Financial analysts push 'Monte Carlo' analytics to solve healthcare issues

Monte Carlo, an old gambling concept, could have a future in healthcare as an approach to data analytics that allows managers to crunch vast amounts of data in order to support fiscally prudent decisionmaking.

The approach, created by Josh Lefcowitz and David Friend, consulting managing directors at financial advisory firm BDO, is starting to gain traction due to the increased processing power of modern computers.

To help explain the idea, Friend said flipping a coin a thousand times would be considered a simplistic Monte Carlo simulation."It's an old statistical tool. That's the origin of this concept," he said.

But where the Monte Carlo simulation really gets its power is in its ability to efficiently deal with complex, variable scenarios. Most outcomes in the world are not binary, said Lefcowitz; there is a multitude of possibilities and any number of variables that can tip the scales to one outcome or another. The Monte Carlo approach is to model multiple variables and run a simulation.

[Also: Hospitals hungry for predictive analytics despite current low utilization]

"Without a computer, it's not possible," said Lefcowitz. "The cost benefit isn't there. Go back 15, 20 years and you're pushing everything into a mainframe, and if you miss one character, it's a bust."

"Now you can do a million times a second," said Friend. "Let's say there are a zillion rules and variables. We literally want to test this hundreds or thousands or millions of times. You can see the outcome -- how the roulette wheel answered your question."

What the approach can achieve depends on the issues a healthcare business is dealing with. An insurer, for example, may want to know how much they should charge, or to know what the cost of an entire population might be. An organization could potentially estimate how many women in a given quarter will become pregnant, or get pneumonia, in order to figure out the best pricing.

[Also: Hospitals hope 'Moneyball' analytics approach can help bend the cost curve]

On the valuation side, BDO has been dealing with clients doing simulations from a case study and market orientation perspective.

"We have a client who has basically invented a new piece of hospital equipment," said Lefcowitz. "There is an incumbent piece of equipment that is already serving a need. So how quickly can this penetrate the market? We modeled out a simulation and went through what their potential market share could be over the next 5, 7, 10 years, depending on a variety of factors.

"Not everything in life falls into a normal bell curve," he said. "Sometimes things are more logarithmic or exponential in nature. We can incorporate all these variables into the simulation. The modeling can be very sophisticated."

Statistical analysis, rather than a physician's individual knowledge, can be used to design treatments for patients, they said. While in many cases there may not be a perfect course of action, the best course can be modeled in a simulation -- although the privacy of data could be a stumbling block in this area.

Still, there are other applications that can be pursued.

Like Healthcare Finance on Facebook

"From a fraud, waste and abuse perspective, the Office of the Inspector General in the Department of Justice is already taking these huge amounts of data and, not doing Monte Carlo simulations, but big data analytics to see where the fraud, waste and abuse is happening," said Lefcowitz. "It may not be used for treatment, or in a personalized context, but it's still an important issue. You don't need to know who the individual is. You need the data."

Friend sees the approach expanding in future years, particularly with the glut of numbers and data that have been swamping the healthcare sector.

"As we rank more programs, as we start to quantify the effectiveness of programs, drugs and skilled nursing facilities, we are invariably going to see an ocean of this," he said.

Twitter: @JELagasse