This AI Startup Helps Hospitals Navigate Operational Troubles Of Covid-19 And Beyond

^In early March, M Health Fairview, one of the largest health systems in Minnesota, got an urgent message from one of its vendors: Mountain View, California-based artificial intelligence startup Qventus. The 13-hospital system had contracted the company to help streamline its operations, but Qventus CEO Mudit Garg saw a more pressing need ahead. “I will never forget this,” says Dr. Karyn Baum, vice president of system clinical operations, recalling the email she received from Garg. “It basically said, ‘We're worried about Covid. And we really think we might be able to help you. Do you want to talk?” 

With the U.S. Covid-19 case count still low but rising fast, Garg knew his company was about to face the ultimate test. “We started doing some early computations and realized that we're going to need to react very differently than our traditional health systems are set up operationally,” says Garg. “This will require us to act and adapt really fast.” 

On Friday, March 6, the same day Minnesota health officials announced the state’s first positive coronavirus case, a team of employees from Qventus and M Health Fairview got to work building what would become the platform’s Covid-19 critical resource mission control. This gave Qventus customers real-time updates on managing Covid-19 patients, especially concerning the availability of ventilators, intensive care and other open beds, and negative airflow isolation rooms, which are critical to treating coronavirus patients. That initial team included Garg, 36, and his cofounder and CTO Ian Christopher, 31, who rolled up their sleeves to help their company move as quickly as possible. The first version of M Health Fairview’s Covid-19 mission control setup was live within 72 hours, with a more refined version operational in under 10 days.

It may seem odd that hospitals don’t already have this type of granular operations data, but it’s rare to worry about how to handle a sudden surge of patients who may require intensive care. At least, it was rare until Covid-19 hit. “It felt like planning and caring for Covid patients went from zero to 60 overnight,” says Baum.

But Qventus was prepared with technology that could help aid the frontlines. The company has built a platform to help hospitals tap into real-time patient dataflows, predict operational bottlenecks and recommend remedies. Now it’s paying off. Qventus expects revenue to hit $18 million this year, a 50% increase from 2019. And it’s not just because of the pandemic. Lost revenue from cancelled elective surgeries means many hospitals are going to be operating on tighter margins and trying to boost efficiency and cut waste. And making healthcare systems more efficient is what Qventus does best.

^Garg believes health systems, much like the patients they treat, need preventive care. So his company’s software helps clinicians adopt habits that make their organizations run more smoothly, using data and behavioral science. It does this by analyzing real-time patient data from electronic health records and overlaying this on top of the entire hospital system. It then offers nudges, similar to a push alert on a phone, suggesting actions to help clinical support staff efficiently manage care. For example, say a patient is ready to be discharged but a very busy clinician is making the rounds and hasn’t seen the papers come through. A nudge will help them prioritize the discharge and free up another bed. Qventus’ analytic and predictive capabilities not only provides nudges on the patient level, but also on a hospital or even system level, enabling administrators  to make more informed decisions about operations.

It seems like a no-brainer that hospitals should have systems in place that give them an overview of everything going on. But it’s a challenging problem, especially for bigger health systems that have grown over the years through mergers and acquisitions. Different hospitals within the same system might be using a multitude of different legacy technologies, which make it tough to share data. “The problem isn't lack of information,” says Baum. “It's a lack of synthesized information that's put together in a form that people can act on quickly.”

Qventus’ goal isn’t to replace decision making by doctors, nurses or administrators, but rather to help them cut through the noise by providing a clearer picture of patient and hospital needs with help from artificial intelligence. This is precisely the reason Ursheet Parikh, a partner at Mayfield, invested in Qventus. “When you use AI to make a person into a superhero, they become your champion, they buy your product, they adopt it into the organization,” says Parikh. “Your product doesn't have to be better than a person, it just has to make somebody better.” Another of the company’s selling points, Parikh says, is the geographic breadth of its customers. “I always feel you can sell anything in the bubbles of New York and San Francisco,” says Parikh. “We saw they had customers in Arkansas and Oklahoma.”

One thing that the current pandemic has demonstrated is that the company’s technology “enables hospitals to care for more patients with the same resources,” says Parikh. This is becoming even more critical as hospitals go from facing cyclical challenges like the flu to unpredictable threats like Covid-19. Qventus has raised $43 million to date from investors, including Mayfield, Bessemer Venture Partners, Northwest Venture Partners and New York-Presbyterian Hospital.

“Think of it as preventative maintenance for an operational process in a hospital, like things you would do for a machine in a factory,” says Qventus CEO Mudit Garg.

^The original idea for the company was born out of Garg’s experience as a McKinsey consultant, when he was first exposed to the U.S. healthcare system. Garg, who had grown up in India, was shocked by the general disorder and extremely long waits in hospital emergency departments despite their world-class clinicians, therapies and equipment. “People [were] working as hard as they could and trying to do everything they could to take care of basic things and everything was happening by chance, not systematically,” he says. 

An engineer by training and in the middle of pursuing his MBA at Stanford, Garg believed there had to be a better way of giving hospitals visibility into operational problems that were clogging up the wards and frustrating providers and patients alike. In 2011, he teamed up with fellow student Ian Christopher, who was getting a master’s in engineering, to build the first prototype. With the addition of another Stanford engineering and MBA grad, Brent Newhouse, the trio cofounded analyticsMD, which was later rebranded as Qventus. The name is a combination of “Q” the physics symbol for electric charge and “ventus” the Latin word for wind. Garg and Christopher still run the day-to-day operations of the company, while Newhouse serves on the board.

Since 2011, Qventus has grown to 80 employees, and has continued to scale its operations and increase its customer base. Its software has been deployed to 40 new locations this year. Aside from developing bespoke Covid-19 mission control for its customers, Qventus engineers also developed a free Covid-19 scenario planner based on 450 epidemiological models to help city, state and federal officials with resource allocation, ranging from ventilators to various personal protective equipment. More than 1,1000 organizations have used it so far. 

The experiences of hospitals in dealing with the first wave of the pandemic have led them to realize that they can’t afford to shut down all other services in the event of a resurgence or future disaster. “Think of it as preventative maintenance for an operational process in a hospital, like things you would do for a machine in a factory,” says Garg. Keeping hospitals running efficiently during unpredictable events has now become a key use case for Qventus, enabling its customers to both care for their patients and keep the lights on. “We're looking at it ahead of time and saying, this process, this habit is deteriorating right now,” he says. “And let's get a hold of it.”