Predictive analytics for audits can help identify missing revenue
There are two parts of maintaining correct audits: ensuring that the charges present on the claim accurately represent the care provided and documented and ensuring the claim is compliant with all applicable billing rules and regulations.
One way to ease the process is to review the charges on the claim with an application or tool that can predict errors using the analytics from outpatient Medicare claim data.
“Predictive analytics don’t give you the definite answer of what might be missing, or what might be wrong. But the analytics give you clues to investigate,” says Felicia Ziomek, MBA, BSN, CHFP, nurse auditor and CDM coordinator at Stanford Healthcare – ValleyCare in Dublin, California. “You then need to examine the medical record to determine if there was a charge omission or if there was a charge error made.”
Ziomek uses an application with predictive analytics that alerts her of claims to review. She advises examining all the evidence before jumping to any conclusions.
Using applications with predictive analytics can help you source the root cause of errors. Ziomek builds a list of all the applications used by each charging department, noting the batch number assigned to each system.
“When I’m looking at charges on a patient account, and I see an inappropriate charge, I can tell by the batch number which system the charge error came from,” says Ziomek. “Some departments like lab can have many different batch numbers, owing to the many different applications their lab tests run through.”
Batch numbers are especially useful on complicated claims where device HCPCS codes come from the clinical department charges, but surgery/procedure CPT® codes come from the coders’ abstracting application. Ziomek advises creating a grid with columns that itemize the following:
- Department charge CPT/HCPCS codes
- Coders’ entered CPT/HCPCS codes
- Codes that arrived on the UB claim form
“It’s important to understand how your system’s CPT and HCPCS code flow settings work, make sure they’re correct, and periodically re-evaluate them,” she says.
An example of Ziomek’s grid is as follows:
Dept. Charged HIM Coded UB Claim Form Code Description
33212 33212 Insert pacemaker generator only (wrong)
33207 (missing) Insert pacemaker w/ventricular lead
C1898 C1898 Pacemaker lead other than transvenous
Tools with predictive analytics can help identify missing net revenue. “When the above claim was rebilled, we received an additional $3,686 net revenue, the difference between CPTs 33212 and 33207,” says Ziomek.
She has found that the revenue uncovered is worth the cost of the software.
“Once I ran a four-month claim data collection through my predictive analytics tool,” she says. “I sorted the data to put the largest estimated net revenue recoupment at the top. I found enough annualized net revenue in just the first 22 lines to offset the cost of the software.”
In the future, Ziomek hopes applications with predictive analytics can evolve to help patient care directly. "Hospitals face penalties of up to 3% of CMS reimbursements, based on the number of readmissions occurring within 30 days of discharge," she says. Ziomek envisions a tool to identify drivers behind patient readmissions, which can help identify appropriate interventions to lower them.