AI is making a fast progress in every filed we talk about these days. But will it take over the Humans jobs 100%? The Answer is NO, for at least few years to come. AI software’s are designed by Humans, which themselves are prone to make mistake so genetically speaking, unless software’s learn and evolve over time and develop the critical thinking, analytical skills, understanding with ever changing regulations, guidelines, coding requirements, correct interpretations of documentation, economical and employment requirements and ability to communicate for feedback, the AI based Medical billing and coding tools and the impact of Artificial intelligence on medical coding may not be widespread instantly. Even once Artificial intelligence (AI) passes it’s infancy stage, It is yet to be seen how governments let the use of AI based tools to spread as countries cannot afford to have it’s people without jobs. But having said that, AI will make a shift in industry with people moving to more specialised, analytical or Auditing job roles and some of the job cuts are imminent.

Is the future of medical coding doomed?

Since long we have been hearing that Medical coding is going away due to the technological enhancement in EMR (Electronic Medical record) and Artificial intelligence (AI). So let’s dive in.

Let’s be honest, it is true that Artificial intelligence will break into the system massively in coming years and more and more once it starts to evolve and gain experience once in operation over the years. Given that, the worst affected area would be employees.

Now coming to our topic today, impact of Artificial intelligence (AI) on medical coding. Computer assisted coding has existed even before 2007. Yes. It’s true that more and more coders are developing every single day and more coding positions are opening. Most providers, Physicians, physician’s assistant, inpatient and outpatient service providers do not want to do their own coding and rather want to focus on their clinical expertise.

Artificial intelligence (AI) in other walks of life and medical coding

We also hear that customer services will be replaced by chat box, your deliveries will be handled by drones but given the fact that one can have his accounts managed with QuickBooks and QuickBooks has already released the Beta version when it comes to Artificial intelligence, but companies still hiring accountants to get the accounts done, why?
The Answer is, companies do not want to risk their business as no one is aware of how correctly and accurately the AI base software’s will cohesively exist with your manual accounts in running, how they can correspond, how good they are picking up deficiencies and whether they can amend it or in trying to be efficient end up making underling changes which can be detrimental since your companies decision are based on information and correct financial health of the company can determine the right decisions.

For our Medical Coding and Billing scenario, since AI based software’s, at least for now in industry, are making progress/replacing hands on job kind of processes, such as manual entries replaced by OCR. OCR ( Optical character recognition) automatically digitize the manual records and as compared to human are more efficient, quick and error free, but when it comes to making right calls on communications, for example defending a client when once gets in trouble with IRS or court, since an employees on the other side of the phone knows all the laws and expertise and can handle the situation better instead of AI making a detrimental mistake. Given this fact, Providers do not want to do their own coding.

Adaptation of Medical coding to Artificial intelligence (AI)

AI can help decrease the human error when it comes to interpretation of Providers note ( If righty understood), and can also increase compliance, also on the surface, Medical coding may look easy, and that there is simply a code and that equates that to certain Diagnosis, Procedures, Durable medical Equipment and so forth. But, the fact and the matter is that there are a lot of changing guidelines and yes you can program the AI for that changes, but Since these changes are coming from WHO – World Health Organization, AMA – American Medical association, CMS – Centre or Medicare and Medicade services and people themselves not the AI is responsible for these changes, the AI will still be reliant on the information feed in them every year after the update and this feed into the system has been going on for years as computer assisted coding did exist for long. However, there are instances where computer will pick the very accurate code based on what’s been input but on other instances it can be completely off.

Check and Balances

The other thing we needs to consider that we need check and balances. Even with the AI taking progressive steps every year and leapfrogging the last year even in it’s infancy stage, and when compared to manual processes (Humans doing it) with all the procedures in place. A lot of providers/organizations end up doing things correctly whether intentionally, to commit fraud, or unintentionally, due to misinterpretation or mistakes, and then are providers who did the coding but did not provide the documentation that supports that code because documentation is what supports or back it up the coding and billing for those services.

Coding Complexities, interpretations and Guidelines

So a lot of providers end up overestimating or underestimating the services they have provided or they pick the right code for what they did but the documentation does not support what was done. Medical coding basically is a science and it’s not a code look up concept. EMRs (Electronic Medical Record) are already doing this and also there are a lot of organizations which provide services where they select the code for you which they think is the right code but later on have it scrubbed and reviewed by medical coder. This happens because there could be different interpretations based on different regional guidelines. These guidelines can change, and these codes can update and then it can become really a complex concept as you can have 5 coders and all five can come up with different codes of the same providers note especially in area like Evaluation and management where certain criteria is not specifically defined by Medicare and you have to developed internal policies on how your organization will comply and interpret these gray scenarios. Also providers documentation is not always one on one match to the code. There are always situations where you need clarification from providers or need feedback on interpretations. Sometimes we need the provider to submit an addenda and AI computer alone is not going to provide that, atleast for sometime to come in future, feedback because of the analytical work that is Involved.

Artificial intelligence (AI) and Medical Coder Future

No matter how we look at it, whether the analytical part of coding, the feedback, interpretations or communications with providers or others, there will still be shift in medical coding and for that matter in medical billing field and it may change. For this who have auditing credentials or can analyze more can have a long carrier in medical coding and billing as there might be Auditing role they take on until the AI can learn and equip itself to that extent but that does not mean that AI will learn and adapt the concepts 100% as human requirement will still be there because since Doctors notes does not a predetermined format and it can have all sort of information in it, misinterpretations by the system can still cause issues.

Down coding, up coding, and Legal issues

Since AI based tools can misinterpret the doctors note then it could also result in Down coding or Up coding. Down coding, which not only result in loss in revenue for provider, but it also skews the healthcare provider’s claims data and with the same breathe if for some reason AI based tool end doing up coding, it can carry potential legal and regulatory ramifications. But with Human since there is cross check of all the data before it is billed to insurance, AI based tool, once the code is finalized, will pass it as finalized code.

Data Breach

Data breach is another aspect we need to consider while we allow the AI to take the precedence. A Study conducted “Healthcare Data breaches, Insight and implications” sighted that IT related issues such as Hacking, malware, ransomware even with the existence of firewalls, and multistage protection, is the leading cause of data breach followed by authorized/unauthorized disclosures.
This means, AI Based solutions, unless they are Hack free, are susceptible to data leak intentional or unintentional and if the bot security is compromised, there can be serious HIPPA compliance repercussions.

Saving the cost on Outsourcing:

Outsourcing vs AI

Medical billing AI software’s will cost not only one off but also on monthly subscription based on the practice and providers requirement. Same as humans, they are not guaranteed to be mistakes free, at least for few years, since there algorithms are based on learning from mistakes same as Human being, If AI does come into Medical billing filed completely, it may still remain costly then outsourcing the medical billing and Coding to the third world where Medical coding and billing cost is 50 to 60% cheaper compared to US.

Economical and employment repercussions

Even if the AI gets to the point where it can do the job more efficiently, effectively then human beings, it will still likely to make mistake, one. Secondly, it is unlikely that any government or country can allow AI to eat up considerable number of jobs since not only AI will kill the employments which is bread earner for the families, it will badly affect the GDP of that country and in that case the outsourcing remains the most viable option.