-
HCA Shutters Practices
-
Subscriptions Offer Low Cost Alternatives
-
23andMe GLP-1 Telehealth Service
-
LLM Assistants Create Shift in Patient Care
- AI Rules Out Abnormal Pathology on X-rays
'It's been devastating': Where a Physician Landed After HCA Shuttered Practices
Tennessee is a Certificate of Need (CON) state, with an eleven member board, five seats being healthcare executives. There are 34 other states with levels of CON laws. While this article is certainly one-sided (bad healthcare org!), it highlights problems with Certificate of Need.
In some cases, they protect bad structures, and in others they limit competition. They originated in the mid-70’s at the national level, and got dumped back on the states in 1987. They served a useful function when hospital systems were all about establishing satellite hospitals and feeder hospitals in outlying areas. They really do continue to serve a function in deal-making with institutions - put beds over there, and we’ll let you put beds over here.
But regulatory capture has taken place in the last 40 years…
'It's been devastating': Where a Physician Landed after HCA Shuttered Practices - Beckers ASC Review
Prepare for the Upcoming ABOG Exam for Less
Med-Challenger's OBGYN Review with CME offers a board pass guarantee and inexpensive subscription options.
23andMe Plans to Launch a GLP-1 Telehealth Service
We love telemedicine, when it’s tied to a physician practice, and part of a holistic patient care system. When structured and reimbursed properly, it’s a lifesaver for those with frequent check-ins and medical management. But something is seriously broken here.
23andMe Plans to Launch a GLP-1 Telehealth Service
NY Hospital Exec: Multimodal LLM Assistants Will Create a “Paradigm Shift” in Patient Care
It’s one of those happy articles about AI, and impacted heavily by the physician interview that centers the article. We don’t disagree with Dr. Beecy - a practicing cardiologist and associate professor. What struck us was that the whole article is not really about AI, it’s about data availability. Data accessibility, security, and storage is not exactly a strength of the current crop of EHRs and hospital data lakes.
While working on the things that AI can do right now, such as near-perfect transcription and exclusionary pattern matching, hospitals really need to manage the trillions of dollars of data they are warehousing.
NY Hospital Exec: Multimodal LLM Assistants Will Create a “Paradigm Shift” in Patient Care
Using AI To Exclude Pathology on X-Rays
This research experiment from Denmark used a CNN tool (Convolutional Neural Network - a form of image pattern AI) to process piles of chest X-rays for a negative result, rather than a positive. The results on exclusion were higher than a radiologist, but the mistakes were more severe. The test patterns weren’t given any clinical or patient history. The severity of the errors kind of follows Challenger’s theory that “1% of the time, the AI is trying to kill you”. 😁
Use of AI tools for exclusionary patterns, rather than trying to use them for diagnostics, is likely to be a much more successful approach to integration of the things that AI does well into clinical workflows. Negative pattern recognition in bulk at superhuman speed, rather than positive recognition.
Using AI To Exclude Pathology on X-Rays - Radiological Society of North America
|
