星空入口

Hint: Documentation may matter more than ever.

While the future of coding may seem scary as technology continues to play a larger role, Stacie L. Buck, RHIA, CCS-P, CPCO, CCC, CIRCC, RCC, RCCIR, believes that coders can find their way forward, according to her presentation 鈥淭he New Era of Medical Coding: Overcoming Workforce Challenges and Harnessing New Opportunities鈥� at AAPC鈥檚 2025 HEATLHCON.

Recognizing that the coding field is going to incorporate more artificial intelligence (AI) is key to a smoother transition.

See How Autonomous Coding Can Work

鈥淎utonomous coding means automatic coding with no human intervention whatsoever; the AI is coding the report and the report is going directly to billing,鈥� Buck explained. While there are other technologies available, which also refer to themselves as autonomous coding tools, they鈥檙e not truly autonomous, she said.

AI learns how to code via analyzing patterns between coding and documentation and its uses through training models utilizing millions of reports. This deep learning process mimics human cognitive processes. You can think of it as if multiple computer programs run on many levels simultaneously, resulting in all the analyses leading to a conclusion. While this technology works well for less complex specialties like diagnostic radiology coding, it鈥檚 not yet advanced enough to analyze conditions or specialties that involve fewer decisive patterns or are more complicated.

Your mileage may vary with these processes, because they, like human coders, need time to learn and effectively navigate updates or other changes. And of course, there are a lot of gray areas in coding and, in such situations, subjectivity may come into play.

Prepare the Future Workforce

One development already becoming apparent is that entry-level coding positions, like in the field of diagnostic radiology, may become less plentiful as AI or other technology can handle the coding. 鈥淭here will still be a need, because AI can鈥檛 code all of it without a human. But a lot of that is going to be automatically done by the AI technology,鈥� Buck said.

Similarly, there鈥檚 probably less of a chance that a coder preparing to enter the workforce will work as a traditional full-time production coder. 鈥淵our job is going to look different in the future. I鈥檓 confident you鈥檒l have a job. I don鈥檛 believe there鈥檚 any reason to think that medical coders are just going to disappear, and we won鈥檛 have a need for them, but it鈥檚 probably going to look different than it does now,鈥� she said.

While this may sound alarming, Buck shared her beliefs that coders in the future will get to focus on the more 鈥渇un鈥� aspects of coding 鈥� the complex situations that require the thoughtful application of knowledge and expertise.

AI and healthcare industry stakeholders see AI as a helpful tool in diminishing the perennial staffing challenges 鈥� especially shortages 鈥� inherent to the healthcare industry at large.

鈥淲ith procedures being done, there鈥檚 this big growth in procedural volume across many specialties. So, what happens? Typically, what you鈥檙e doing is you鈥檙e looking for more employees. You鈥檙e looking for more coders as you have to code more procedures. You have to add more staff or outsource overseas 鈥� which is something that has been occurring for a very long time in our industry, with a lot of jobs for medical coding going overseas. It鈥檚 a staffing issue 鈥� and also a cost-containment issue,鈥� she explained.

Other aspects of employee dissatisfaction or transition also come into play, like when employees feel overwhelmed by a job and choose to leave or reach retirement age. Buck believes that AI can help mitigate all of these challenges, especially because it can be used at various points of the revenue management cycle.

Documentation Still Matters 鈥� a Lot

The training models used to develop AI require accuracy, and since computer automated coding (CAC) is trained on documentation, it has to be extremely accurate. Problematic documentation may include documentation that is ambiguous or incomplete.

Complex documentation can also be problematic at this point because it鈥檚 so much less straightforward. Specialties, including those that perform surgical procedures, have a lot of complexity; so while some AI platforms can accommodate such data, more human oversight is required.

鈥淭hose are the types of things that you would still have to have complete oversight over for more difficult cases,鈥� she said.

Plus, with the gray areas that naturally arise when coding procedures or diagnoses concerning the many variables of the human body, AI can sometimes get things wrong. When documentation is incomplete or there aren鈥檛 enough examples of certain cases to give AI a comprehensive perspective, AI can assume biases, which can produce wrong results.

There are also so many diagnostic and coding situations where human intuition is crucial to accurately reporting a patient鈥檚 health and condition. At this point, while AI processes can do some of the routine and repetitive coding work, humans are superior in accurately portraying a person鈥檚 medical narrative, as well as auditing the results.

Take These Steps, Personally

While the future of coding may look different, human coders will remain a keystone in the healthcare industry. According to Buck, some roles that will become increasingly important include coding quality analyst, revenue cycle specialist, revenue integrity specialist, charge capture analyst, denials and appeals specialist, and healthcare data analyst.

If you have a skill gap now, look into a professional development plan; do some research and define your goals, Buck recommended. Be aware of opportunities that may arise at your organization, especially since many organizations like to promote from within. And Buck鈥檚 last piece of advice: Don鈥檛 forget to network, because then you鈥檙e casting a wider net for opportunities and connections.

Rachel Dorrell, MA, MS, CPC-A, CPPM, Development Editor, AAPC