Back in the late 1990s, a business conglomerate called BASF conveyed its value with a simple slogan that went something like this: “We don’t make X. We make X better.” That may hold a lesson for those who itch to create new health care apps: Start by examining where current practices could be improved with mobile IT devices. It doesn’t have to be costly or complicated to produce a return on investment quickly.


Geisinger Health System designed an automatic phone-based process to remind patients to show up for their appointments. But no-shows continued to be a multimillion-dollar drain on the Danville, Pa.-based network. When Geisinger shifted to a text-message reminder system, it shaved the no-show rate by a few percentage points, saving about $700,000 in the first year, and anticipated plenty of additional savings moving forward.

Vanderbilt University Medical Center implemented a methodical process for observing and fostering proper hand hygiene, and it improved the compliance rate from 50 percent in 2009 to 80 percent by 2012. As impressive as that was, hand-washing protocols still were being ignored one out of every five times. An app was built to make the work of observers easier and more data-rich, and that pushed the compliance rate to 95 percent.

The simplicity of Geisinger’s reminder technology and quick evidence of its impact made it relatively easy to scale up from 5,000 patients in an early 2013 pilot project to 55,000 when it went networkwide by January 2014, and more than 130,000 today. “I think text messaging has been overshadowed by a lot of the cool mobile apps and devices,” says Chanin Wendling, director of e-health at Geisinger. “That’s kind of sad, because it’s not that hard to do. Our results from text messaging of appointments … have been very positive. And the amount of effort we’ve expended hasn’t been dramatic, whereas, trying to navigate the mobile app world is much more complicated.”

Geisinger developed a representative figure of $176 lost per no-show, which includes opportunity costs, as well as the expense of setting up for a visit that doesn’t happen. It takes into account the time to book the appointment, prep a room and every other task that has to be done in advance. The provider also loses out on an opportunity to see a patient. A study done on patients with and without text reminders determined that if the no-show rate in the texted group were the same as the population of patients not getting a text reminder, it would amount to 4,000 additional no-shows, Wendling says. That number multiplied by $176 is $704,000, against the program’s $50,000 cost.

Vanderbilt’s progress in hand-washing compliance between 2009 and 2012 began with a systematic mobilization of firsthand observers stationed in all 36 inpatient units and 112 outpatient offices across the medical center. Every unit donated an observer, usually at management level, who was trained and assigned a nearby venue. The volume of observation increased dramatically enough to allow valid analysis of compliance. During the previous seven years, a total of 3,000 observations were made; now, they were coming in at 3,000 to 4,000 a month, says Tom Talbot, M.D., chief hospital epidemiologist. Total observations from July 2009 through August 2012: 109,988.

Every unit’s score was available to all staff. Rewards in the form of rebates on malpractice premiums created an incentive. However, the performance of the whole network had to hit a certain target to gain the rebate for everyone paying into the self-insured plan, so underperformance in some units created peer pressure to improve. Hand hygiene was becoming part of an overall culture of safety.

Nevertheless, the routine bogged down in certain areas. Transferring written observations to a Web interface after a shift was a big-time commitment. “Observer fatigue” could affect the observing or reporting. There were specific details to document: morning, afternoon or evening shift; time and place of observation; room entry or departure; type of provider; compliant or not. Any missing information undermined the process. The system had to keep track of the trackers, rotating where they should be and getting the schedule out.

An app loaded onto any type of mobile device told observers where they were assigned. When someone reported to the same unit multiple times, the app auto-populated such details as day of the week or time of day. Color-coded job descriptions captured in one click whether the subject was a doctor, nurse or student. “The design really did streamline the efficiency for the observer,” Talbot says.

Case Study

Such precautions as wearing gloves guard against contracting infections — but only for the person donning them. Gloves can be as covered in germs as bare hands, so hand-washing rules still apply. But the message wasn’t always getting across at Vanderbilt University Medical Center.

The gloved compliance rate had not been tracked, but “we realized that we could very easily capture that in the app,” says Talbot. Early on, the compliance rate for washing gloved hands was shown to be about 10 percent lower than for bare hands. So the medical center added this message on patient-isolation carts outside each room: “If you’re wearing gloves, that does not exempt you from washing hands. Wash before and after putting on gloves.” Finally, the lesson got through, Talbot says, and “our data helped to drive that.”

Program managers had to make sure they didn’t exploit too much of a good thing. “We had to keep it simple,” he says. “We couldn’t ask 15 [additional] questions; but if we could ask two questions — or three questions in this case — that would be very helpful.”

Case Study

In an age of ubiquitous cellphones, the blinking light of a received text message can be easier on a patient and better for appointment compliance than a phone call. But both? That might be pushing a patient’s patience, says Wendling.

Geisinger is working through a way for patients to opt for either one. Until then, “the text message program exists on top of every other communication the patient already gets,” she says. “That’s not ideal. We would prefer to have the patient be able to express preferences, so they can say, ‘I like text messages, I don’t like IVR [interactive voice response] calls.’ That will be coming, but it’s been a bit of a process internally to let people work through letting go.”

The IVR program is the long-established way to get patients to their appointments. In some cases, patients also get a letter in the mail. “You have to balance that business need against [the prospect that] patients are going to tune you out, because you’ve now told them the same thing four times and they can’t stand you anymore,” Wendling says.

It is now getting to the point that all those responsible for reminding patients are satisfied that the text messages are working as effectively as other methods and that it’s all right to allow a patient preference of one or the other, she says.