By Aaron Krol 

July 3, 2014 | Once every month, a few dozen teenagers in the Boston area all receive the same Facebook notification. The message is fed to Facebook by a local healthcare provider, Massachusetts General Hospital (MGH), but it’s not an ad or a public service announcement. It’s a link to a five-item test that asks questions like: “During the past four weeks, how often did your asthma symptoms wake you up at night?”

Teenage asthma is a persistent problem for MGH, with sufferers repeatedly ending up in the emergency room for life-threatening events associated with this manageable condition. These events are preventable, but not without the involvement of the patients. Despite the serious risks of poorly controlled asthma, teenagers in the MGH pediatric asthma clinic keep their follow-up appointments less than a fifth of the time.

“On average, they land back in the ER about four times a year, because of exacerbated asthma,” says Kamal Jethwani, the principal investigator for the study MGH is conducting through Facebook. “Ninety percent of the time the cause is that they —haven’t been able to take their medication correctly, or they forgot to refill it… Adherence to appointments became such a big issue that we had to figure out a way to make this easier for teenagers, and go where they are.”

For Jethwani and his colleagues, going where the patients are has become a guiding principle. Jethwani is the Corporate Manager for Research and Innovation at the Partners HealthCare Center for Connected Health, which experiments with technology-driven programs for tackling stubborn health problems on behalf of the Partners HealthCare network in Massachusetts.

The philosophy at the Center for Connected Health is that if a program can engage patients in their own care, positive clinical outcomes will inevitably follow. Research at the Center does look at the standard metrics hospitals worry about — in the case of the Facebook study, it’s tracking hospital admissions — but it also wants to know whether patients enjoy using the programs, and their rate of participation. “We always measure engagement, and if we have good engagement, we consider that successful,” says Jethwani. “It improves clinical outcomes, it improves conversations and communication with physicians, and engagement also improves… your own ability to take care of yourself over time. We’ve shown that time and again.”

That theory has been borne out in the group of teenagers with asthma. One year into the program, four-fifths of these patients were still taking their monthly asthma control tests online, allowing their doctors to see immediately if their scores were falling below a dangerous threshold. That’s led to a thirty percent reduction in hospital readmissions among the Facebook group.

“It’s something really simple, just making them answer a survey every month,” says Jethwani. “But doing it on Facebook, because they’re already there, they’re already engaging with it every day and they see it as part of their daily newsfeed, made it so much easier to get follow-up from these kids.”

Marketing Health 

This April, the Director of the Center for Connected Health, Joseph Kvedar, appeared at the Medical Informatics World Conference to explain the Center’s vision for boosting patients’ personal involvement in their healthcare. With new technologies like mobile phones and social networking, he told attendees, “we can make health addictive.”

Kvedar wants healthcare providers to learn a few tricks from the advertising industry, an idea that can be a hard sell. “Healthcare people think of that as soft science and manipulative,” he told Clinical Informatics News, but in many regards, the ad industry respects how people live their lives in a way that the health system rarely has. “We think there’s a lot of good evidence that the way we approach illness and prevention is quite broken,” he continues. “We’re constantly telling you the bad things you’re trying to avoid, reminding you that you’re overweight, or your blood pressure’s high, and for God’s sake get that under control. And then if you don’t, we call you non-compliant.”

In contrast, ad agencies try to seamlessly work their messages into people’s daily routines; and if the message doesn’t take, the responsibility lies with the messenger, not the recipient. That’s a much more constructive attitude to take toward the recurring problems in American healthcare, almost all of which can only be addressed with sustained changes in behavior. “There are six or seven use cases that keep coming back,” says Jethwani, including diabetes, heart failure, asthma, cancer care, and smoking. None of them can be effectively tackled during short, infrequent visits to the clinic.

“If you think about things like chronic illness, coming in for an office visit twice a year to get your blood pressure checked doesn’t make sense at all,” adds Khinlei Myint-U, the Center’s Corporate Manager for Product Development and Communications. Before coming to Partners, Myint-U worked in online retail and online banking, a background that helps her put programs in the context of patients’ needs.

“The consumers for banking and for other media are the same people in healthcare,” she says. “Unfortunately, sometimes I think in healthcare we’ve looked at patients differently, but really they’re the same people. They want to do things that they’ll find of value, they want to do things that are going to be easy to use… It’s about creating products and programs that feel relevant and personal to a person.”

Moving Information 

The Center didn’t always have the technology to live up to its ambitions. When Kvedar founded the Center in the mid-1990’s, it was called Partners Telemedicine, and its methods of bringing healthcare into patients’ homes could be byzantine and unwieldy.

“It was a different world,” says Kvedar. “We didn’t have Netscape Navigator yet, so no form of web browser was around or at least popular. I think people hadn’t envisioned the Internet as a place to actually do transactions and work yet, so we thought about moving information over private networks like ISDM lines.”

The first programs focused on heart disease, a good model for conditions that need to be addressed at home and on a daily basis. “We had a Windows CE clamshell device, and we hooked some sensors up to it like a weight scale and a blood pressure cuff, and created a little home monitoring ensemble,” Kvedar remembers. “We were able to show that patients would interact with it, could upload their vital signs, and we could collect their vital signs.” It was an interesting proof of concept, but far from a scalable program for delivering system-wide improvements in cardiac health.

Ultimately, the big changes to the Center’s capabilities were driven by even bigger changes in the country at large. The skyrocketing market for cell phones, smart devices, and web conferencing has meant that connected health programs no longer need to jury-rig hardware to get a new platform off the ground. “Now, the technology exists and people already own it,” says Myint-U. “And that’s huge, because we’re not talking about deploying technology.” In a newer heart failure monitoring program, she recalls sending care workers to elderly patients’ homes to distribute tablets. “What we found when we went out there was, these grandmas were showing us their iPads, and saying why can’t I use this instead?”

Texting was the first mobile technology to take off at the Center, and it’s now become a staple of behavior-changing programs, for everything from helping patients quit smoking to reminding them to use sunscreen. Kvedar calls it the lingua franca of eHealth, a medium virtually everyone today can engage with. Often, a program doesn’t need any more complex interaction than text messaging to make a difference.

A good ongoing example is a text-based campaign to send information and support to pregnant teens, which began at Lynn Community Health Center in a small city south of Salem. “They said, our teens don’t want to read pamphlets, they don’t answer their phones, they don’t often come in for their prenatal visits, and we want to be able to reach them,” says Myint-U. “We created a very simple campaign that was personalized to each patient, based on their due date and delivery date and who they were getting care from, and there was a series of messages that would go out to them proactively.”

Messages, which go out one to three times a week, let the participants know what to expect at each stage of their pregnancy, remind them to stay on top of important preparations like practice runs to the hospital, and offer them chances to connect to their obstetrics teams. Different clinics can also add twists to address local priorities — say, post-partum depression, or breastfeeding.

“The great thing about texting is that you’re forced to create these bite-sized pieces of information,” says Myint-U, which is an important motivator for physicians who are often tempted to over-explain or use inaccessible language. A message on a cell phone that can be read and understood in seconds can get the essentials across where a detailed pamphlet in a doctor’s office will almost always go unread.

The Personal Touch 

The other advantage of a text campaign is that it can be shaped to fit each individual patient’s circumstances. For the pregnancy program, the Center can send messages based on the location of a patient’s care center, and how they’ll need to plan a route to get there for delivery. It can also respond to users’ interest levels — a patient who rarely reaches out the clinic may start ignoring texts if they come too frequently, whereas on who engages frequently might like to hear more about the support services available.

In recent studies, the Center has been learning to take this flexibility to a new level. “The advertising industry has done a lot of work being able to segment people, and understand how to target products to certain groups,” says Jethwani. “We’re trying to really get into that psyche, and think of patients as consumers, and appeal to them in the way the advertising industry has been able to do.”

In a fitness program for patients with diabetes, the Center is experimenting with ways that the detailed patient information tracked through Partners HealthCare’s electronic medical records can enhance mobile messaging. This program, called Diabetes Connect, starts with a questionnaire for patients, which tries to gauge their baseline level of engagement in managing their diabetes through physical activity. At the lowest level, a patient is considered in “pre-contemplation,” not even aware that being active could help with their condition. At the highest level, a patient is in “maintenance,” making changes to their lifestyle and sustaining them over time.

It may seem obvious that patients at opposite ends of this scale will need to be engaged in different ways, but in practice these questions are too rarely asked. Through the Diabetes Connect pilot, which monitors patients’ activity levels over time using freely-distributed Fitbits, the Center for Connected Health can speak to each patient’s current needs. For a pre-contemplation patient, says Kvedar, “you’re going to get a lot of educational messages about ways that you can add to your step count without rearranging your life. It might mean parking your car farther away from the restaurant, or taking the stairs in a building.” Meanwhile, for a patient in maintenance, “your messaging will be more congratulatory – keep up the good work – but also ways you can extend your enthusiasm: challenge yourself, set a new goal.”

Diabetes Connect also layers in medical data from Partners’ records, which is Jethwani’s area of expertise. “Our EMR data allows us to be a little more clinically accurate, in order to understand or predict how quickly these patients will move,” he says. “For example, recently diagnosed diabetics usually take much longer to make changes than people who have internalized their diagnosis over a number of years. Insulin makes it very real for patients, because they know they have to take an injection every day… We see a very big difference.”

Overall, there are twenty-five dynamic categories for patients in Diabetes Connect, each with its unique system of messages. The live Fitbit data on activity levels lets the Center shift patients up the scale in real time, tweaking their messaging to meet them at each new level. The program even responds to the weather, sending out different messages on rainy days to help participants get in their daily activity indoors.

While the study is still in its pilot phase, and data on clinical outcomes are not yet available, Diabetes Connect has been hugely successful on the metric that matters most to the Center: 80% of patients have stayed engaged with the program through its duration.

Understanding Incentives 

In many ways, the Center’s work is a study of the incentives that move people to change their behavior, which are not always as simple as the vague knowledge that we need to live healthier lives. “The transition from the hospital to the home is an area we’re really interested in,” says Myint-U. “It’s a kind of trigger point, where people are really interested in their health, and open to something that’s new,” and that gives the Center a higher success rate with appropriate interventions. People with strong social networks, and lots of social interactions, are also more likely to stick with new behaviors, which suggests new avenues for the Center to pursue in social media or online support forums.

There are also strange disincentives at work in healthcare. The clinicians themselves can be an obstacle to the uptake of new programs, if tools aren’t designed with the doctors’ ease of use in mind as well as the patients’. “Doctors don’t want to use a separate website,” says Myint-U. “They don’t want to start any more log-ins. They just won’t use it. We did a study early on that showed the engagement of the doctor was directly correlated with the engagement of the patient.” That has guided the design principles behind new platforms ever since.

The social barriers between patients and their physicians can also stand in the way of delivering the best care. The Center has gotten involved in cancer pain precisely because of this sometimes fraught relationship, where patients don’t always feel able to be totally honest with their doctors.

“Eighty percent of cancer pain is either underdiagnosed or undiagnosed,” explains Jethwani. “Patients have these feelings about cancer pain, that they need to feel the pain in order to make the cancer go away, or if they tell the doctor about their pain they’ll stop the chemotherapy.” To take away the temptation for patients to underplay their pain for their doctors, the Center introduced a pain-tracking app that prompts regular pain reports every few days.

Using an app not only leads to better reporting of cancer pain; it’s also more responsive to how pain can change over time. The app can tweak medications, ask patients questions about their side effects and medication adherence, and even be used to order prescription refills in a fraction of the time it takes the hospital system.

That variety of uses shows how extensible tools developed at the Center for Connected Health can be, especially when they’re built for smart devices. “We have all this rich information now being driven from patients into the app that we’re making available to our clinicians,” says Jethwani. “So I can easily see this being deployed in oncology centers, and training physicians, or creating decision support, to improve how we manage cancer pain overall.” That rich dataset could also be used in the future to target this kind of program to the patients who most stand to benefit. At present, oncologists recommend their patients for the cancer pain app, but there’s no reason that Partners couldn’t use data from medical records to predict which patients would see the most value from the program.

Not every program started at the Center is a success. They nearly all begin as small pilot studies, tracking one or two hundred patients over six months or a year, and often they’re discontinued after that point. But the most effective tools become implemented system-wide at Partners, or get spun out into new companies or licensing deals. And each study brings the Center closer to understanding the incentives that move patients to take better care of their health, and what kinds of outreach they’re most likely to interact with.

The rapid uptake of mobile technology is turning the platforms created at the Center into models for care that could soon be broadly adopted. Kvedar hopes that the ease with which these campaigns can reach patients at home, and the evidence his Center is collecting that patients are interested and willing to participate, will help shift perceptions of how doctors can connect to their patients.

“As healthcare providers, we manage everything the same way,” he says. “You come see me in the office, and I take care of you. And if I don’t see you, I don’t think of you, and you come more often when you’re sick than when you’re not. That just doesn’t work well for chronic illness management and prevention.” For the biggest challenges facing the American healthcare system, care centers will have to learn to reach outside their walls.