Out today: our simulation study on how AI ambient scribe tech performs across specialties
#AI #AmbientScribe technology produced moderate-quality notes in simulated encounters, with highest ratings in #Neurology and lower ratings in #Cardiology; clinicians highlighted workflow integration and accuracy as key challenges. ja.ma/3N5LJTF
Excited to share this paper I wrote with @seppo_rinne for @AnnalsATS. Tried to cut through hype on LLMs and talk about how they're relevant to the work of health services researchers. Not "what they might eventually do" but what they are good for (and bad for) RIGHT NOW.
Large Language Models as a Tool for Health Services Researchers: An Exploration of High-Value Applications
🔗 tiny.cc/8trcyz
JulianBrunner retweeted
We at YIMBY Denver pride ourselves in having a message that appeals to everyone, but it's sometimes important to tailor the message to the audience.
So, here’s a cheat sheet for how to speak our most important phrase in various dialects, a 🧵!
JulianBrunner retweeted
Have the words "eliminate exclusionary zoning" ever appeared in a Presidential transition plan before? You love to see it.
JulianBrunner retweeted
Worth reading this whole interview with @BillGates
He's frank and plain-spoken, for example in this section, echoing what @DrTomFrieden and I have been saying about using payment levers to get faster test results.
wired.com/story/bill-gates-o…
Meet Talia Avery Brunner!
Born July 20th at 1:11pm. 20 inches long, 8lbs 12oz! Mom and baby healthy and happy. Dad blissed out. Hashtag pending. #taliab ? #babytal ? @ Kaiser Permanente Labor&Delivery instagram.com/p/CC9IZrQHf15/… – at Los Angeles, CA
JulianBrunner retweeted
For those who are interested in research-based solutions to stop police violence, here’s what you need to know - based on the facts and data. A thread. (1/x)
public health infrastructure > medical infrastructure newyorker.com/news/news-desk…
This article confirms what I already believed about how people are responding to covid newyorker.com/news/daily-com…
JulianBrunner retweeted
Exclusive to NPR's @FrancoOrdonez -- the bipartisan proposal to fund a national contract tracing program. Led by @ASlavitt @ScottGottliebMD
npr.org/2020/04/27/845165404…
Quick fix for foggy glasses while wearing a face mask
ncbi.nlm.nih.gov/pmc/article…
This podcast needs to exist! Pledge at least a dollar to make it happen, and become part of the origin story of "speculative journalism"
This quoted post is unavailable.
JulianBrunner retweeted
This is the most useful & compelling recommendation for conference poster redesign I have ever seen! Great job bringing #HCD/#UX to academia, @mikemorrison! I suggest we adopt this fully for all #SIRC2019 poster sessions!! @ImplementCollab #ImpSci youtube.com/watch?v=1RwJbhkC…
Another turnaround for evidence on head start vox.com/future-perfect/2019/…
Qualitative researchers: tired of myopic responses? See our new paper where we asked people what “ideal” would look like. It’s @alisonh3’s famous “queen for a day” question. authors.elsevier.com/a/1YCng…
At #DIscience18 @dramyabernethy emphasizes SPEED of implementation. Accept more uncertainty up-front, then continuously monitor & re-evaluate w/ real-world evidence. Do D&I ppl know CMS has done this for years via “Coverage with Evidence Development”? disq.us/t/2w7za32
At #DIscience18 session on #DeImp, @helfrich_c says PCPs report they don’t have time to keep up w/ COPD Rx guidelines, yet COPD Rx is not worthy of pulmonologists' time. That’s a big deal! We need to create and entrench optimal norms for when (and how) to engage specialists.










