Intermountain Health executives have ramped up the health system’s revenue cycle management (RCM) AI and machine learning to prevent what they believe are excessive billing denials from insurers.
It’s been working so far.
The Salt Lake City-based health system has created an RCM strategy that relies on both machine learning and human oversight. The approach has reduced the number of denials that persist at a time when Intermountain has experienced a sharp uptick in initially rejected claims. AI is used to make it easier for doctors to appeal denials.
“It just seems to us like the payers are denying more and more and more,” said Sathya Vijayakumar, MS, MBA, LSSGB, director of appeals and denials with Intermountain Physician Advisor Services.
Given that this problem affects the patient experience, Intermountain was eager to get ahead of the denials.
“The reason we do this is at the end of the day, when a patient leaves our hospital, we don’t want them to be stuck with a bill that’s going to make their lives miserable,” Vijayakumar said.
That means trying to stop a denial before it starts.
“We’ve been trying to think ahead and kind of predict where or what kind of cases might get denied,” she said.
Identifying trends
Prior to adopting its predictive model, Intermountain would focus on retrospective methods that tended to identify process issues that can lead to denials, such as using an incorrect template or regularly entering information incorrectly.sing an incorrect template or regularly entering information incorrectly.
Now, by employing a machine learning model, Intermountain can break down denial trends by payer, explaining why one payer might be denying a certain diagnosis-related group (DRG) and another might not.
“Payers have scrubbers,” Vijayakumar said. “When they see a claim come in, they’re checking boxes. And if something is not met, they push it right back.”
But Intermountain’s team can build its own scrubber based on historical data that doesn’t let the claim go out the door before Intermountain has checked all the boxes, thus preventing a denial.
That dynamic, forward-looking approach is important because managing denials can turn into a game of whack-a-mole as payers reject different DRGs for a new reason.
“The insurance companies are going to change,” said Vicki Flores, RN, CPC, COC, director, Intermountain Physician Advisor Services. “They’re going to change what they’re doing. They’re going to change what they’re denying. You can’t be in a spot where you’re just stuck. You have to be able to change and adapt with them to stay on top of it.”
The new process, which Intermountain calls the Financial Impact Tool, was introduced just as the number of denials among physicians grew 2.7 times. Yet the overturn rate fell less than one might expect, dropping to 57% from 62%.
“That is really a good outcome considering our volumes are 2.7 times higher,” Vijayakumar said.
Breaking down trends by payer
Now, by employing a machine learning model, Intermountain can break down denial trends by payer, explaining why one payer might be denying a certain diagnosis-related group (DRG) and another might not.
“Payers have scrubbers,” Vijayakumar said. “When they see a claim come in, they’re checking boxes. And if something is not met, they push it right back.”
But Intermountain’s team can build its own scrubber based on historical data that doesn’t let the claim go out the door before Intermountain has checked all the boxes, thus preventing a denial.
That dynamic, forward-looking approach is important because managing denials can turn into a game of whack-a-mole as payers reject different DRGs for a new reason.
“The insurance companies are going to change,” said Vicki Flores, RN, CPC, COC, director, Intermountain Physician Advisor Services. “They’re going to change what they’re doing. They’re going to change what they’re denying. You can’t be in a spot where you’re just stuck. You have to be able to change and adapt with them to stay on top of it.”
The new process, which Intermountain calls the Financial Impact Tool, was introduced just as the number of denials among physicians grew 2.7 times. Yet the overturn rate fell less than one might expect, dropping to 57% from 62%.
“That is really a good outcome considering our volumes are 2.7 times higher,” Vijayakumar said.
Using AI
Intermountain also is using AI to assist with the written appeals to consolidate data and bring it to the physician, adding more efficiency to its revenue cycle.
The time spent on appeals has fallen precipitously. It used to take two hours for physicians to write an appeal for almost half of all their appeals. Now it’s taking 30 to 45 minutes, according to Vijayakumar.
Intermountain has adopted the tool in all 33 of its hospitals without adding full-time equivalent employees, with the intent of making the patient experience better across the system.
“The No. 1 reason we do this is for our patients,” Flores said. “It’s very important to us.”
