Artificial intelligence technology can take on the burdensome tasks of RCM.
When Valerie Barckhoff, a principal at tax advisory firm Windham Brannon, in Atlanta, Ga., got a first look at what artificial intelligence (AI) can do for simplifying revenue cycle processes, she said: “I became ridiculously geeky excited at the possibilities.”
AI is incredibly good at breaking down repetitive processes, she says, which is what revenue cycle management is all about. “A lot of what we do in the revenue cycle is predictable and repeatable and we’re just hiring some entry level staff to do the work because it needs to get done.”
Windham Brannon piloted a program to create a digital employee for a client hospital to work on what she calls “my biggest pain point,” pre-certifications, notices sent to payers alerting them that patients want to have a healthcare service. “We were doing a poor job of initiating and obtaining the pre-cert and patients were getting delayed or rescheduled, or we were getting denials because we didn’t pre-cert properly.”
She partnered with the cardiology department to create a digital employee they’ve named “LIA” (for Learned Intelligence Applied)that has a login just like any other employee. “She can get the schedule. She gets procedure information and diagnoses, and can log into the payer portal and initiate the pre-cert,” Barckhoff says.
The benefits of having a digital “employee” that, like all artificial intelligence programs, learns as it goes and gets smarter as it receives more data, is that it can take on some of the mundane and repetitive work out of employees’ schedules.
While a live staff person still has conversations with payers, the virtual employee helps with follow-ups and pulling back information.
Barckhoff believes this is just the beginning of what the AI can do, and sees it as a crucial way to free up employees from burdensome tasks.
“There’s no way any revenue cycle operation can touch every claim that comes through,” she says.
She envisions that the digital employee will free up time so that real employees can go after the low dollar claims that practices rarely get to follow up on, typically leading to a high-volume back log.
“I see [AI] as having the ability to up-skill [employees] and have them do more value-added work,” she says.
Because AI programs learn as they go, the virtual employee can be programmed with a confidence rate to free it up to do tasks. For example, it can be trained to pull specific records and input them into the medical record. “So if they need an EKG, she can find a file and say ‘I’m 99 percent confident this is an EKG so I’m going to submit it,’” Barckhoff explains. “If it’s below the confidence level, it just kicks to a manual work queue.”
While Windham Brannon is beginning to use its digital employee with pre-certifications, she believes that AI will soon be able to help with handling “first level appeals of denial management” by going into the payer portals and pulling back information. “We’re going to be able to make sure the work is getting done consistently,” Barckhoff says.
Additionally, an AI assistant can help with scheduling. “We’re expecting to have to reschedule [fewer] patients because of administrative issues,” she says.
The response she’s received from executives has been overwhelmingly positive, so she urges physicians to be the drivers of this technology within their medical groups or hospitals-once they receive access to this technology.
“I think this is truly revolutionary for the revenue cycle,” she says.