YiDing Yu, MD, chief medical officer at Olive, provides an overview of how AI is successfully being used in the revenue cycle.
Revenue cycle executives have a lot of conflicting thoughts about artificial intelligence (AI). They're intrigued about it, but also sometimes confused and a bit skeptical. Not only do many vendors market tools like bots as AI, but even the experts struggle to provide real-world use cases illustrating how AI is being successfully used in the revenue cycle.
"Confusion and skepticism is incredibly natural, especially with newer technology," says YiDing Yu, MD, chief medical officer at Olive, an artificial intelligence and process automation solution designed specifically for healthcare.
Dr. Yu is our guest on the new episode of the HealthLeaders Revenue Cycle Podcast, where she provides an overview of how AI works, how it can help the revenue cycle, and most importantly, concrete examples of how AI is being used in revenue cycles right now.
To listen to the full episode, click here.
Here are three real-world ways that hospitals are using AI in the revenue cycle. Dr. Yu's replies are edited slightly for clarity:
Example 1: End-to-end optimization for prior authorizations
Dr. Yu: One of our most successful deployments is our end-to-end optimization suite … for prior authorizations. If you are in a hospital revenue cycle you might have centralized this, it might be decentralized at your physician offices, or it might be a combination of all the above.
We're an enterprise AI suite that supports all of those capabilities. Whether you're a hospital pre-cert team or you're a nurse at a physician's office, our prior authorizations suite essentially plugs into and connects seamlessly to your EMR. We support all the top dozen of the major EMRs, so that we automatically act the moment an order is placed.
So, imagine that a clinician sees a patient in the room. She thinks that the patient is going to need surgery and orders that surgery. The moment that order is placed, Olive goes into action. We check if a prior authorization is needed against the insurance, and we write that back into your EMR.
If a prior auth isn't necessary, we can close it out right there and then in the EMR, but if it is necessary, then it will pull through medical necessity criteria for over 40,000 health plans so that we know exactly what's required. We will also scrape through 13-18 months of clinical documents in the medical record looking for the relevant clinical documents.
Then, we put it all together so that when your staff does submit to the payer, everything was pre-done for you. You know all the rules were looked up, all the clinical documents were pre-packaged for you, and all you have to do is send it over.
We'll even complete that process. For many payers we have direct connections, so we could actually send it over. And then, even after a prior authorization is submitted, we can check the payer's website for the statuses, retrieve that, and give you the authorization information directly back into EMR.
When you think about something like that, the only interventions that a human would need to work are the exceptions; it's really an exception-based workflow. We're automating all the work. We're also augmenting the human for the work that they would normally have to do with those clinical documents review and natural language processing. We're making it easier for that person not to have to delve into those complex policies.
The outcome of this is really substantial. I'd say throughput for the average prior authorization takes anywhere on average 10-15 days from the day it is ordered by the physician to when it is finally approved. We've been able to accelerate that by over eight days, and so a patient is able to get care more than a week faster.
And also, if you have a last-minute opening in your operating room or in your MRI machine, you're able to accommodate these patients faster, whereas before, if it took you 10-15 days to get that authorization, a last-minute opening might be bypassed and unused, and it's another example of healthcare waste.
By accelerating care by over 80%, we've raised revenues. Organizations that have done this have been able to see an increase in top-line revenue. We had a hospital that for its single-hospital location raised revenue by $3 million in a year. And we're also able to reduce downstream challenges like write-offs that happen because small things get mistaken, and humans might make a mistake here and there.
The beauty of AI is it's all automated; there are no errors. We consistently show [that] our customers have a write-off reduction of 35%. In revenue cycle, prior authorization-related write-offs account for about a third of all the health system write-offs. It's a substantial ROI to get this right, not only for your patients but for the health of your health system.
Example 2: Claim statusing that works across all participating hospitals
Dr. Yu: True artificial intelligence needs to adapt and connect across multiple…networks and bring information that you might not have access to.
If we build a claim-statusing solution—we support currently over 700 hospitals—any single hospital would benefit from a change that another hospital would encounter. For example, [let's say] you are in Florida, and we've updated all of our claim statuses for Florida Blue. If a hospital in Michigan has a visit from a snowbird and they have Florida Blue, even though [the Michigan hospital] might never have built automations or tools for Florida Blue because they’re an out-of-state payer, because you have Olive, we're giving you access to that entire network of information.
Example 3: Upfront, guaranteed payments because of confidence in automation
Dr. Yu: [With] anything that is simple, that is easily automated, that we can easily predict and actually do for you, and the AI is handling, we're getting all the way to the point where we just launched Olive Assures, which is our platform that will actually guarantee payments. As a health system, you are paid the moment you deliver care, and do not have to worry about all that claims remittance and fighting with the health insurance companies.
The power of AI is that when we deploy in a [health] system, we can be so confident that a health system's revenue cycle will be reimbursed. We are so confident automating and being able to predict the outcome, that Olive is beginning to assure payments, so that you can accelerate your own cash flow, but more importantly have the confidence, knowing that you have an insurance policy where you won't lose money.
To listen to the full podcast episode with YiDing Yu, MD, chief medical officer at Olive, click here.
Alexandra Wilson Pecci is an editor for HealthLeaders.
