Taking the time to do it right: Automation in medical coding comes of age, thanks to Fathom

Lightspeed backs Fathom’s automation platform to reduce billions of dollars of waste in the health system

Sometimes, the best solution for a complicated problem takes time.

For years, hospital administrators, physician practice owners, and healthcare investors alike have viewed medical coding as one of the most expensive problems in their revenue cycle. Thanks to a tangled web of inefficient and error-prone practices, over $10 billion is spent annually in the U.S. on medical coding: the process of translating clinicians’ notes into numerical codes that represent the services provided, which the healthcare provider then uses to bill a patient and their insurance.

After looking for a differentiated solution to this problem for some time, we at Lightspeed are excited to announce our investment today in Fathom. Founded by Andrew Lockhart and Chris Bockman, Fathom automates medical coding and delivers superior efficiency gains to healthcare organizations across the industry. We’re thrilled to co-lead Fathom’s $46 million Series B alongside Alkeon Capital.

During my clinical rotations, I saw how recent attempts to reduce medical coding costs were Band-Aids that didn’t address the root of the problem. Healthcare organizations have tried hiring more coders offshore, offloading the entire coding function to revenue cycle management companies, and worse yet — shifting admin responsibilities to already overworked clinicians.

Unfortunately, all of these approaches still rely on human labor. With over 200,000 possible codes — and growing — that have varying guidelines and provider preferences, even with highly trained coders, the process is slow, expensive, and inaccurate. Concurrently, the pool of medical coders in the U.S., estimated at about 125,000, continues to shrink as these professionals choose to seek other roles or retire (the average age of a medical coder pre-pandemic was 54). To reduce cost and streamline medical coding for healthcare organizations already operating at razor-thin margins, technology is the way forward.

A truly automated solution has been elusive, however. Computer-assisted coding (CAC) has been around for nearly two decades, but today is still essentially a rudimentary speech-scanning technology that relies on medical coder usage with modest productivity gains. As a result, only about 10–20% of providers are estimated to have adopted CAC. In the last decade, some have attempted automation, but those attempts were built on a foundation of rules-based engines that require constant and laborious up-keep, resulting in limited cross-specialty adoption.

The promise of AI-driven solutions certainly generated investor interest back then, but neither the technologies behind the start-ups nor a willingness from healthcare organizations to adopt new technologies was as far along as most had hoped. As a result, some investors felt burned, and others were simply disillusioned by the pace of development and abandoned the opportunity. However, having invested in a number of core health IT companies over the last decade, I always believed that among the myriad functions of the back-end RCM web, medical coding automation was a matter of when, not if. One has to look at the fundamental levers of what could lead to real automation and track the underlying principles that could enable it. When I met co-founders Andrew and Chris, I realized quickly that they created Fathom in perfect alignment with these principles.

Andrew and Chris stayed disciplined on the core thesis: that a data-driven, deep-learning based foundation will achieve real automation at scale. Today, Fathom’s automation rates are industry leading and 30–50% higher than competition, and we anticipate this lead to grow. Fathom achieved this because Andrew and Chris believed the engineering and data processing architecture was as important as the data itself to enable cross-specialty model training and deployment. This focus enabled deployment across more than 2,000 provider sites and is a result of the founders not settling for incremental improvement early on at the expense of leap-frog productivity gains in the long-run.

These may seem common sense now, but over the course of the last five years, Andrew and Chris persevered and stayed true to the north star of real automation when many others did not. The data required for this ambitious approach was more than anyone anticipated, the engineering more challenging and the patience required more than most founders and investors could bear. To overcome and succeed in all three parameters makes Fathom special, and has positioned the company to lead the industry in this critical opportunity for medical coding.

We look forward to supporting Fathom as it scales in depth and breadth of its impact across healthcare organizations and we welcome those who would like to join and contribute to the journey.