By Kumar Venkatesiah
Up until the 1980s, medical coders were responsible for manually transcribing every piece of information pertaining to a patient’s visit into an electronic code that would be stored in the hospitals’ patient database for insurance claims and other purposes.
The job of coders has evolved dramatically in the past couple of decades with the arrival of encoders. With encoders, medical coders today have their job greatly simplified since a lot of the coding process has been streamlined, and accuracy in patient records has thus been significantly enhanced.
However, as any healthcare professional would tell you, relying merely on encoders can be a risky proposition. Medical coders continue to play an important role in accurate documentation, owing to their knowledge of medical procedures and terminology. But change is inevitable. As technology continues to disrupt healthcare, there are bound to be changes in the way medical billing and coding is done.
The United States moved to the latest ICD-10-CM codes in 2015, which is now mandatory for all inpatient medical reporting. Variations of this code have been present in other countries, like the United Kingdom, Netherlands and Canada, from as early as the 1990s. The introduction of the new codes was opposed by many small physicians who estimated the cost of the transition to be anywhere between $56,639 to $226,105. Besides, the early days of the implementation also saw a noticeable increase in the number of claims being rejected due to wrong ICD-10 codes.
Despite the cost of implementation, there are already concerns that the introduction of ICD-10-CM could be too little, too late. Although the Centers for Medicare & Medicaid Services (CMS) — the organization responsible for releasing the ICD-10 codes — has been providing updated guidelines frequently, there are concerns that this is still not agile enough to capture all the latest diagnostic tools and procedures.
As medical research continues to grow at an astonishing pace, we may soon be reaching a breaking point where introduction of new guidelines every few months may simply not be adequate to capture all the latest technologies available to treat patients. From a coder’s perspective, the sheer number of new codes introduced to the system makes it difficult to accurately document patient diagnosis. Cloud-based healthcare IT systems, however, ensure that the code database remains updated at all times.
Artificial Intelligence In Medical Coding
The US Bureau of Labor Statistics predicts the number of medical coders in the United States to grow 15 percent in the ten year period between 2014 and 2024. But there are also concerns that the growing need for technology to take up coding could potentially restrict their growth.
Many hospitals today rely on what is known as ‘cliff notes’. This is a document that is generated by a computer after processing the medical records of a patient. Although the objective of a cliff note is to speed up the coding process, there are inevitably dozens of mistakes that a coder needs to rectify. According to one study, up to 90 percent of medical bills contain errors of one kind or the other. These errors are cumulatively estimated to be worth $750 billion a year. The role of a coder has thus remained as important as ever.
But a number of large hospitals have now begun experimenting with machine learning and artificial intelligence-backed computer assisted coding (CAC) systems. These solutions are helpful in identifying mistakes, fixing codes and generally assisting the coders with real-time feedback to improve their coding process. Not surprisingly then, a Frost & Sullivan report estimates the market for healthcare AI tools to be worth $6 billion by 2021.
Although they are a potential threat to medical coding as a profession, experts believe that technology shall never be allowed to completely take over coding and human intervention would always be deemed necessary.
Bringing Costs Down With Technology
Machine learning and artificial intelligence may never completely replace humans in the medical billing and coding functions. However, they may play an increasingly important role in bringing down the cost of healthcare. According to one estimate, the United States spends nearly twice as much as countries like Canada, Scotland and France when it comes to healthcare. Nearly a third of this goes simply towards medical billing and administrative costs.
Machine learning tools could efficiently interpret every line in a patient’s record — this helps a hospital avoid duplicate charges, upcoding and unbundling related errors. As a matter of fact, the push towards this may be coming from insurance agencies instead of the hospitals themselves. Issues like upcoding — where a patient is inaccurately mapped to a condition that is more serious than what they really are ailing from — can tend to shoot bills upwards, a major portion of which is borne by insurance companies.
Finding A Balance
A survey commissioned by the Royal Society found that when compared to functions like policing, personalized learning and transportation, consumers were least worried about the implications of artificial intelligence in healthcare. This is mostly because a significantly large chunk of “front-end” healthcare is still handled by medical professionals. As patients, you don’t expect to be consulting a robot anytime soon.
However, the implications of ML/AI on back-end healthcare cannot be overstated. In addition to making diagnosis more accurate and improving documentation, they could also potentially bring down the average costs of healthcare in the country.
At the same time, as we pointed out earlier in this article, machine learning and artificial intelligence could not be expected to completely take over the coding and billing processes in healthcare. The potential risks with wrongful diagnosis and the potential documentation errors are too high for these technologies to be run without human supervision. The ideal future is where these modern technologies would co-exist in perfect harmony with human coders that oversee the process to make the system as free from errors as possible.
About The Author
Kumar Venkatesiah is an eLearning consultant from India with more than nine years of experience. He works with telemedicine startups to help onboard new clients to the digital platform. You may reach him at firstname.lastname@example.org.