Aaron Patzer, founder of personal finance platform Mint, is back in the startup world. This time, he’s building an artificial intelligence platform to simplify perhaps the only industry more complicated than finance — healthcare. And, he’s bringing a couple of Emory University physicians with him.
Patzer’s entrepreneurial drive stems from dissatisfaction with overly-complex legacy systems. He founded Mint.com in 2006 after becoming frustrated with how difficult it was to use the most common personal financial platform of the time, Intuit Quicken.
In 2009, when Mint hit 1.5 million users and had raised a little over $30 million, Intuit bought the startup for $170 million.
In 2013, Patzer left Intuit to found Fountain, an on-demand answers app that connects users with curated experts in almost any field — health, financial, legal, home improvement, even veterinarian medicine. After raising $4 million, Fountain was purchased by home services company Porch.com.
While taking a break from high-stress startup life and partially relocating to New Zealand, Patzer began helping his sister, an Emory University medicine professor who specializes in kidney transplant research. He used his expertise in artificial intelligence and predictive algorithms to create models that could predict the success of kidney transplants.
“So that got me onto what machine learning and good software could do in healthcare,” Patzer tells Hypepotamus. He had the connections to really dive into the subject — in addition to his sister, his brother-in-law, Dr. Justin Schrager, is an Emory emergency room doctor. His sister-in-law also practices emergency medicine.
Patzer learned that, though average ER waits are three to four hours long, 90 percent of patients have “relatively straightforward cases.” Nurses and doctors spend much of the patient’s and their own time asking the same basic questions: why are you here, when did this happen, what have you tried, what makes it better or worse?
His family members wanted to know: couldn’t Patzer come up with a way to make the process faster and easier? They proposed a check-in app that could be filled out by patients when they first come into the ER waiting room. Could it be that simple?
“Well, it turns out that it can, and we can do it quite accurately,” says Patzer. He started visiting emergency rooms around Atlanta to learn more about the roadblocks and obstacles providers faced. He searched for replicable models that he could use to build an algorithm.
In 2016, Patzer first established Vital. Like his previous startups, the software is entirely geared towards making tedious processes easier. When patients come into the ER, they register through the cloud-based, HIPAA-compliant web app and enter all of the information about their current condition — what hurts, what helps, and what they’ve done about it thus far. They can even take a photo of a health insurance card to speed up that process.
Using their name, the platform finds the patient’s existing electronic health records, any past tests and medication regimens, and any other relevant medical history. As doctors order tests and labs, the patient gets text-messaged about their status.
On the provider side, doctors and nurses are able to see all of their registered patients in one place and assign levels of priority. If a patient needs to be admitted overnight, the platform can alert other hospital departments to hasten the process of reserving a bed.
All of that by itself would speed up and simplify ER processes. But Vital also utilizes the AI and natural language processing technologies that made Mint and Fountain so useful. The platform looks at the symptoms the patient reported, along with their medical history, and uses that to make predictions that will hopefully make their visit much smoother.
For example, it can determine the patient’s severity, so the doctor knows to assign certain patients higher priority.
“Higher severity patients are sent back into the rooms work more quickly, whereas somebody who was just in for a fever or vomiting can get in and out of the emergency room much faster,” explains Patzer. Lower severity patients can be “fast-tracked” and helped in the front rooms rather than transferred to the back, likely saving them hours of waiting time.
“Our research shows that maybe 20 percent of lower severity patients are missed,” Patzer says. “So it’s both better for the patients, and much better for the hospital who saves a lot of time by identifying those patients sooner.”
For more serious patients, Vital can predict which lab tests may need to be run. The technicians can do them immediately, rather than waiting the extra time for a doctor to order those tests.
“Our ultimate goal at Vital is to reduce wait times for patients by half,” Patzer says.
It’s an ambitious goal, but the company is already running pilots with several hospitals. Along with CEO Patzer, the company is being led by his brother-in-law, now business partner, Dr. Schrager.
Vital currently has eight employees. Medical research, sales, and support are in Atlanta with engineering largely in New Zealand, though they will begin to hire a tech team in Atlanta as they grow.
Patzer initially self-funded the company with about $1 million, but is now raising a funding round from institutional investors. That will be used to roughly double the team and power sales and marketing to get the system into more hospital systems.
Of course, Patzer acknowledges that in the end, AI can’t solve all the challenges in the healthcare system. His goal is to automate “the easy stuff”, but ultimately, leave the hard decisions up to the clinicians.
“This is already such a stressful day. You’ll sit there for three or four hours and not have any idea what’s going on in the background,” says Patzer. “So this app communicates how long you should expect to wait, what’s going to happen next… it provides that expectation setting to make the journey feel just a little more bearable.”