So the article mentions AI in the headline and the first paragraph, but there's no specific mention anywhere else in the article of these vendors using AI for assessing PAs - I'd love to know if...
So the article mentions AI in the headline and the first paragraph, but there's no specific mention anywhere else in the article of these vendors using AI for assessing PAs - I'd love to know if there's more information out there about this aspect of the issue. Instead, what's described sounds more like a pretty typical rules-based process, and it sounds to me like the bigger issue is that the payment structure for these vendors incentivizes them to deny as many PA requests as they can get away with. On top of that, these types of vendors will almost never hire adequate staff to deal with support needs in a timely manner, because that cuts into their profitability, so if the gov't doesn't have robust SLAs and meaningful financial penalties attached to them, then the experience of trying to get issues resolved is going to be an absolute nightmare.
Medicare’s AI Push Snarls Patients and Doctors in Errors and Delays - KFF Health News https://kffhealthnews.org/medicare/medicare-ai-prior-authorization-wiser-delays-errors/ I looked up a few of...
I looked up a few of the vendors and other articles seem to show that these are systems with integrated AI and that AI is being used in the process. I don't have time to dig deeper but it seems likely to me they'd be clearly stating how they're not using AI if that were a false allegation.
Hmmm. The evidence presented in this article is pretty scant as well. I'm not saying it's not happening, and if it is it's entirely inappropriate, but to me it sounds like a pretty typical case of...
Hmmm. The evidence presented in this article is pretty scant as well. I'm not saying it's not happening, and if it is it's entirely inappropriate, but to me it sounds like a pretty typical case of low bidding vendors cheaping out and building shit rules engines, and not hiring enough (or competent enough) staff to handle the volume of requests.
I tend to think that prior-authorization is a pretty shit practice in general when it comes to health coverage, personally. I recognize that provider fraud, upcoding, etc are very real issues but I think in general post-payment auditing is a more humane way of handling it that doesn't deny or delay healthcare recommended by the providers who are closest to the patient. I think the focus on AI in these headlines is distracting from a practice that is just designed to make healthcare worse and put more of a burden on providers and patients while lining the pockets of middle-men.
EDIT: I just want to be clear that I'm not just trying to be argumentative here. If these companies are using AI in PA or Claims review is absolutely needs to be found out and stopped unless there are very robust safeguards and testing on the type and capability of the models/systems being used - specifically trained machine learning models might be better than a human reviewer, but there needs to be a human in the loop and those systems need to be rigorously tested. And a general purpose LLM should absolutely not be used for decision making in healthcare. That said, I think the real problem here is that the profit motive is so deeply embedded in every level of american healthcare. I would argue that pretty much every corner of US healthcare is bloated, inefficient, rent-seeking, and inhumane. I know that providers are generally trying their best but the deck is stacked against them and the burden of dealing with all of the middlemen means they have to spend ungodly amounts of time and money just to be able to stay afloat. I work in this industry and I would be happy if the whole industry became irrelevant tomorrow.
It's the fact that they've added it to Medicare as a practice and are also using these tools that really create a default level of acceptance or rejection that doesn't take into consideration...
It's the fact that they've added it to Medicare as a practice and are also using these tools that really create a default level of acceptance or rejection that doesn't take into consideration those provider specific knowledge pieces. Like what is even the point of submitting if you know the first time we'll always get rejected for XYZ procedure. Or if the state patient will get rejected each time for ABC because they don't meet the profile but they have been approved before and they really shouldn't be rejected again.
This directly impacts my partner because he is on disability and while the articles tend to focus on seniors, it absolutely impacts disabled folks of all ages.
So yes, it's a bad practice but it's being worsened when there's any sort of automated non-humanistic system. Especially when you're adding a bad practice into a system for the most vulnerable while trying to limit the practice among private healthcare companies. But if these vendors would like to clarify that they don't use AI in this practice, they are free to do so. I will remain quite suspicious even if they do, but the fact that they aren't bothering suggests to me that this is an accurate accusation.
Yeah I edited my comment to try and clarify a bit because I felt like I was arguing without really even disagreeing with you. The way the healthcare system (if you can call it that) in the US...
Yeah I edited my comment to try and clarify a bit because I felt like I was arguing without really even disagreeing with you. The way the healthcare system (if you can call it that) in the US works is already an absolute dumpster fire without adding AI hallucinations into the mix.
In January, the Trump administration rolled out a pilot program that uses artificial intelligence to authorize or deny certain types of care for patients with Medicare—a federal healthcare program for seniors that previously hasn’t required doctors to get any such pre-approval, called prior authorization.
Not long after, media outlets began relaying the disastrous outcomes: technical difficulties, long delays in decisions and care, puzzling denials, frustrated doctors, and patients suffering in pain. Those reports were largely confirmed earlier this month when the Electronic Frontier Foundation released a tranche of federal documents about the program, called WISeR, that the group obtained amid litigation. The documents included feedback from healthcare providers, including one who called the program “a disgrace to the human race” and reported seeing patients crying in pain as they waited for care.
Amid the worrying reports, lawmakers have been trying to get answers and shut the program down. Meanwhile, the Government Accountability Office determined in May that Trump officials did not follow proper procedure in setting up the program, calling its legality into question. Still, it appears to be moving ahead unabated, with plans to expand in the years to come.
[...]
WISeR, which stands for Wasteful and Inappropriate Service Reduction, aims to use AI and machine learning to ensure “appropriate Medicare payment” for select services while benefiting taxpayers “by decreasing fraud, waste and abuse.” It was rolled out in January in six states—New Jersey, Ohio, Oklahoma, Texas, Arizona, and Washington—and intends to run until the end of 2031.
So far, the program requires prior authorization for around a dozen medical services, including nerve stimulation, epidural steroid injections for pain, cervical fusions (permanently connecting bones in the neck), skin substitutes for wounds, and treatments for incontinence and impotence.
According to the documents released by the EFF, the program’s implementation was rushed, and the technology wasn’t ready. One vendor, Innovaccer, was so unprepared that it asked the government to delay the rollout, and when it didn’t, the vendor set up its program to automatically approve all requests temporarily.
“[A]uto-affirming is the only path available that avoids creating a backlog of unprocessed prior authorizations and claims while we finalize, validate, and deploy the full rules-based solution,” Innovaccer wrote, according to a letter it sent to government officials.
But others carried on. One vendor, Zyter, had data discrepancies for months because it apparently didn’t understand the difference between Medicare Part A, which covers inpatient/hospital care, and Part B, which covers outpatient services. Zyter’s CEO, Sundar Subramanian, told Ars Technica in a written statement that the company is now “fully functional across Medicare Part A and Part B claims.” The company is working closely with federal partners and the healthcare providers to “enhance the experience,” Subramanian said.
[...]
In addition to a questionable number of denials, decisions from the various contracted companies often took too long. WISeR is intended to provide authorization decisions within 72 hours, but many have taken weeks, some even months. The documents from EFF reveal at least one request was still pending after 83 days. In a survey, one healthcare provider said it “had a surgery pushed back almost two months due to zero communication.”
So the article mentions AI in the headline and the first paragraph, but there's no specific mention anywhere else in the article of these vendors using AI for assessing PAs - I'd love to know if there's more information out there about this aspect of the issue. Instead, what's described sounds more like a pretty typical rules-based process, and it sounds to me like the bigger issue is that the payment structure for these vendors incentivizes them to deny as many PA requests as they can get away with. On top of that, these types of vendors will almost never hire adequate staff to deal with support needs in a timely manner, because that cuts into their profitability, so if the gov't doesn't have robust SLAs and meaningful financial penalties attached to them, then the experience of trying to get issues resolved is going to be an absolute nightmare.
Medicare’s AI Push Snarls Patients and Doctors in Errors and Delays - KFF Health News https://kffhealthnews.org/medicare/medicare-ai-prior-authorization-wiser-delays-errors/
I looked up a few of the vendors and other articles seem to show that these are systems with integrated AI and that AI is being used in the process. I don't have time to dig deeper but it seems likely to me they'd be clearly stating how they're not using AI if that were a false allegation.
Hmmm. The evidence presented in this article is pretty scant as well. I'm not saying it's not happening, and if it is it's entirely inappropriate, but to me it sounds like a pretty typical case of low bidding vendors cheaping out and building shit rules engines, and not hiring enough (or competent enough) staff to handle the volume of requests.
I tend to think that prior-authorization is a pretty shit practice in general when it comes to health coverage, personally. I recognize that provider fraud, upcoding, etc are very real issues but I think in general post-payment auditing is a more humane way of handling it that doesn't deny or delay healthcare recommended by the providers who are closest to the patient. I think the focus on AI in these headlines is distracting from a practice that is just designed to make healthcare worse and put more of a burden on providers and patients while lining the pockets of middle-men.
EDIT: I just want to be clear that I'm not just trying to be argumentative here. If these companies are using AI in PA or Claims review is absolutely needs to be found out and stopped unless there are very robust safeguards and testing on the type and capability of the models/systems being used - specifically trained machine learning models might be better than a human reviewer, but there needs to be a human in the loop and those systems need to be rigorously tested. And a general purpose LLM should absolutely not be used for decision making in healthcare. That said, I think the real problem here is that the profit motive is so deeply embedded in every level of american healthcare. I would argue that pretty much every corner of US healthcare is bloated, inefficient, rent-seeking, and inhumane. I know that providers are generally trying their best but the deck is stacked against them and the burden of dealing with all of the middlemen means they have to spend ungodly amounts of time and money just to be able to stay afloat. I work in this industry and I would be happy if the whole industry became irrelevant tomorrow.
It's the fact that they've added it to Medicare as a practice and are also using these tools that really create a default level of acceptance or rejection that doesn't take into consideration those provider specific knowledge pieces. Like what is even the point of submitting if you know the first time we'll always get rejected for XYZ procedure. Or if the state patient will get rejected each time for ABC because they don't meet the profile but they have been approved before and they really shouldn't be rejected again.
This directly impacts my partner because he is on disability and while the articles tend to focus on seniors, it absolutely impacts disabled folks of all ages.
So yes, it's a bad practice but it's being worsened when there's any sort of automated non-humanistic system. Especially when you're adding a bad practice into a system for the most vulnerable while trying to limit the practice among private healthcare companies. But if these vendors would like to clarify that they don't use AI in this practice, they are free to do so. I will remain quite suspicious even if they do, but the fact that they aren't bothering suggests to me that this is an accurate accusation.
Yeah I edited my comment to try and clarify a bit because I felt like I was arguing without really even disagreeing with you. The way the healthcare system (if you can call it that) in the US works is already an absolute dumpster fire without adding AI hallucinations into the mix.
Yeah I got you. It's a trash system with a new fun element of trash!
From the article:
[...]
[...]