is doing some amazing things. like DALL-E and ChatGPT
What's Up With That: Why It's So Hard to Catch Your Own Typos
Paris in the the Spring. Why do such typos often elude us? The explanation is simple, obvious and comforting. When we are proof-reading our own work we know what we mean to say and we expect to find the meaning there in the text. The words we see on the page must compete, in effect, with a more authoritative set of words already in our minds. One remedy: Read the text backwards
From Letters to a Young Poet by Rilke
I want to beg you, as much as I can, dear sir, to be patient toward all that is unsolved in your heart and to try to love the questions themselves like locked rooms and like books that are written in a very foreign tongue. Do not now seek the answers, which cannot be given you because you would not be able to live them. And the point is, to live everything. Live the questions now. Perhaps you will then gradually, without noticing it, live along some distant day into the answer.
Art of the week
The Races at Longchamp: Édouard Manet (French, 1832–1883)
From the The Art Institute of Chicago
Album of the week
I recently watched Tar (recommended) and it got me into a Mahler rabbit whole. This was my favorite and I had it on repeat for a few weeks after the movie.
Thank you again for reading my newsletter.
If you enjoy reading this newsletter, please share it with someone else who might enjoy reading it.