Augmenting Coding And Billing With AI

July 10, 2018
732 Views

 

Medical coding and billing is one of the fastest growing jobs in the healthcare field, with a predicted 15% job growth expected between 2014 and 2024. But can those numbers hold up to the rise of AI in the workplace? Though on the surface it may seem like coding and billing may be the perfect job to automate, in fact, it’s unlikely that AI will ever do more than enhance the process. In practice, medical coding and billing is too complex and variable to lose the human touch.

The Nuances of Medical Billing

Part of what makes medical coding and billing so difficult to automate is that each coder needs to build an intimate relationship with their employer. That’s because there’s not enough time for doctors to write detailed notes during each encounter. Instead, they jot down generic notes, and then coders interpret that information and break it down into the appropriate medical codes for insurance and billing purposes – and accuracy is vital.

To perform this job well, medical coders and billing specialists are trained on a range of topics, including HIPAA issues, EHR administration, and medical billing software, but they also need to be familiar with the thousands of unique codes that describe ailments and procedures. It’s a vast skill and knowledge set, and that’s precisely why the human element is so important.

Supplementing With AI

For healthcare staffing experts faced with hiring coders, understanding the interaction between coding and billing experts and AI is critical to assessing hiring efforts in the next several years. Should hospitals be hiring more coders, or will AI help them to scale back?

READ
Are Healthcare Leaders Ready For AI-Driven Change?

The best description of next-generation medical offices is one that combines both AI and coding experts, but those experts will still dominate. Individual coders will still do the interpretive work of translating medical records into diagnostic codes, and then pass that content off to software that can check for errors and enhance coding accuracy.

Another billing and coding task suited to AI is record reconciliation. Many patients have conflicting diagnoses in their files, or medications that are out of date. AI can easily sort through past files and documentation to determine which information is out of date, what may need to be added to patient files, and what files are no longer relevant at all. This reconciliation work can also make patient files much more manageable for doctors and coders.

Better Health Through AI

There’s a lot of information packed into diagnostic codes and billing documents, but when doctors and coders look at these documents, they tend to become a blur of numbers and letters. In other words, no one is studying these documents to learn more about patient health patterns. And this is exactly where AI has an advantage.

AI technology can be trained to identify repeating patterns in medical coding and billing and identify those patients who were readmitted within a 30-day period. Based on those patterns, the technology can then look at single case billing codes and predict the likelihood of patient readmission. Since this kind of turnaround is bad for patient health – and bad for doctor’s performance statistics and reimbursement rates – the combination of excellent manual coding and AI analysis could provide important treatment breakthroughs for the sickest patients.

READ
The Impact of Big Data In Healthcare Analytics Career

Payment Productivity

From an administrative perspective, AI augmentation for medical billing and coding may also provide significant support for billing and insurance reimbursement. First, since AI can help identify errors in medical coding, which can prevent lengthy reimbursement cycles. When an insurance company spots a flaw in coding, they may reject the claim, leading to a cycle of claims and bills and phone calls that can leave patients stranded and healthcare providers unpaid.

In the near future, insurers may make decisions about what organizations and providers to partner with based on long-term data mining called “care transition analysis.” This information can help them identify suspicious billing patterns, separate hospitals with a high rate of rehabilitative success from those with extreme readmission rates, and even develop prepayment solutions for quicker transactions.

Healthcare is constantly changing and billing and coding practices shift at least as quickly – but it’s not the coding that is preventing AIs dominance. No, billing and coding will remain firmly in the human domain precisely because the field demands intuition and interpretation. Though new AI can read medical scans and interpret test data, in the case of coding, what they’re interpreting are doctor’s notes and impressions. The work is subjective and messy – and it requires a human touch.