Mohan Giridharadas: Lessons Healthcare Can Take from the Airline Industry

Ryan Black
JANUARY 11, 2018

Yesterday, Healthcare Analytics News™ published the first part of our conversation with Mohan Giridharadas, the CEO of growing analytics firm LeanTaaS. He outlined several major inefficiencies that plague health systems. In the second half of the interview, he details how similar shortcomings are reflected in other industries, and why healthcare would be well-served to learn from them. 

We were talking about inefficiencies and wasted resources. How do other industries deal with this issue?

The only time an airline makes money is when the passengers are in the seats and the plane is in the air.  A plane lands at an airport, pulls up to the gate, connects to the jet bridge, and then what happens? They kick out 200 passengers and the crew, clean the cabin, restock the food, and put in 200 new passengers and a new crew. Meanwhile, they refuel, check the engines and the tires and the flaps, unload 10 households’ worth of furniture, and reload 10 households’ worth of furniture. No service interferes with another.

This all happens in 45 minutes. It happens 1000 times per day in hundreds of airports around the country. And it’s done perfectly: With even 99.9% accuracy, we’d have a dozen plane crashes per day in the US.

Each does their own job under a super tight choreography. Think of the capacity it has unlocked. Back in the 1980s, when I went to graduate school at Georgia Tech, Atlanta’s airport could do a few hundred flights per day. Now it does thousands. The airspace around Atlanta hasn’t changed between then and now, and they’ve only added 1 runway.

They were able to increase the velocity by 10 times. That was done by optimizing each service and synchronizing them. If hospitals were to do that, think of the efficiency we could unlock.

So how does this metaphor translate to healthcare?

Operating rooms (ORs) are massive assets for health systems, representing 40% of revenues and 60% of profits. They also contribute enormously to inpatient beds, because the next step following complex surgery is usually a couple of days in the hospital.

An OR is also an incredibly expensive asset. You’re talking about rooms that cost $10 million or $20 million. Every minute in an OR is like a 747 on the ground. A wasted minute in the OR is a $100 or $300 event. A day of wasted OR time is a $50,000 to $100,000 event.

The transaction costs for swapping block time are horrendous. If a surgeon knows they have the Monday block, but 6 Mondays from now they’ll be at a conference, the process by which that time gets used is horrible.

We put out a template for chemotherapy, suggesting the pattern of appointments for Mondays. We watch how the next 5 or 6 Mondays unfold, and our algorithms pick up on patterns and suggest we unlock more slots in that time. Our algorithms learn and adapt.

Beyond the OR, clinics are an opportunity. We’re going to start in oncology clinics because it’s a rapidly growing area. The clinic appointment is linked to the use of other assets in oncology, like infusion and radiation oncology, so optimizing its flow would help a lot.

How do you change a system’s ingrained operations?

What healthcare does is complicated. They perform magic every day. The tendency to just do things as they’ve always done them is a powerful force that’s hard to argue with.

We draw a very bright line. We don’t opine on clinical matters; we opine on operational excellence. A lot of data science tries to form pattern recognition and prescribe a drug or tell the clinician what to do. The moment you cross that line, you’re trying to tell doctors how to do their job.

How are you able to tell MD Anderson, “I’ve got better oncologists on my team than you have on yours?” That’s a difficult game to play.

What other barriers prevent scheduling reform?

In health systems, it often leads to analysis paralysis. They say, “We must fix the patient experience from labs to pharmacies to clinics.” And 5 years later, nothing has been fixed.

In an interconnected system, people believe that you can’t optimize 1 node without them being involved. The oncologist has a hard time believing, in some instances, that you can fix infusion without asking them to behave differently. But we can.

We are trying to solve the service delivery problem in a hospital without getting caught up in the drama of the hundreds of individuals and doctors. In a complex, interconnected set of services, you have to optimize the nodes before the edges. FedEx will make the warehouse automated and productive before it worries about faster drivers. Amazon’s warehouses will be mostly human-free before they worry about the drones flying faster.

Become a contributor