Unsurprisingly, I have seen similar trends in other fields like medicine where LLM-powered tools are adopted. One has to have the knowledge to recognize hallucinations and effectively leverage these tools.
University of Chicago sued for sharing “de identified” data with Google. Federal District Court and subsequent Appeals Court dismissed the lawsuit. Reason is that patients do not have property interest in personal information and impossible to show harm. I’m impressed they were able to “de identify”clinic notes. Curious to hear thoughts on the suit, seems like this opens door for way more healthcare data sharing.
Yes. Also, even within rare disease this application has desirable features. Obvious bio marker (hearing), relatively isolated organ system with obvious delivery method (inject into cochlea). main downside is that there is an existing alternative (implant) so risk/benefit question is less obvious compared to eg neurodegenerative diseases.
Largely agree w article. However I think it's important to mention that there are other validated ways to detect abuse and that clinical decision making is, or at least should be, based on multiple lines of evidence. Especially with abuse.
In ED and clinic I have had kids that screened positive for TEN 4 FACESp (see https://jamanetwork.com/journals/jamanetworkopen/fullarticle...). One of these babies ended up having b/l subdurals, likely abuse. In the end they may have called that "shaken baby", however there were multiple other red flags, the subdurals was only one piece, and the suspected mechanism of the subdurals was repeated drops / falls.
Agreed "shaking" may not be the mechanism but the author lists other mechanisms that can lead to subdurals and retinal hemorrhage where NAT could play role, eg hypoxia via choking, repeated falls from pushing/drops, neglect in unsafe motor vehicle situations.
Conversely, has anyone currently working in an IT position had to deal with physicians? What was the context? What skills did you wish the physician had? What made it easy or difficult to collaborate on a project?
I did. It is really interesting working with them, even if they change their mind very often.
However, some of them could be afraid of using new developments that you may think they could actually improve their working lives.
When I was in high school, I had a little business running on the side to install servers to backup files for physicians. I got paid about 300-400 for about 2-3 hours of work. Though I told them it usually takes about a full day.
