How Automation Could Address the Medical Coder Shortage

As artificial intelligence (AI) continues to make inroads in virtually every industry, there is legitimate concern over what it will mean to jobs that have historically been labor intensive. Concerns over AI implementation have certainly been raised in medical coding. Right now, the biggest AI influence is automation.

 

A case in point is an automated coding platform now in use by New York's Mount Sinai Health System. The platform already codes about half of the system's pathology cases. Mount Sinai officials say they hope to eventually reach 70%. What does that mean for medical coders?

 

At the current time, medical coding jobs are not being threatened. Automation is actually helping Mount Sinai as it continues to grapple with the ongoing medical coder shortage. The human coders they do employ can continue working on cases not appropriate for automation. Meanwhile, the automated platform can handle mundane coding tasks that otherwise slow coders down.

 

It is important to point out that not all cases can be automated without compromising accuracy. If you are already a medical coder or biller, you know that some cases are so complex that it can take hours to pick everything apart and code it correctly. Automation does not change that.

 

Mount Sinai has been able to automate a rather large portion of its pathology cases because, by and large, those cases are not complex. Diagnostic procedures within the pathology ecosystem are pretty straightforward. Likewise for diagnoses resulting from diagnostic testing.

 

On the other hand, consider a case in which a patient is being treated for multiple comorbidities. Procedures and services overlap. Procedures for the same illness may be carried out over several days even though treatment is considered a single event. As a coder or biller, you know the drill. Complex cases need the human factor that AI just cannot provide.

 

To date, AI and automation have proven their worth in terms of cutting down on manual work. In the simplest of cases, medical coding is a routine exercise for a computer algorithm. Turning simple cases over to automation means human coders have to spend less time looking things up.

 

Even when an experienced medical coder knows exactly where to look for a set of specific codes, she still needs to take time to read through a digital document or conduct an online search. A computer algorithm can do that same work much faster.

 

Automation appears to be very effective for certain types of routine tasks. But as health systems are discovering, their platforms still need help by way of accurate clinician notes. Even AI-driven platforms are incapable of deciphering vague notes that do not clearly point to diagnoses, treatments, etc.

 

This part of the equation is not likely to change. As every medical coder and biller knows, clinician notes are not always what they appear to be. Human coders need to pick through page after page to try to understand a clinician's intent. The human brain can accomplish such a task given enough time and experience. Not so for an AI algorithm.

 

AI and automation can help the medical coding industry address the ongoing shortage of qualified workers. It is as simple as using automated platforms to manage mundane tasks that don't really require human intervention. That reduces the demand on coders and the need to hire more.

 

Will AI and automation eventually replace human coders? It is not likely. Computer algorithms have their limits. When it comes to medical coding, human brains can do a lot more even though it takes more time.