Kenya's Startup Boom

Erick Njenga, a 21-year-old college senior wrapping up his business IT degree at Nairobi’s Strathmore University, has a gap-toothed grin and a scraggly goatee. A mild-mannered son of auditors, he didn’t say much as we tucked into a lunch of grilled steak, rice, and fruit juice at an outdoor café amid the din of the city’s awful traffic. But his code had done the talking. Last year Njenga and three classmates developed a program that will let thousands of Kenyan health workers use mobile phones to report and track the spread of diseases in real time—and they’d done it for a tiny fraction of what the government had been on the verge of paying for such an application. Their success—and that of others in the nation’s fast-growing startup scene—demonstrates the emergence of a tech-savvy generation able to address Kenya’s public-health problems in ways that donors, nongovernmental organizations, and multinational companies alone cannot.

Njenga was humble about the project, but the problem he had tackled was critical in a nation where one in 25 is HIV-positive (10 times the U.S. rate) and AIDS, tuberculosis, and malaria are among the leading killers. In 2010, the Kenyan government realized it had to do something about its chaotic system for tracking infectious diseases in order to improve the response to outbreaks and report cases to the World Health Organization. Handwritten reports and text messages describing deaths and new cases of disease would stream in from more than 5,000 clinics around the nation and pivot through more than 100 district offices before being manually entered into a database in Nairobi. The health ministry wanted to let community health workers put information into the database directly from mobile phones, which are ubiquitous in Kenya. The ministry initially sought a solution the usual way: it explored hiring a multinational contractor. It drafted a contract with the Netherlands office of Bharti Airtel, the Indian telecommunications giant that also operates a mobile network in Kenya. The company proposed spending tens of thousands of dollars on mobile phones and SIM cards for the data-gathering task, and it said it would need another $300,000 to develop the data application on the phones. The total package ran to $1.9 million.

The contract was never signed; Kenya’s attorney general stopped the deal over questions about its reliance on one mobile carrier. Not very many years ago, there wouldn’t have been any options within the country. But Kenya’s director of public health made an urgent call to Gerald Macharia, the East Africa director for the Clinton Health Access Initiative (CHAI), a wing of the foundation started by former president Bill Clinton. Macharia then called an instructor at Strathmore, who quickly rounded up the four students. They spent the spring of 2011 at the CHAI offices, receiving internship pay of about $150 a month. They sat for days with the staff in the health ministry to understand the traditional way of gathering information. Then they pounded out the app and polished up the database software to allow disease reporting from any mobile Web interface. By last summer their “Integrated Disease Surveillance and Response” system was up and running at the ministry, obviating much of Bharti Airtel’s proposed costs. The process was “rough—but not too bad,” Njenga says. “There were some nights we worked until 2 a.m.” He and his colleagues are now finishing an SMS version so that health workers without Web access can make reports via text message from mobile phones of any make or model. The students are also working on another key problem: coming up with ways for the health ministry to track pharmaceuticals it ships to the government’s hospitals and clinics, to avoid shortages or waste.

Mobile phones are lifelines for Kenyans. Some 26 million of the nation’s 41 million people have phones, and 18 million use them to do their everyday banking and conduct other business; most use a service called M-Pesa, which is offered by the country’s dominant wireless provider, Safaricom. If mobile phones could play as big a role in Kenyan health care as they do in Kenyan financial transactions, the effects could be profound. A growing body of research worldwide is showing that beyond disease surveillance, mobile phones can improve public health by connecting people with doctors for the first time, reminding people to take medications or bring children in for vaccinations, and even enabling doctors in remote areas to view, update, and manage crucial clinical records.

Still, there are big gaps between the promise of mobile health technologies, or “m-health,” and their actual implementation. According to the mHealth Alliance, a Washington-based group, 45 mobile health projects are active or have already been completed in Kenya alone—more than in any other country. Most have been devised and paid for by philanthropies, aid agencies, and NGOs. The projects vary widely: one delivers money via M-Pesa to pay for repair of fistulas, a damaging complication of childbirth; another verifies the authenticity of drugs when workers text their serial numbers. Some have had substantial impact. But most are limited in scope and time frame. And there’s often no business model for sustaining them when the funding runs out, leaving the field suffering from a bad case of “pilotitis,” says Patricia Mechael, executive director of the mHealth Alliance. “The space is incredibly fragmented, unfortunately,” she says. “You have a lot of bits and pieces coming from different angles and lots of pilots going on.”

Meanwhile, IT contracts for government websites, electronic registries, and other large projects are typically conceived by NGOs or donors and carried out by contractors who may be remote from the specific needs of workers at the front lines. “You have people thinking at 30,000 feet: ‘Let’s do websites for every government ministry,’” Jackson Hungu, CHAI’s country director, told me over dinner in Nairobi. That’s good, he says, but it may not meet the needs on the ground: “Have we gone to that pharmacist and asked, ‘Look, what do you do? You are the one who meets the patient and feels the pain.’ Have we understood it thoroughly from that guy’s point of view? Or are we building something so donors can say, ‘Oh, we are online’?” Successful national technology strategies, he argues, require people like the Strathmore students, who have the code-writing chops, can readily work with the people who need to use the technology, and are likely to remain in Kenya to sustain the effort.

THE M-HEALTH GAP

Nairobi’s Prestige Plaza shopping center would look familiar to anyone from a rich country: it’s got an anchor megastore, called Nakumatt, and a food court (the Swahili Plate concession, which dishes up beef stews and curries, is a clue you’re not in Kansas). But one block away, a rutted dirt road perpendicular to the complex leads to the maw of one of Africa’s largest urban slums, Kibera, with 170,000 residents. The outskirts bustle with stalls selling kale, peanuts, sugarcane, herbs, and cell-phone SIM cards. The ground is hard-packed mud littered with stones and garbage. Single-story huts flank alleyways. Rusty corrugated-metal roofs shed the rains. At the nicer huts, curtains blow through openings. But the smells of smoke and feces linger, and children play near fetid rivulets lined with plastic refuse. The river at the lowest end of the slum becomes like a sewer when it rains. Kibera is, unsurprisingly, a hotbed of infectious disease, including HIV and tuberculosis.

Zuhura Hussein was born in Kibera 38 years ago and never left. (Her roots in Kenya are deep: she is descended from the Nubians conscripted in the Sudan by British colonial forces a century ago and permitted to settle in what was then a lush forest.) A mother of three and grandmother of one, Hussein is one of 140 community health workers attached to one of the clinics that serve the slum. She encourages the people of Kibera to venture out for medical checkups and vaccinations; she also urges patients with HIV or tuberculosis to take their medications every day. She and thousands of Kenyan workers like her are crucial to the success of many global health initiatives, such as the U.S. President’s Emergency Plan for AIDS Relief (PEPFAR), which spends $500 million per year in Kenya alone.

The day of my visit, Hussein and I snaked through a four-foot-wide alleyway, past two girls peering into a fragment of shattered mirror as they braided their hair next to a pot of boiling meat. We ducked into one of Kibera’s dim, cramped habitations. When our eyes adjusted, a bed came into focus. A figure wrapped in blankets stirred. The woman (who wasn’t well enough to give permission to use her name) was 48 years old but looked 75. She was HIV-­positive and was struggling with a severe case of tuberculosis. “The TB has come back—so many times, I don’t know why,” Hussein said. Asked what she needed, the woman whispered in Swahili: “I want just food—only food.” Amid this scene of despair, a phone rang; Hussein reached into her dress and produced a Nokia model 6070. Later, I scrolled through her contact list and found more than 300 names, from Abdala to Zubeda. Many, she said, were patients she’s worked with.

Phones like Hussein’s hold great potential to improve the way health services are delivered. One major study demonstrating as much was started five years ago by Richard Lester, a Canadian infectious-disease specialist. After arriving in Kenya for a research fellowship, noting the ubiquity of mobile phones, and recognizing that the country has only one doctor for each 6,000 citizens, Lester and his team developed a communication link with HIV-positive patients at three health centers, asking them weekly by text message whether they needed any assistance with their antiretroviral drugs (ARVs). Once 500 people were participating, Lester conducted a clinical trial. The results, published in 2010, showed not only that a higher percentage of those receiving the reminders said they took their drugs regularly, but also that viral loads were suppressed in 57 percent of them, compared with only 48 percent of the control group. Today he estimates that expanding that system to all 410,000 Kenyans on ARVs would suppress HIV in 36,000 people, saving $17.4 million in health-care costs by averting the onset of AIDS or making more expensive drugs unnecessary.

More evidence is streaming in. In western Kenya, a research project called Academic Model for Providing Access to Healthcare (AMPATH), led by the Indiana University School of Medicine and the local Moi University, recently began keeping track of 130,000 HIV-positive patients using electronic health records and automated reminders on Android phones. Now workers in 55 clinics can quickly and easily see what tests or drugs patients need. Published research suggests that the proportion of HIV-positive mothers passing the infection to their babies has dropped below 3 percent, compared with nearly 15 percent in other areas, probably because more of the pregnant women are receiving antiretroviral drugs consistently. “These reminder systems are an extremely important way to make sure all of the ts are crossed and better quality of care is provided,” says Paul Biondich, a research scientist at Indiana’s Regenstrief Institute, who co-developed the underlying open-source records system platform, called OpenMRS.

But all this groundbreaking work is still reaching only a fraction of the people who need help. Sub-Saharan Africa is home to more than two-thirds of the 33 million people estimated to have HIV worldwide. Health IT projects established the usual way—funded by donors or NGOs and run by international contractors—are benefiting relatively few of them, and they are vulnerable to financial cutoffs. Indeed, when Lester’s research funding—$719,000 from PEPFAR—ran out in 2009, two of the three sites he was servicing stopped providing the SMS messages. Lester is now back at his desk at the University of British Columbia, doing what most people do when trying to fix health care in Africa: seeking more grants. “That is the unfortunate fate of the study,” he says. “It’s been very frustrating to go from research finding to programmatic funding. I think there is an ethical obligation, when you have a clinical trial with positive results, to do everything in our power to provide it as a service.”

I had expected that Hussein—a community health worker right in Nairobi, in the heart of a much-studied slum—would have some mobile health technology on her phone. I was mistaken. She can call patients to keep in touch with them, but Hussein had no automated SMS system to remind them to take their drugs. She was not using her Nokia to report any newly discovered disease cases. She received no formal instructions or updates from it. If the phone slips into one of Kibera’s ditches, or if Hussein jumps at an NGO’s offer for a paid stint of community work (these opportunities sometimes arise only to evaporate after a year or two), the crucial human connection to dozens of people like the patient we called on may be lost.

RISE OF A STARTUP CULTURE

If you travel back up a rutted dirt road from Kibera and turn right on the Ngong Road, just past the Uchumi Hypermarket, you’ll see a five-story office building completed in 2009. From the patio ringing the top floor, a haze from diesel fumes and the cooking fires of Kibera’s shacks is visible just beyond the crest of a hill. But step inside, and it feels as if you’ve been transported to a Silicon Valley startup. Dozens of twentysomethings toil away on laptops; a few blow off steam at a foosball table; Pete’s coffee bar (not to be confused with Peet’s of the United States) doles out cappuccinos, milk shakes, and slabs of banana bread. This is a business incubator called iHub, the fruit of a homegrown information technology culture that had its coming-of-age moment in December 2007. That month, ethnic violence broke out after a disputed presidential election; at least 1,100 people died and 300,000 were displaced. Ory Okolloh, a human-rights activist, put out a call to Kenya’s loosely knit blogging and technology community to help report on the fighting (see “Frustrated Innovation”). Several people responded, including Erik Hersman, Juliana Rotich, and David Kobia. In 48 hours, Kobia had written the first draft of an incident-reporting platform called Ushahidi, the Swahili word for “testimony.” Now any Kenyan could send in an eyewitness report by text message, and it would be reviewed and then posted on an online map. Ushahidi has since been used widely, in countries including Haiti, South Africa, Russia, and the United States (where it helped map flood-related problems on the Missouri River).

An incident in Ushahidi’s formative days planted the seed for iHub. Ushahidi’s developers had initially offered the technology free of charge to the Kenya Red Cross Society and other NGOs monitoring the violence. But the NGOs didn’t want it; it wasn’t part of their existing plans and funding models. “We had so much resistance,” Hersman recalls. “We kept trying to say, ‘It’s free, we will hold your hand, we will help you communicate with the public to say how you are providing a service.’ They weren’t willing to do anything with it.” The experience taught Hersman, 36, that more might get done if local hackers would get together, write more code, and start some companies that had sustainable business models. He pitched the idea of a corporate-funded space for the tech community to companies including Google and Nokia. “Nobody wanted to set up a hub/lab space in Africa,” he says. “That sounded crazy in 2008.” Finally, the Omidyar Network, the philanthropy founded by eBay founder Pierre Omidyar, donated $200,000 to fund iHub for two years. Other donors, including Nokia, Google, and the African ISP Wananchi, stepped up with equipment and high-speed Internet service.

The incubator opened in 2010 and now counts more than 6,000 members, with an average of 1,000 new applications a year. Most members are merely part of iHub’s online community, but more than 250 of them use the space. Some 40 companies have launched from iHub, and 10 have received seed funding from venture capitalists. The most successful so far is Kopo Kopo, which helps merchants manage payments from M-Pesa and similar services. One key to iHub’s growth is that Kenya’s IT infrastructure has improved significantly. The first Internet fiber connection landed at the Kenyan coast in 2009 (previous service had come through satellite dishes in the Rift Valley), and the country’s first truly mass-market Android smart phone went on sale in 2010, for $80. Safaricom now counts 600,000 smart phones of all kinds on its network and expects them to make up 80 percent of the market by 2014.

Inevitably, this petri dish produced a mobile health startup. Shimba Technologies, led by a couple of University of Nairobi graduates named Steve Mutinda Kyalo and Keziah Mumo, created a platform called MedAfrica with the simple goal of providing basic health information to Kenyans in the face of the national doctor shortage. So far, MedAfrica offers lists of doctors and dentists taken from government registries, plus menus for finding basic first-aid and diagnostic information. “What we want is for the common man to have the right information in his hand,” says Kyalo, the company’s CEO. “We can’t replace the doctors, can’t replace the hospitals, but we can improve access to relevant information.”

MedAfrica illustrates the power of local entrepreneurship. Though it has few connections with the medical community or the health ministry, its health-care app has been downloaded on  43,000 phones, and the company is still only halfway through $100,000 in seed funding. The service can be delivered through an app or through a mobile Web interface (nearly all Kenyans who access the Internet do so through mobile devices). Soon it will be available through SMS—an essential feature, because 85 percent of Kenyan mobile-phone owners don’t yet have Web access. Kyalo hopes to aggregate other medical apps on the platform and ultimately sell sponsored messages from pharmaceutical companies, health-care providers, and others.

I joined Kyalo and one of his colleagues, Jackie Cheruiyot, a leukemia survivor who has firsthand experience with the scarcity of Kenyan health services, as they hit the road to pitch the app to potential users. They faced skepticism as they made cold calls in Narok, two hours west of Nairobi. In Narok’s bustling downtown, with a mosque at its center, women picked over vegetables and potatoes while men hauled sacks of grain on carts or peddled small red sausages; a three-man team dug out a culvert to clear green water from a roadside ditch. At the Narok District Hospital, a government-run facility festooned with posters and stickers from the U.S. Agency for International Development, the Centers for Disease Control, and other donor groups, Cheruiyot knocked on the door of the head nursing administrator, disarmingly singing out “Welcome!” But the administrator shooed us away with this reproach: “You must get authorization from the health ministry.” The idea of a startup company trying to participate in health care is still far too alien. “Government is the hardest nut to crack,” Kyalo said as we beat our retreat.

Other startups have emerged without iHub’s help. Changamka Microhealth sells health payment cards that can be charged up through M-Pesa. You buy a card, preloaded with 450 shillings, for 500 shillings—Changamka makes 50 shillings, or about 60 cents, on each—and then you make M-Pesa payments to add money until you’ve got enough for a given procedure. (There’s even a special card for pregnant women; a hospital childbirth costs about 4,000 shillings, or $50, a sum many cannot easily or safely save at home.) And Intellisoft Consulting builds electronic medical records platforms, which many clinics throughout Kenya employ. It provides this infrastructure using OpenMRS, the open-source platform initially developed by Indiana University researchers and the NGO Partners in Health, which continues to evolve with help from dozens of developers, many of them Kenyans. Such companies are developing a crucial local capacity to improve health care, says Paul Biondich, the Indiana University researcher. “We have to do things like iHub, help the local people get organized, and set up a way that the cash coming in to support health is increasingly available to these kinds of startup entities,” he says.

FROM M-PESA TO M-HEALTH

Larger Kenyan companies are starting to figure out how to do that—using the promising ideas and developing business models from them. On the same day Kyalo launched MedAfrica last November—with doctors’ phone numbers as a major selling point—Safaricom announced that it was launching its own doctor-calling service. In a nation with few doctors and no free 911 service for medical emergencies, residents can now at least speak to a doctor for about 25 cents per minute. The service already fields 500 calls per day, but while it’s helping, it also painfully illustrates the challenges facing its potential users. Nzioka Waita, Safaricom’s director of corporate responsibility, described a call that came through in January from a woman desperate for help because her husband wouldn’t wake up. During the course of the call, her mobile-phone credits ran out, though the doctor was able to call her back. Safaricom says it is in discussions with a partner willing to subsidize future emergency calls so they can’t be cut off.

Safaricom is also working with partner companies to do for health care what it did for banking with M-Pesa. A system now being designed—initially for pregnant women in several rural districts—would let community health workers create an electronic medical record for each patient, update the records, and send health information and reminders to the patients’ phones. In many ways, the project would adapt technologies pioneered by groups like AMPATH and allow them to be rapidly scaled up.

The $2.3 million project is expected to roll out this spring. The idea is for community workers, armed with a phone and sheets of ID cards bearing bar codes, to issue a card to a woman and scan the code with the phone’s camera, registering the woman’s identity. The woman, if she has a phone, would then receive text messages offering health advice and reminders of upcoming appointments. On each return visit, new information, tied to the bar-code identifier, would be uploaded by SMS to a central database. Crucially, the system would build on existing mobile billing and banking platforms. Each transaction uses phone minutes, which are mostly prepaid in Kenya and could be subsidized by donors.

And in a nation where 75 percent of the population is not covered by any health insurance, Safaricom envisions enrolling people in insurance programs and letting them make payments via M-Pesa. About 50,000 laborers have recently started doing just that. Handling the financial side of health care with mobile phones, say Biondich and others, would make it possible to bring more people into the system and thus improve the nation’s health. Mobile payment also provides a potentially efficient way for donors to fund health care.

Nairobi’s Kenyatta National Hospital, one of the largest hospitals in sub-Saharan Africa, has a distinctly 1930s feel, with painted wooden doors and hand-painted signs. One day Ambrose Kwale, the hospital’s director of IT, showed me around. There was a new 25-bed isolation unit for multiple-drug-resistant tuberculosis, and a grassy spot outdoors where several people who appeared to be in their 50s or 60s were sprawled, some curled in the fetal position. These were cancer patients. Many had traveled overnight, referrals in hand, for appointments with some of the few oncologists in East Africa. (One hospital IT initiative is to install a telemedicine facility to help patients at regional medical centers avoid the trip to Nairobi to see specialists.) A woman who appeared to be in her 30s, wearing a pink jacket and a flowered shawl, leaned against a concrete pillar, short of breath. When Kwale approached, she weakly handed him a piece of paper marked up in blue pen. She had traveled 50 kilometers to see a specialist for her breast cancer, and now she was alone, exhausted, and at the wrong place on the campus. A pale blue cataract blighted her left eye, and a look of fear and pain shadowed her face as she rested her head against the pillar. Kwale could only call for an orderly to help the woman find her way.

Mobile technologies offer great potential to help patients like her—to keep track of their care, provide reminders, and give them broader access to expertise. And experience is showing that local talent can create the technology.

The challenge lies in organizing this emerging talent so that it can tackle large-scale projects. Last year USAID, a major funder of health projects in Kenya and other developing countries, requested proposals for help creating a unified, Web-based national health information system that would be “host country owned.” The five-year, $32 million contract went to Abt Associates, a consultancy based in Cambridge, Massachusetts, which has done extensive work in global development projects. But although it has expertise, so does the new tech class back in the host country—which also has a long-term stake in the solution and no U.S. overhead. “If you talked about an RFP for $32 million at iHub, people would go nuts! You’d fund 500 startups for that,” CHAI’s Jackson Hungu says. “And this country’s public health delivery would be changed forever. I have no doubt about that.”

David Talbot is Technology Review’s chief correspondent.