Predicting No-Shows to Put an End to Waiting at the Doctor's Office

What if you could accurately predict which patients show up for doctor's appointments and which ones don't based on past behavior? You could put an end to long waits at the doctor's office.
Image may contain Footwear Shoe Clothing Apparel Human Person Furniture and Chair
Tobyotter/Flickr

When you're stuck waiting an hour past your appointment time at the doctor's office you can thank those patients who decided to never show up at all. That's because when too many people are no-shows one day, doctor's offices often overbook another day. But just as with overbooked flights, that strategy often backfires, causing too many people to show up on the overbooked day. In the scheduling business it's called "naive booking." While a fitting term, it's little comfort as you cool your heels with hopelessly outdated magazines while your now overwhelmed doctor tries to get to everyone.

Enter stealth startup Smart Scheduling. At MIT's H@cking Medicine conference last winter, Smart Scheduling CEO Christopher Moses met his future co-founders Andrea Ippolito, Donald Misquitta and Gabriel Belfort. Belfort, a M.D. and Ph.D. himself, told Moses about the scheduling woes his wife faced at her pediatrics practice. It was the whole no-shows leading to overbookings, leading to long waits scenario in spades.

But what if you could predict which patients would show up and which ones wouldn't, based on their past behavior? Naive booking could become downright smart booking, Moses (right), Ippolito, Misquitta and Belfort reasoned.

To test out their idea, the four joined the Healthbox Accelerator in August 2012 to try and develop on an algorithm that could make those predictions. While there, the foursome was introduced to Athenahealth, the public-traded medical cloud services company founded by former President George W. Bush's cousin Jonathan Bush.

It was through Athenahealth that Smart Scheduling got its hands on the data it needed to test and tune its algorithm. Athenahealth gave Smart Scheduling three years' worth of patient scheduling data from 12 clinics to get started. Of the more than 700,000 appointments they looked through, there were 30,000 no-shows. "After looking at the data, we've been building prediction models to show who's going to miss their appointment and why they're going to miss it, (either because they're likely to cancel last-minute or tend to fail to show up without warning)," says Moses.

Smart Scheduling's first web application running the algorithm is being used in a pilot program with Boston-based medical group Steward Health, which uses Athenahealth software to run its medical practices. The web app shows an office's existing Athenahealth schedule and Smart Scheduling's suggestions for when they should book patients and when they shouldn't. When the office has to schedule a historically flaky patient, Smart Scheduling will suggest the best time slot so that if they don't show up, it won't burden the office. The thinking is that if Smart Scheduling's tools can warn the person making the schedule of potential no-shows and the ideal overbook slots, it will help the medical practice make better scheduling decisions that ultimately affect how long you'll be stuck in the waiting room.

Rich Fernandez, the COO of Steward Health, jumped at the opportunity to participate in the pilot. For him, no-shows cost his medical group revenue, but also disrupt patient experience. "It's a revenue thing, but it's also about providing better quality and service," Fernandez says. "A struggle for a large medical group like ours is when patients don't show up. What we were impressed by with Smart Scheduling was their innovative way of helping us predict who's really going to show up so we can provide great access to our patients and keep our providers in our offices running as smooth as possible."

Moses says that during Smart Scheduling's private alpha in a few medical offices, the algorithm was 95 percent accurate at predicting the total volume of patients that would show up to a medical office on any particular day, accounting for no-shows and last-minute cancellations. "That allows us to tell medical offices how many patients they should be booking each day to hit their optimal number of patients," Moses says.

You can imagine applying the same Smart Scheduling approach to all kinds of waiting-room rich scenarioses, from mechanics to barber shops. And indeed, Moses has been approached by people who want to use his technology for their hair salons or spas. But Smart Scheduling is focusing on medical offices for now, Moses says, with an eye toward branching out to dental offices next. The goal, of course, is to have Smart Scheduling running in every stripe of physician's office, Moses says. For the sake of our dental, physical, and mental health, here's hoping they get there - and fast.