Interesting. I think there is good with use of AI but I just found out its quite imbedded in many hospitals particularly Palantir. The problem is it shows data and can solve some issues, but does it address critical staffing or is it more for efficiency of the hospital leaving out patients? Also notice Palantir has joined collectively with Joint Commission.
There's definitely potential for AI to improve medical quality and efficiency. When I first starting being involved in AI development a few years ago, I was really excited to have something help us triage away negatives and junk cases, and to hone our eyes to rare findings. Unfortunately, the model in vetmed is cost savings by trying to replace people, rather than augmentation. I hate being put in the position of reflexive skeptic/crank, but so much of what these companies are putting out there are just crummy products with little to no validation.
One thing that definitely that definitely protects human medicine are numerous regulations. The FDA and CMS-CLIA treat AI algorithms like medical devices that need approval, and there is a relatively high barrier to clear. No similar organizations or rules exist for vet diagnostics, so it's truly the Wild West :/
Alex Karp of Palantir strikes me as a nut, and I don't love all the work they do with ICE and the Pentagon, although it sounds like what they essentially do is custom modify other AI applications for specific contracts. They're basically glorified middlemen, rather than an all-powerful company they are sometimes hyped to be
This is frightening and makes this pet owner upset. If the systems are not validated and are not accurate, why can they be sold and used? Why would any DVM trust my pet’s health to this? How much pet misery continues due to these errors? How many euthanasia decisions are made based on errors? Seems like the AVMA is asleep at the switch on this. Seems like a profit motive has eclipsed the academic rigor.
You are right to be upset, as many of us in vet med are.
"If the systems are not validated and are not accurate, why can they be sold and used?"
The simple answer is there are absolutely no actual regulatory or legal standards for accuracy in veterinary diagnostic labs. There is voluntary accreditation by the AAVLD, but it's a low bar, and pretty much the only labs that go through with the process are university and state labs.
The FDA has minimal oversight for animal devices, but it is primarily related to false advertising, and is reactive: if there are complaints that something is grossly defective or misleading, they MAY take action, but they don't have to prospectively validate anything for animals like CLIA standards that apply to human medicine
“Why would any DVM trust my pet’s health to this?”
People are overwhelmed. Caseloads have exploded, many vets are wearing 15 hats and don’t feel comfortable interpreting X-rays (we receive a modest amount of training in vet school, but it’s not a 3 year long focused residency). Sending rads out for specialist consultation is super expensive—it can more than double the cost to clients, and many decline the add-on. Finally, a lot of busy vets in practice get bamboozled by slick salespeople, and there is minimal independent data to contradict the marketing hype.
"How many euthanasia decisions are made based on errors?"
I shudder to think. Probably too many.
"Seems like the AVMA is asleep at the switch on this. Seems like a profit motive has eclipsed the academic rigor."
The AVMA is mostly a lobbying and publishing body. If there is anyone "asleep" at the wheel it is the state boards of veterinary medicine that actually regulate the practice and licensing of DVMs in their states. They could conceivably change the rules in their states to require more human oversight. That said, until we have more published data, even they are flying blind.
Specialty organizations like the American College of Veterinary Radiologists have been forming task forces and trying to shape diagnostic AI in a safer direction, but there's a lot of money on the line for the private companies. My society, the ACVP, has been much more lackadaisical, and if anything, uncritically embraced AI. Hopefully the tide is changing...
Jesus Christmas. I cannot even with that
Interesting. I think there is good with use of AI but I just found out its quite imbedded in many hospitals particularly Palantir. The problem is it shows data and can solve some issues, but does it address critical staffing or is it more for efficiency of the hospital leaving out patients? Also notice Palantir has joined collectively with Joint Commission.
There's definitely potential for AI to improve medical quality and efficiency. When I first starting being involved in AI development a few years ago, I was really excited to have something help us triage away negatives and junk cases, and to hone our eyes to rare findings. Unfortunately, the model in vetmed is cost savings by trying to replace people, rather than augmentation. I hate being put in the position of reflexive skeptic/crank, but so much of what these companies are putting out there are just crummy products with little to no validation.
One thing that definitely that definitely protects human medicine are numerous regulations. The FDA and CMS-CLIA treat AI algorithms like medical devices that need approval, and there is a relatively high barrier to clear. No similar organizations or rules exist for vet diagnostics, so it's truly the Wild West :/
Alex Karp of Palantir strikes me as a nut, and I don't love all the work they do with ICE and the Pentagon, although it sounds like what they essentially do is custom modify other AI applications for specific contracts. They're basically glorified middlemen, rather than an all-powerful company they are sometimes hyped to be
Look at this FDA has a partnership with Palantir https://investors.palantir.com/news-details/2020/Palantir-Technologies-Expands-FDA-Partnership-With-44m-Contract/
This is frightening and makes this pet owner upset. If the systems are not validated and are not accurate, why can they be sold and used? Why would any DVM trust my pet’s health to this? How much pet misery continues due to these errors? How many euthanasia decisions are made based on errors? Seems like the AVMA is asleep at the switch on this. Seems like a profit motive has eclipsed the academic rigor.
You are right to be upset, as many of us in vet med are.
"If the systems are not validated and are not accurate, why can they be sold and used?"
The simple answer is there are absolutely no actual regulatory or legal standards for accuracy in veterinary diagnostic labs. There is voluntary accreditation by the AAVLD, but it's a low bar, and pretty much the only labs that go through with the process are university and state labs.
The FDA has minimal oversight for animal devices, but it is primarily related to false advertising, and is reactive: if there are complaints that something is grossly defective or misleading, they MAY take action, but they don't have to prospectively validate anything for animals like CLIA standards that apply to human medicine
“Why would any DVM trust my pet’s health to this?”
People are overwhelmed. Caseloads have exploded, many vets are wearing 15 hats and don’t feel comfortable interpreting X-rays (we receive a modest amount of training in vet school, but it’s not a 3 year long focused residency). Sending rads out for specialist consultation is super expensive—it can more than double the cost to clients, and many decline the add-on. Finally, a lot of busy vets in practice get bamboozled by slick salespeople, and there is minimal independent data to contradict the marketing hype.
"How many euthanasia decisions are made based on errors?"
I shudder to think. Probably too many.
"Seems like the AVMA is asleep at the switch on this. Seems like a profit motive has eclipsed the academic rigor."
The AVMA is mostly a lobbying and publishing body. If there is anyone "asleep" at the wheel it is the state boards of veterinary medicine that actually regulate the practice and licensing of DVMs in their states. They could conceivably change the rules in their states to require more human oversight. That said, until we have more published data, even they are flying blind.
Specialty organizations like the American College of Veterinary Radiologists have been forming task forces and trying to shape diagnostic AI in a safer direction, but there's a lot of money on the line for the private companies. My society, the ACVP, has been much more lackadaisical, and if anything, uncritically embraced AI. Hopefully the tide is changing...
Yes