Coding of physician office visits can be very confusing – and is ever-changing.

Every year, regulations and guidelines change, and it is difficult for providers to keep up. Most groups have coding staff with limited training, so some changes go unrecognized.

The flipside of coding is compliance – and the ever-rising risk of compliance audits that may have grave financial consequences to provider groups if violations are found.

Increasing top-line revenue

"In healthcare, you get paid only for what you document and the codes you submit," said Dr. Bruce Cohen, a surgeon and former CEO at OrthoCarolina in Charlotte, North Carolina. "Proper coding of encounters in association with proper documentation can truly increase top-line revenue.

"Often, providers are working on old rules and procedures and are missing significant revenue opportunities through their coding of office visits," he continued. "This is especially evident in primary care as most of the practice income is derived from office visits."

AI-based programs that provide proper coding guidelines and leveling in real time can be invaluable additions to a practice, Cohen said.

"This doesn't mean eliminating the jobs of coders; it expands the oversight and accuracy of every charge going out based on evaluation and management (E/M) coding," he explained. "It also brings a level of compliance to documentation that isn't present in current practice environments. As annual coding requirements are instituted, an AI-based system will integrate and implement those changes in real time."

Moving to an AI system

OrthoCarolina went with the Calm Waters AI coding system from vendor MontecitoPLUS to help with its coding efforts.

"Calm Waters AI is used by physicians, physician assistants, medical coders and certain others who review the physician's documentation and assign CPT and ICD-10 codes for appropriate billing for services provided," Cohen explained. "Coding is a complex and often-confusing process due to both the sheer volume of distinct codes and the number of regulations and guidelines that change from year to year, and within each year.

"For E/M services, part of the physician's daily workflow is to review each patient encounter and provide documentation of the patient's symptoms, history and diagnosis, along with treatment recommendations," he continued. "Based on the diagnosis – including the acuity and complexity of problems diagnosed – the complexity of data required to make the diagnosis, and the level of risk of complications and morbidity/mortality in the patient, the physician assigns a level of medical decision-making (MDM) to the encounter."

The types of MDM – low complexity, moderate complexity and high complexity – guide coding and billing decisions for the providers' services.

Integrating AI into the Epic EHR

"After the provider has documented the types of history, exam and MDM, the group's coders can assign E/M codes based on this information," Cohen said. "Give-and-take between coders and providers often hinders this from being a seamless and time-efficient process. Coders may disagree with the MDM level assigned by the physician, or they may ask for more documentation from the physician to justify a certain MDM level and the billing associated with it.

"Because Calm Waters AI is integrated into our Epic EHR, it is seamlessly part of the physicians' workflow," he continued. "The system relies on artificial intelligence to review physicians' documentation for each encounter and suggests the appropriate level and coding."

Physicians then can review the system's recommendations and decide in a matter of seconds whether to accept or override them. The system helps them identify potential coding compliance and documentation issues while records are still on the physician's desktop, before they reach the coding and billing stages and errors become more difficult and time-consuming to correct.

"In this way, the system helps increase accuracy and compliance, reducing our risk of payer denials, delays and audits," Cohen noted. "In addition, the system saves documentation time for providers and makes the jobs of coders simpler and faster."

What they expect from the system

Cohen said it is too early in the implementation to report any hard results, but believes the AI-powered system will deliver a significant return on investment by enabling OrthoCarolina to receive reimbursement for more of the time physicians spend caring for patients and by improving the accuracy of E/M coding.

Cohen has a few words of advice for peers when implementing or using AI-based systems for coding.

"The first is to reassure your coding staff that this should not be seen as threatening from a job security standpoint," he said. "It should reaffirm the importance of their job and give them the ability to expand their scope without exceeding their current capacity.

"Most groups survey the accuracy of their coding through audits or spot-checks; by contrast, we can use these AI-based systems on every encounter, which would be unattainable with current staffing," he added.

Embrace technology

The second piece of advice he has is to accept the use of technology as soon as possible.

"In healthcare, we will continue to have downward pressure on reimbursement and increased demands for documentation and medical decision making," he said. "Why wouldn't we use AI-based systems to optimize our reimbursement for these encounters?

"The providers are doing the work, and are being required to do more and more for the same reimbursement code; why shouldn't they be fairly compensated for that encounter?" he concluded. "I would strongly encourage healthcare administrators and providers to look at technology to assist us in areas of practice management and documentation."