Artificial Intelligence Takes On Big Pharma
My nose is pressed to the glass as I watch thousands of human cellular samples whisk by, running through a large, elaborate machine. There are lab workers bustling around in neatly pressed lab coats, jotting down notes on their clipboards. It looks like a scene from a futuristic movie. But it’s not. This is real life, and the technology in use is poised to change the world of healthcare as we know it.
Fixing drug discovery using AI
“I never wanted to work for a pharmaceutical company,” says Amanda Guisbond, the director of corporate communications at Recursion Pharmaceuticals. “I hated the excess. I hated the idea of wasted efficiency.” But yet, here she is, fresh-faced, bright-eyed, and leading me around the Recursion office building. “What changed?” I ask. For her, it was simple: “I came to work here because Recursion is different. Unlike the other companies, we are all about fixing those inefficiencies.”
Often referred to as “broken, inefficient, and expensive” our healthcare system is one desperately need of innovation. Riddled with expensive problems that require expensive solutions, many are struggling to find the proper care they need at an affordable price. And Big Pharma companies are a huge part of the problem.
According to an article from the Washington Post, 78 percent of the newly patented drugs approved by the USDA correspond to medications already on the market. Essentially, these Big Pharma companies are rebranding (and redistributing) the same medications over and over again to make an easy profit, without actually bringing any new, innovative medications to market.
That being said, it can be hard for pharmaceutical companies to develop any kind of drug innovation at all since it costs so much time and money to do so. On average, it takes about fifteen years and $2.5 billion to fund a new prescription drug and bring it to market. And yet, 95 percent of those drugs will never make it to patients in need, failing their clinical trials in a manner of months. This kind of wasted efficiency, explains Ms. Guisbond, is exactly why many prescription drugs on the market are so expensive.
In order to be profitable, pharmaceutical companies have to make up for that cost by charging more for the medicines that actually work, or re-patenting the same ones over and over again. Americans spend about $1,443 out-of-pocket annually on prescription drugs. That’s more than any other developed country in the world―and patients are still suffering. Unfortunately, most pharmaceutical companies are content with approaching drug discovery in the same inefficient ways they always have.
Not Recursion. Focused on fixing some of the costly inefficiencies plaguing the industry, they’re using artificial intelligence (AI) technology to dramatically alter the entire drug discovery process. “If you look at the trends, it’s getting more and more expensive to get a new drug to the market. All of the low hanging fruits have already been picked,” says Lina Nilsson, senior director of data science product.
Using artificial intelligence, she says, Recursion reinvents the system, cutting costs and making it easier to bring newly discovered drugs to the people who need them most. And they’re one of the only companies in the pharmaceutical industry using this innovative drug discovery method. Using big data, Recursion quickly reads and evaluates all of the potential connections. Even the ones that may have not occurred to a human mind. “It’s just not [humanly] possible to write every rule about how biology and disease works,” says Ms. Nilsson, “[artificial intelligence] completely changes the game.”
Indeed, it does. Tina Larson, COO at Recursion, explains that the company uses artificial intelligence to run between 400,000 and 500,000 different experiments a week. In the past, numbers like this would have been completely unfathomable, taking years, if not lifetimes, to compute.
All of those hundreds of thousands of experiments have resulted in two clinical trials for new medications, or new uses for old medications. Though, the medications in the trial have yet to be approved for patient usage. The number might not sound like much, the team assures me this is a whopping success. And even if the test was one of the thousands that didn’t result in a new drug discovery, all of those data points can be used to drive other experiments down the road.
So far, Ms. Larson is proud of the way they’ve influenced the industry, “There are just so many inefficiencies in this existing [healthcare] system” she says, mentioning problems with insurance and billing, “and ultimately, in the end, there are only so many hundreds of dollars available. Any inefficiency that you can solve, in my opinion, helps.”
Fixing healthcare billing with AI
“Don’t even get me started on billing,” I chuckle. I was diagnosed with type one diabetes at a young age, and as soon as I was old enough, I was the one on the phone discussing billing problems with the insurance companies to learn how the system worked. Even as a teenager I found it frustrating that you could never walk into the doctor’s office, receive a bill, and know precisely what you were going to pay for treatment before heading home.
Instead, I often sat waiting around until a bill―usually one for more than I anticipated―showed up in my mailbox. Come to find out, a company in Utah is using AI technology to fix that, too.
These inefficiencies in the billing system, explains Usama Fayaad, advisor at OODA Health, are largely due to the sheer amount of manual labor required to process healthcare claims and payments. Instead of waiting for the manual back-and-forth between healthcare providers and insurance companies (this is often what takes so much time in the first place), OODA Health automates billing by using artificial intelligence to process healthcare claims in real-time, instead of waiting for a human employee to do the same thing.
With OODA Health’s software, patients can pay their bills immediately and physicians can worry less about when they’ll get paid and more about providing top-notch care. They’re automating the entire billing process with the click of a button.
Like Recursion, OODA Health, Mr. Fayaad says, aims to use AI to liberate the healthcare system of waste, inaccuracies, and a lack of transparency by doing the job cheaper, faster, and more efficiently. “In the US healthcare industry,” he says, “of the $3.3 trillion spent on healthcare, we estimate, some $500 billion is spent on administrative overhead. We believe that 80 percent of this cost is waste.” And their intelligence system can help eliminate that.
Though I am all about reducing waste, I can’t help but wonder about the AI naysayers. “What about the ones who still think AI will one day take over the world?” I ask. The naysayers, Mr. Fayaad says, are usually confused about what AI really is, and often believe AI is somehow simulating (or replacing) human intelligence. When in reality, he says, AI is about coming up with a practical solution to a problem that has nothing to do with actual human intelligence.
In the healthcare industry, AI technology is not an alternative to human intelligence. It’s about using it in conjunction with human intelligence to solve real-life problems. “Successful AI is hybrid AI, with most of the ‘intelligence’ being done by humans and most of the mundane processing being done by machines (robots),” says Mr. Fayaad. “In my opinion, ‘AI is not about replacing humans with robots, AI is about taking the robots out of humans.’” He’s referencing one of the mottos coined by MIT Media Lab.
Ms. Larson agrees, because without the help of AI to lead the drug discovery process, there’s no way Recursion employees could perform those thousands of experiments every week. “[At Recursion] we’re really applying AI to do things that supplement our people’s ability to do what they do best,” she says, and she’s confident the implementation of AI will change the landscape of the healthcare industry to be more patient-centric and a lot less wasteful.
“We need to get to a place where, you know, the patient truly is the customer of the healthcare system. Not the insurance companies or drug companies,” Ms. Larson says, mentioning that a lot of activity and expense in the industry doesn’t ever benefit patients in need. “Using [AI] technology to reduce the amount of waste or misspent energy in the healthcare system [is crucial] so that it can really be directed at better patient care.”
And Ms. Larson thinks that these ways of utilizing AI technology in the healthcare industry are only the beginning. One day, she expects, artificial intelligence machines will be the ones making the diagnosis, though they will never replace doctors or other healthcare providers.
“[These technologies] are impacting the healthcare [industry] in so many different ways,” says Ms. Larson. “Now increasingly, artificial intelligence algorithms are beating the best-in-class physicians and other experts in the world at diagnosing disease. And so that’s a really exciting [way of applying] AI to get patients better diagnoses in the first place.”