The Evolution of Medical Records: AI & Autonomous Coding
Search For Schools
“We need to embrace this technology. It will become the standard.”
Keith Olenik, Chief Health Information Officer for AHIMA
Administrative costs account for up to a quarter of all US health expenditures, with billing coding costs being one of the top drivers (JAMA 2021). Staffing shortages and coding complexity have worsened matters. But major breakthroughs in automating aspects of the medical coding process may point toward a long-term solution. Already, some provider organizations have implemented autonomous coding in high-volume outpatient specialties such as radiology, pathology, emergency medicine, urgent care, and primary care (AHIMA 2023).
The rise of autonomous coding may appear at first glance to target the role of the medical coder for extinction. In fact, the opposite is true. Increasing integration of autonomous coding systems will allow medical coders to undergo a major evolution, focusing on complex cases and leveraging their expertise in healthcare, data analytics, and technology to help organizations, clinicians, and researchers level up. Auditing will be another prime opportunity for coders to validate the result of new technology involved in the code assignment process.
To learn more about the evolution of medical coding, and the important role that automation plays, read on.
Meet the Expert: Keith L. Olenik, MA, RHIA, CHP
Keith Olenik is the Chief Health Information Officer for AHIMA. He has more than 35 years of experience working with healthcare delivery systems as a member of executive leadership and a consultant. He specializes in streamlining business operations, evaluating and implementing information technology solutions, and enhancing productivity through process improvement.
Olenik holds a bachelor’s degree in health information management from the University of Kansas and a master’s degree in health services management with an emphasis in computer resources management from Webster University. He has a long history of volunteer leadership at AHIMA, having served on the boards of directors for AHIMA, the AHIMA Foundation, and the Council for Excellence in Education. Olenik also served as chair of the CHPS exam construction committee and the EHR Practice Council.
The Evolution of Medical Coding
“The medical coding function drastically changed in the mid-1980s, when the government implemented what we now call Diagnostic Related Groups, or DRGs, to reimburse Medicare patients,” Olenik says. “Prior to that, coding was basically a back-office function used to track diseases and procedures.”
The advent of DRGs attached a financial aspect to medical coding. In Olenik’s view, this might have derailed some of the more clinically-relevant aspects of medical coding and reoriented it toward being more reimbursement-focused. As DRGs evolved, other payment methodologies connected to codes added to the complexity that medical coders experience today.
Contemporary medical coding is a non-trivial task, with coders expected to decipher a large number of documents related to a patient’s episode of care and select the appropriate codes from a large classification system that corresponds to a variety of contexts in the associated documents and their guidelines (Nature 2022).
“The government has standard coding guidelines that every organization submitting specifically to Medicare must follow,” Olenik says. “Then every payer has also created their own list of coverage determinations because we’re not all covered by Medicare. It’s like you’re playing football and every team and stadium has its own rules. The complexity of what a coder has to deal with to get the organization paid for services provided, or for an individual to understand what they’re going to have covered by their payer based on what the doctor says and then what gets coded, is pretty crazy.”
The Rise of Autonomous Coding
According to AHIMA, chronic medical coding challenges and revenue cycle management staffing shortages have left healthcare organizations struggling to contain costs while ensuring coding accuracy and continuous compliance. Autonomous coding, which leverages machine learning algorithms, has emerged to meet those challenges.
“Autonomous coding has a significant opportunity in that individuals are human, and humans make mistakes,” Olenik says. “With automation, you reduce your labor costs. You also increase the potential for increased quality, as the autonomous coding algorithm will not get distracted. It’s not going to misinterpret things. It has the potential over time to learn and do a better job than humans.”
Autonomous coding can handle the complexity of edits in coding more easily. But it isn’t good at handling unique situations or outliers, and there will still be the need for human intervention by the medical coder. Ideally, autonomous coding fits into a coder’s existing workflow, running in the background and taking a first pass at coding charts. Any event which can’t be handled through automation gets flagged for review. This removes an element of monotony and reduces the number of overall charts a coder has to examine, both of which should, in turn, boost accuracy and efficiency.
It’s still relatively early days for autonomous coding; Olenik estimates that less than 15 percent of healthcare organizations have implemented autonomous coding solutions. Upfront cost remains a major deterrent for many healthcare organizations, but prices should come down as major vendors of autonomous coding solutions begin to recoup some of their initial investments. Yet another deterrent is psychological: the discomfort that comes from replacing human jobs.
“People are still leery of technology taking over jobs,” Olenik says. “Some of it is self-preservation. But there’s a lot of upskilling, and the knowledge that a coder has can be applied to many other places. We need to embrace this technology. It will become the standard.”
The Future of Medical Coding (and the Medical Coder)
The future of medical coding will revolve around the evolution of the medical coder’s role. Today’s coder is already well equipped to pivot into more complex and advanced responsibilities, with an educational and experiential background that includes medical terminology, pharmacology, disease processes, reimbursement methodologies, regulatory requirements, and the general healthcare delivery system.
That knowledge and expertise can apply to several other areas. Olenik envisions medical coders utilizing their experience as practical data analysts to act as translators and bridges between agnostic data and clinicians and researchers.
“A coder has the ability to help organizations make sense of all that data that’s potentially being collected outside of the healthcare organization, which might be related to population health or public health,” Olenik says. “They also can help researchers make sure they’re pulling the right data to answer their questions. They can help with outcomes for the researchers who are doing clinical studies. They can help them make sure they’re using the right data because they understand the various sources, and they understand what the data means through that terminology of pharmacology and disease processes.”
Coders have yet to be given the full level of respect that they deserve. Their role in financial reimbursement is vital, and their job also has an enormous clinical impact: the wrong code can have a significant effect. Perhaps coders have even undervalued themselves, overlooking some of the clinical purposes their coding can have in patient outcomes, care quality, and public health research. Autonomous coding should allow the current and coming generation of medical coders to fulfill their latent potential.
“The bottom line is there will be jobs for medical coders,” Olenik says. “They’re just not necessarily going to be called medical coders. Competencies are more important than titles.”