Thrive in The AI-Rich World

February 10, 2020

Whether you code in house or outsource domestically (or even internationally), you have probably heard the talk about Artificial Intelligence (AI) and wondered about its impact on medical coding. In this brief, we will propose a historical analogy to explore AI’s likely impact on coding, and how you can benefit from the wave of automation that is coming.

Code like a Chessmaster

In many ways, medical coding bears a similarity to chess. Both have a defined set of rules, a set of adversaries, final outcomes that rely on humans executing not just a science but a form of art, and finally, both fields are fundamentally changing due to the influence of powerful computers.

In the 1980’s and 90’s, Gary Kasparov was the most dominating, masterful chess player the world had seen in generations. In 1996, he staged a famous exhibition match against Deep Blue, a computer built by IBM. He narrowly defeated the machine. The following year, it defeated him. That machine cost millions of dollars and tens of thousands of man-hours to build – but today, a chess playing algorithm vastly more talented and powerful lives in the chess app on your smart phone.

This exponential trend not only in computing power, but in automation capability and cost savings is coming for every industry, and medical coding is no exception.

The Centaur Approach

Given Gary Kasparov’s infamous – some might say ignominious – defeat to Deep Blue, you might be surprised to learn what he’s working on today. He’s pioneering the development and use of what has been called the “centaur” approach; humans and computers working together. Even today, these teams perform far better than either computers or human chess players working alone. The same approach is valid outside the world of chess in many other fields where AI will prove a powerful automation tool in the coming years.

The computer that beat Gary Kasparov could best be described as employing brute force – the capacity to calculate millions of outcomes in seconds. This capacity has only grown in the 20 years since that famous match, and today’s AI algorithms employ techniques orders of magnitude more sophisticated. But they are still not, in any way, “intelligent” – and good outcomes in both chess and coding still require human intuition, strategy, and judgment – which cannot now and may never be automated. The most successful outcomes for the future will almost certainly be achieved by the centaur approach – the wisest, most experienced humans, with the best, most comprehensive data, utilizing the most powerful AI algorithms.

According to Kasparov, a “weak human + machine + better process was superior to a strong computer alone and, more remarkable, superior to a strong human + machine + inferior process.” (1). This holds true in the medical and billing field as well. According to a recent publication in The Lancet, “Machines will not replace physicians, but physicians using AI will soon replace those not using it. “ (3). The same goes for coders and billers.

Riding the Oncoming Wave

As computing becomes more powerful, cheaper, and more reliable, the risks of outsourced coding will exceed the costs. Humans working in concert with highly trained, sophisticated AI can expect to execute the majority of medical coding in the near future. (2)

So what can you do to thrive in the AI-rich world?

Eliminate manual steps. Are your human coding staff manually scanning reports, or keying charges? Take advantage of the capabilities of electronic interfaces to eliminate unnecessary intermediate steps. Take Kasparov’s advice to heart: weak humans with good processes and machines beat strong humans with weak processes and machines. Clean up your processes!
Promote/retask your human staff from front-line coders to auditors, denials follow-up, managers, and optimizers. Only code from scratch those encounters an AI doesn’t have high confidence in coding independently.
When employing an AI tool, look for a team with a deep understanding of EMR data. AI is only as powerful as the data it has access to is deep, comprehensive, and structured usefully. If all you have is a free text file of notes, you don’t have much.

News

Code like a Chessmaster

The Centaur Approach

Riding the Oncoming Wave

CHECK OUT HOW ARTIFICIAL INTELLIGENCE CODING WORKS IN THIS VIDEO

References