Unshackling medical practices: How AI can reshape coding
Navigating recent evaluation and management coding updates with automation can reduce errors, improve efficiency and combat physician burnout.
Evaluation and management coding guidelines don’t change often. In fact, the recent changes in 2021 and 2023 were the first guideline updates since 1997.
Adjustments after such a long period of stability can be frustrating for coders, not to mention practice administrators who may see an uptick in evaluation and management (EM) errors after the updates take effect.
Many coding teams may be caught in this struggle after the 2023 EM coding changes. But what if you could accelerate the learning curve of guideline updates and minimize the subsequent rise in EM errors?
The truth is, you can. The secret is to leverage artificial intelligence (AI) for coding automation. Let’s explore five scenarios where AI streamlines guideline updates and avoids common EM errors to show how this works. But first, how does AI coding provide value to a practice?
Improve coding, transform a practice
According to a study co-authored by the American Medical Association, burnout among physicians hit an all-time high at the end of 2021 – with 62.8 percent of physicians reporting some symptoms of burnout. One of the main drivers cited for this burnout is administrative tasks. In fact, according to a Doximity poll, 46 percent of physicians believe reducing administrative work is the best way to alleviate burnout.
Reducing these tasks is where AI technology can help.
Coding automation helps to minimize physicians’ time spent on the coding process. Instead of stressing over selecting the correct EM code in the few minutes that physicians have between patients, AI completes this task based on the full clinical documentation. With this assistance, physicians can be more productive, focused and at ease, ready to devote their complete attention to the next patient.
However, accuracy is essential. EM errors can cause denied claims that result in delays and time-consuming resubmissions, while incorrect leveling can leave deserved reimbursement dollars on the table. Coding automation helps resolve these problems. With its enhanced accuracy, the technology has the potential to improve revenue capture and increase margins. What’s more, physicians potentially can have less administrative burden – freeing up time for patients and improving overall well-being.
Lastly, coding automation can eliminate the need to retrain staff when guidelines change. Because autonomous coding technology updates nearly instantly, it can integrate any new guidelines that impact EM code rules and accuracy. As a result, the whole practice benefits. The revenue cycle accelerates, and physicians and staff are less preoccupied with EM errors that distract them from patients.
Scenarios in which AI can trim errors
How does AI-coding technology work in practice? Here are five scenarios where coding automation can help prevent errors and adjust quickly to guideline changes across specialties.
Changes to COPA. One of the main 2023 guideline changes involves the number and complexity of problems addressed (COPA). Consider the following scenario. If a patient is rushed to an emergency department with a potential heart attack, doctors may consider a percutaneous coronary intervention procedure or thrombolytic medications to rule out the condition. As part of the 2023 guideline updates, the EM code should reflect the clinical team’s effort when considering those options, regardless of whether either treatment is actually administered. Coding automation technology reflects this change and picks up evidence of any treatments considered.
Additional resources. The new guidelines also credit physicians for information obtained from an independent historian – an individual who provides history about the patient or the patient’s condition. This individual can be a parent, child, spouse or other witness who provides information when the patient is unable to do so. This information can be buried within different sections of the note, depending on where the physician decides to document, and it can be easy to miss. AI will pick up this important component of the visit and make sure the EM level reflects all of the work, thought and discussion required.
Risk updates. While assessing risk has always been a part of EM coding, the 2023 guidelines have made it a key component, with some subtle differences among straightforward, low, moderate and high risk. For example, in the emergency department, orally administering morphine falls under moderate risk, while administering morphine through an IV is high risk. The method of administration may seem like a small detail, but it is an important one that inexperienced coders could easily overlook when assigning the appropriate EM code. AI won’t miss it.
Double counting. With the new guideline changes, some tests are separately reportable while others are not. For example, if an EKG interpretation is counted as part of the EM, then an EKG procedure such as 93010 should not be coded, as the physician receives credit for it through the EM. In other words, it’s important to avoid double counting. As with the examples above, coding automation technology applies these rules comprehensively to ensure compliance with guidelines.
Missing elements for EM levels. Burnout is a major problem among physicians, and primary care doctors are not an exception to this trend. After seeing a patient, it’s their job to document the encounter. Of course, there is usually a long line of patients waiting to be seen. In this environment, physicians can easily overlook elements like point-of-care tests, complexity of comorbidities and patient history, which all matter for correct EM coding. However, coding automation can systematically examine all of these details, assigning the appropriate EM codes for practices to receive proper reimbursement.
Guideline changes no longer need to be a cumbersome event that disrupts a coding team, causing mistakes and payment delays. As you can see, AI technology can help a practice adapt effortlessly to changes and ease the burden on physicians and coders.
Taylor (Ross) Webster is head of coding quality at Fathom, a health technology company that uses deep learning AI to automate medical coding.