In 1997, Jonathan Bush (a first cousin of George W. Bush) founded a women’s health and birth practice called Athena Health in San Diego, but he was unprepared for the costs and burdens of reimbursement by insurers. The bitter experience with that failed business suggested a different venture: Athenahealth, which he founded later that same year in order to offer doctors a service that processed medical bills over the Internet. In 2006, the company launched another service: the first electronic medical records software that physicians didn’t buy as a released product, but subscribed to as a service on the Internet. Other health-care management services have followed. Using the cloud for health-care management is an innovation that has rewarded Athenahealth with a public offering and swift adoption by many medical practices: around 30,000 of them now use its services. Bush, an ardent free-market enthusiast, believes these services are popular because they tap into an overwhelming sense of frustration with the complexity and opacity of the U.S. health-care system, which he calls “The Beast.”
Jason Pontin, the editor in chief of Technology Review, spoke to Jonathan Bush at Athenahealth’s headquarters in Watertown, Massachusetts.
TR: Explain why managing electronic medical records and processing insurance claims in the cloud is so superior to traditional methods.
Bush: Well, of course the other guys don’t suck. They make the best darn hammer and chisel in the railway business, but when you show up with a steam shovel, even a shitty one, you know, John Henry’s going to die.
Do I understand you?
When we began, we sucked! When you look at the early versions of AthenaNet and compare it to all the wonderful features and modules and other things that get sold on [software] discs, it sucks. But our developers got paid to stick around and watch, see what happens, and tweak and re-tweak and re-re-tweak.
You’re saying: the cloud has encouraged continuous improvement of your service, whereas software companies are committed to the interrupted, incremental improvements of product releases.
New code goes out every night, and the front end of the application, the actual layer that customers use, changes eight times a year. That gives us the opportunity to try stuff, have it not work, and change it later. The quality-assurance process for traditional old-school software is exhausting. ‘Cause once you ship a million discs, you’d better not have anything broken. We get to actually have things be broken, because it changes the next morning.
And the old kind of product development suppresses the desire to do anything innovative.
You’re singing it! You’re singing it!
Why is being paid for services, rather than just selling software, a positive advantage?
Our big discovery is we’ve built a business model that asks us to build the right technology. It isn’t that we are uniquely able to build technologies; it’s that we are required by our business model to keep focusing on making them work. Because we’re only a software-enabled service: we only get paid when the service is completed, when the customer actually has money in the bank. If Athena didn’t agree to do all of the work associated with settling a claim or managing a medical record, then we wouldn’t have that selfish incentive to build out technologies that are invisible to our clients.
Is it true that you work around the unwillingness of doctors to do data entry by scanning anything they’ve written on paper and faxed you?
Right, but they’re not as troglodytic as people think. I bristle a tiny bit when these insiders talk about health care and tell the world that the problem is that doctors are troglodytes—well, they’re businesspeople. They are being asked to pay $50K up front for software that will make them move patients through the office more slowly, and whose only benefit is to digitize the information that they are writing down. I could codify my e-mails, you know: I could categorize them all into 73 dozen different buckets and run through them all and make sure that they were all categorized properly and color-code them. But I don’t, because I already know what they said. And I don’t really care.
And it’s not your job.
I’m sure that some constituency could benefit from color-coding all of my e-mails, but not me. And doctors, I think, are in the same position. Clinical data is increasingly required for doctors to get full payment. But we don’t care how the doctor collects that data. If the doctor needs first to collect her notes by dictating, they talk. If the doctor prefers to color in little forms that they’ve been using for years, then color in the form. If they’re the kind of doctor that’s got two iPhones on their belt, then they click.
With traditional insurance claim processing, how often are claims rejected?
The last time we looked at the outside world, about 33 percent of claims did not go through on the first try. And they had to be redone some number of times. I think it was 2.6 cycles of claims per encounter.
Imagine the impatience, the frustration, and the sheer costs associated with that.
We’re now below 5 percent of claims that need to be rehandled.
Do you benefit from the complexity and opacity of the U.S. health-care system?
We call it “The Beast.” Every doctor goes through life with this slight prickle on the back of her neck that there is a giant black primordial octopus whose tentacles are connected underneath the desk and in the exam room and in the file room and it’s slowly kind of wrapping these tentacles around their life and crushing. Whether they make a million dollars a year or ten dollars a year, they all have this sense of being crushed by some mysterious force that is outside of their control and that does not have a heart.
Why is the system so perverse?
These rules were all the well-intentioned campaigns of some congressman back when some mobster in Brooklyn did too many mammograms, right? And now all these rules, these never-let-this-happen-again laws, have piled up in the bottom of the doctor’s file room and it’s just almost impossible to breathe. It’s impossible to get excited about moving your practice to a new place, a new metaphysical place. What we are doing is taming that beast.
You don��t think President Obama’s health-care reform will reduce the complexity in the U.S. health-care system. You believe it’s just going to increase the complexity?
Look, a system can be complex but beautiful. The most beautiful systems in the world—like biological systems, or retail supply chains, or equity trading markets—run more cleanly and sophisticatedly than many others. These systems that we take for granted are insanely complex: probably a lifetime can be spent studying them, and you’d still not understand them. We’re calling this health-care reform, but it really looks more like one of Mao’s five-year plans: a “great leap forward.” It is a conscientiously-thought-through set of improvements, but they’re conscientiously thought through by a central authority that is going to force everyone to comply. So the natural forces that keep complex systems running smoothly will not be in play, and you’ll end up with these Kafkaesque moments. Any system with so many players that is inflexible, unresponsive, and carries the force of law is going to create unfortunate side effects that won’t go away.
Technology helped increase the costs of health care. Is technology part of the solution in bending the cost curve and controlling costs, too?
The first thing that technology can do is tame the Beast. But then, technology can be used in more creative and subtle ways. Doctors are the people who are doing Pap smear samples on women to screen for cervical cancer—$250. Absurd, right? Doctors are doing routine surgeries that would be much better done by a technician whose only job is to execute that one surgery, like we’ve seen with eye surgeries. But all of that will require a level of process control. Today, we don’t have a healthy technology system around process control. We’re the first player in that, but there’s going to be others now that the cloud-based business service model exists. We expect to see lots more competition.
This new data poisoning tool lets artists fight back against generative AI
The tool, called Nightshade, messes up training data in ways that could cause serious damage to image-generating AI models.
The Biggest Questions: What is death?
New neuroscience is challenging our understanding of the dying process—bringing opportunities for the living.
Rogue superintelligence and merging with machines: Inside the mind of OpenAI’s chief scientist
An exclusive conversation with Ilya Sutskever on his fears for the future of AI and why they’ve made him change the focus of his life’s work.
How to fix the internet
If we want online discourse to improve, we need to move beyond the big platforms.
Get the latest updates from
MIT Technology Review
Discover special offers, top stories, upcoming events, and more.