Can You Hear Me Now?

An Evolving Landscape

The mHealth movement has taken hold of public health almost as fast as the exponential rise of cell phones themselves. As of early 2010, the number of cell phones in use worldwide had hit more than 4.6 billion, according to the International Telecommunication Union, a UN agency. (To add some context, the world’s population hit 7 billion in late 2011.) A recent search of PubMed, the NIH biomedical research database, yielded hundreds of articles focused on the use of cell phones to improve health or gather health information, most added in the last three years.

Labrique recalls seeing the change himself over the past decade in rural Bangladesh. When he started work 11 years ago on the JiVitA project, a study designed to understand the effects of supplementing pregnant women’s diets with vitamin A, Labrique remembers the abysmal communication among members of the research team scattered throughout the rural countryside dotted with green rice paddies. The people were quick to offer a place to sit and a betel nut to chew, but were stunningly isolated.

“We couldn’t make a phone call to the next town,” Labrique says. The only reliable way to pass information among team members was to pay messengers to carry written memos by bus, so getting a simple answer to a question could be an all-day affair. “I joke when I lecture about this that we were seriously contemplating carrier pigeons,” he adds.

By 2004, the first cell phones started making their way through the area. With just a single tower nearby, it still wasn’t a useful way for Labrique and his colleagues to connect—it worked better as a landmark. His research team counseled visitors driving to their site to travel north until they saw that cell phone tower, then take a left to reach the field site.

But in a few short years, the landscape changed. As 30 new cell phone towers popped up around the JiVitA site, more and more of his local colleagues began using cell phones themselves—not just senior managers in the study, who could easily afford what started out as a luxury item, but, eventually, grassroots-level field workers as well.

“In the span of two years, these field workers—who usually have no more than an eighth-grade education—went from having no phones whatsoever, to almost every single one carrying a personal phone,” Labrique says.

Eventually, he and his colleagues started noticing that cell phones were infiltrating the narratives they were collecting from women and their families to describe obstetric crises and maternal or infant deaths. When they crunched the numbers, they found that about half the women in their study who’d experienced an obstetrical crisis had used a mobile phone to try to turn their situation around—by calling a provider, arranging transportation to a clinic, getting financial aid to pay providers or seeking out medical advice.

With access to cell phones skyrocketing in the area, either through direct ownership or access to a village phone, Labrique and his colleagues decided to start up a mobile phone–based labor and birth notification system. In a recent study, led by International Health Professor Parul Christian, when pregnant women went into labor, they or their families called or sent text messages to a central number. This action dispatched nurse-midwife teams to the women’s homes, where 90 percent of births take place in rural Bangladesh. Results showed that about 89 percent of these births—which would normally have taken place without any medical care—were attended by highly skilled health care workers with the new system.

Empowered by this success, Labrique’s team will launch a new project this year called mCARE that takes these previous studies to a whole new level. Working closely with the Bangladeshi Ministry of Health and Family Welfare, complementing the government’s vision of a “Digital Bangladesh,” supported by the UBS Optimus Foundation, the researchers will be supplying cell phones to the community health workers who visit women periodically to get those who are pregnant into prenatal care as soon as possible. The phone is an in-expensive Chinese-made Android model—an operating system well suited to mHealth applications because its open-source nature makes it highly customizable to users’ needs.

On their regular pregnancy surveillance visits, these workers can use these phones to register their clients, possibly even snapping a quick picture so supervisors can verify who they’re talking with in subsequent visits. Guided by a customized app on the phone, the workers will then ask a series of questions incorporating lunar calendars and local events, to sort out when the woman’s last menstrual period took place. If it was more than five weeks ago, the app notifies the worker that this client is potentially pregnant.

That pivotal revelation will automatically trigger a series of other events. Based on the woman’s expected due date, the app schedules several prenatal appointments. It will send her reminders on her own phone, if she owns one, and to the community health worker, who will stop by a couple of days before appointments to emphasize the importance of each visit to the woman and her family. As with the previous study, each woman and her family will be encouraged to notify the study by text when they go into labor and if they need help, spurring a mobile health care team into action to attend the birth or facilitate a referral to clinical care. If labor appears premature according to the system’s records, it then signals a special alert to the health care team that they may be dealing with a preterm baby that may have more intense medical needs. Another text when the baby is born will trigger another series of visits one, three and five days later, to make sure that mother and baby are doing well.

“Each action here stimulates a reaction,” Labrique explains. “Rather than waiting for a crisis to happen, we’re using mobile technology to respond to potential problems before they occur.”

The study’s impact on mortality is yet to be measured, but based on the pilot work with labor and birth notification systems and emergency dispatches of nurse-midwife teams, Labrique expects these efforts will pay off through better prenatal care for mothers, more attended births and targeted care for infants (especially high-risk, preterm babies)—ultimately saving the lives of mothers and their infants.

“The groundwork has been done to demonstrate that these systems can work in this challenging, resource-limited, remote context,” he says.

A Game Changer?

As Labrique and his astute colleagues noticed, cell phones are an ideal solution to connecting with low-resource populations. But mHealth isn’t just for the developing world, according to Betty Jordan, an assistant professor in the Johns Hopkins School of Nursing.

In 2009, when Jordan was serving on the board of directors at the National Healthy Mothers, Healthy Babies Coalition, she heard of a project that would send text messages with health advice to pregnant women three times a week. Then, once they gave birth, it would switch to health advice for newborns, all based on the due date that enrollees provide when they sign up for the service.

While many low-income women may not have computers or pregnancy books, the coalition reasoned, many of them do have phones—providing a way to deliver information that could have enormous impact on their health knowledge and behavior and the health of their babies.

“I thought it was a fabulous idea,” Jordan recalls. “A 16-year-old inner city pregnant teen may not be going to the library to read a pregnancy website or be able to afford childbirth classes, but she might be willing to read the message that comes across her phone."

In early 2010, text4baby launched across the country. Since then, more than 260,000 have enrolled. But she and her colleagues wanted to make sure that text4baby was a success by other standards as well, so they built in measures to evaluate the program from the start.

According to Piers Bocock, project director for the Knowledge for Health Project, run by the Bloomberg School’s Center for Communication Programs (CCP), mHealth evaluation remains a huge hurdle. Governments and donors want to make sure that mHealth interventions can be measured so they can make the right decisions about funding comprehensive mHealth programs. “There are a lot of pilots out there,” says Bocock, “but not a lot at scale.”

CCP, which includes mHealth components in more than a dozen of its projects around the world, is constantly working to understand how mobile efforts are adding to the effectiveness of its programs.

“We all realize mHealth can be a game-changer, especially when it is part of other social and behavior change communication activities,” says Bocock. “The question we want to answer is how to quantify its effectiveness within the context of broader public health interventions.”

Garrett Mehl, PhD ’00, MHS ’94, a WHO scientist and a chair of its Health Data Forum Working Group on mHealth, notes that insufficient attention to the role of research has been the downfall of countless other mHealth projects.

“I think we can definitely say that there have been a considerable number of pilot mHealth projects, and a lot of them have failed in either their ability to demonstrate some health impact or in their ability to find a mechanism to sustain them,” he says. Mehl adds that in a joint project with a Bloomberg School intern that assessed the global state of evidence generation among mHealth projects, a considerable proportion of the projects were struggling with research—and donor support—needed to validate their efforts. Despite their presence in public databases, many mHealth projects were found to have already ended, suggesting their inability to transition from pilots to scaled-up programs. Often projects, Mehl notes, are driven by the pleas of donors to get implementations running quickly, without any forethought about how to judge success or pressure to plan for scale-up and sustainability.

“You begin to worry that a lot of investments are being made in this area, and if they fail, you worry that people won’t want to continue to invest,” he says. Fortunately, Mehl notes that donors are now beginning to pay more attention to evaluation and invest in mHealth evidence generation and synthesis.

To head off these problems, Jordan and her colleagues incorporated some unique evaluation methods into the fabric of text4baby. For example, to see whether the program is reaching its intended audience, researchers ask participants for their ZIP codes during registration. The result is a real-time map across the country that text4baby’s partners, including local health offices, can access and watch enrollment numbers change on a minute-by-minute basis. They can also instantly see whether ads to entice women to sign up have the desired effect. An ad for text4baby during the popular MTV program 16 and Pregnant, caused a huge spike in enrollment.

“It’s a huge strength of the program to see whether we’re hitting our intended audience,” Jordan says.

But demonstrating whether these texts are improving outcomes for mothers and babies is a much tougher problem to tackle, Jordan notes. “It’s easy to tell whether women are enrolling, find out whether they like the messages or see if the number of texts they get each week is acceptable,” she says. “It takes a lot more time, effort and evaluation strategies to demonstrate knowledge and behavior change.”

One step toward judging whether they’re achieving this goal, Jordan adds, is a series of interactive modules that the text4baby team recently began inserting into the typical texts that users receive. Around the end of October 2011, they sent their first interactive module: a questionnaire on whether users had received the flu shot, and if not, why. Within 48 hours, nearly a third of the 96,000 users who received the module responded, giving Jordan and other researchers involved with the project reassurance that users were engaged and interested in sharing information, as well as lending insight into their health behaviors.

Larry Cheskin, MD, an associate professor of Health, Behavior and Society, and director of the Johns Hopkins Weight Management Center, is hoping to get around the evaluation problem by incorporating mHealth into a randomized study—the gold standard for other health interventions.

He explains that the typical program at the Weight Management Center is a relatively time- and resource-intense affair. On their first visit, patients see a series of health care providers—a dietitian, psychologist, exercise expert and Cheskin himself—and come back frequently for follow-up. This care usually isn’t covered by insurance. Since those of low socioeconomic status are more likely to be obese in the U.S., it places the program out of reach for those who probably need it the most.

“It’s not translatable to the U.S. as a whole,” Cheskin notes.

Seeking a better way, he and his colleagues launched the TRIMM study—short for Tailored Rapid Interactive Mobile Messaging—in 2011. They’re recruiting 150 minority participants from inner city Baltimore who are interested in losing weight. All the participants will receive comprehensive counseling on diet and exercise, but half will receive customized text messages several times a day that address their self-identified problem areas. Cheskin and his colleagues plan to see how the two groups compare after six months—and then after another six months, when the text messages are shut off.

“It’s well known in this new field of mHealth that there’s not a lot of control data,” Cheskin says. “Doing a randomized controlled trial is a high quality way of seeing whether the outcome you’re hoping for is really there.”

Evaluation isn’t the only tough problem in mHealth—scalability and sustainability are issues that have doomed many other mHealth projects, notes Patricia Mechael, PhD ’98. She recently became the executive director of the mHealth Alliance, a Washington, D.C.–based organization hosted by the United Nations Foundation that serves as a convener of the mHealth community and provides guidance and support for those using mHealth tools. For example, giving out phones to researchers and subjects alike might be the kiss of death for many mHealth projects, according to Mechael. For a small pilot project, maintaining equipment and airtime might be manageable, but continuing to provide equipment and airtime for a full-scale project is oftentimes financially unsustainable. Unless a country’s government or private sector investor can invest in buying a phone and minutes for the target population, Mechael explains, that model simply won’t work for the long haul.

Similarly, Mechael says, multiple projects have failed because there is no standard for them to integrate with one another. For example, she explains, there’s a missed opportunity if one mHealth intervention evaluates patients for tuberculosis symptoms while another assesses HIV risk, but the two aren’t designed to easily combine their findings. Governments that are seeking a complete picture of these two diseases in the populations they serve will likely discard both programs.

From the outset, Mechael says, programs should examine how mobile technology can be leveraged to strengthen the health system as a whole and interact with other platforms, even if the initial funding is specifically targeting a particular health condition.

“mHealth is a lot more complicated than just giving out phones or developing apps,” Mechael explains. “Technology is only as good as the systems that it supports.”