How Do We Cure mHealth Pilotitis? Critical Lessons in Reaching Scale

Successful pilots are in abundance, but most of the sector has been slow to reach scale. This series of posts offers nine different perspectives from leaders in mHealth.

mHealth has the potential to transform healthcare, particularly for the hardest-to-reach women and children around the world. The debate about exactly how, when, and in what form is alive and well. Successful pilots are in abundance, but most of the sector has been slow to reach scale. In short, the sector has a case of mHealth Pilotitis. In the first debate of a series on mobile health, the Skoll World Forum on Social Entrepreneurship partnered with Johnson & Johnson and Stanford Social Innovation Review to surface important lessons and learning from some of the world’s leading organizations who have taken mHealth services to scale. This debate will also set the stage for a larger discussion on mobile for development at this year’s Skoll World Forum in Oxford, UK.

As a co-founder of the UNICEF Innovation Team I had the opportunity to design, build, pilot, evaluate, and scale an mHealth program in Zambia called Programme Mwana. It was our team’s 9th project that utilized mobile phones to strengthen government service delivery and UNICEF’s operations in the field. I was able to leverage our team’s previous experiences to design a system and program that could scale to address maternal and newborn health issues in rural settings.

With an HIV prevalence of 16.1 percent among women in Zambia, prevention of maternal to child transmission (PMTCT) of HIV is a major issue. Within PMTCT, early infant diagnosis (EID) of HIV is a particularly significant problem. More than 50 percent of infants are not tested, leading to much higher infant mortality rates. When a child is born to an HIV positive mother, normal HIV tests are not possible for the child, as the mother’s antibodies are in their blood. Rather than a rapid-test, a technique known as PCR analysis is used to look at the DNA of the antibodies. In Zambia, there are only three PCR machines for a country roughly the size of France. Additionally, the country has very poor road infrastructure, and logistics is cited as a major bottleneck in EID. We designed Programme Mwana to see if we could improve infant mortality by addressing these particular bottlenecks.

We created two SMS systems: Results160 and RemindMi. The first, used by clinic staff, is used to deliver results from the lab to health clinics. The second is for community health workers (CHWs) and is used to bring the mother back to the clinics to receive the results. During our evaluation, we found that we were able to speed up the time between when the samples were collected to when the mother received the results by 56 percent. Furthermore, 30 percent more results were being delivered to the mothers because they were digital—paper copies often get lost.

In the spring of 2011, the Zambian Ministry of Health officially decided to scale Programme Mwana to 414 health facilities in Zambia that provide EID services. The scale-up is taking place over three years, assisted by a wide range of government and NGO partners. Programme Mwana is currently in more than 364 facilities.

From the very beginning, we designed the system with scale in mind. We chose to see if we could create significant health impacts using only SMS and clinic workers’ personal phones. Purchasing phones and creating the IT capacity to support them, may not seem that expensive, but when multiplied by the number of clinics in the entire country, it becomes quite costly. And while SMS is an expensive information medium, as we scaled, we were able to negotiate much lower rates than what could be purchased on the street. We spent the time negotiating with all of the major telecom companies in Zambia, knowing that we couldn’t scale to the whole country unless it worked on all networks. We also used the open source software RapidSMS, eliminating licensing fees and allowing us to replicate the program in other countries like Malawi. We could use every dollar saved in procurement to add another facility during scale up.

We also designed the system in the most rural communities we were seeking to serve. Our team of researchers and programmers moved for a month to the town of Mansa, twelve hours from the capital of Lusaka, to create the Programme in an iterative way, involving the local community members who would be using it. This human-centered design process, assisted by Frog Design, was critical to scale, as it assured the software would make sense in the local context and be easy for new health facilities to adopt. Through this participatory process, we learned that the system could be used in facilities that had no mobile coverage if we designed it to allow staff to use it at nearby markets where there was coverage.

Most importantly, technology was the foundation, not the focus, of the program. The team went to great lengths to understand and strengthen the existing health interventions, not replace them with something new. And we did this in close partnership with the government and partner NGOs.

Despite this, we initially had issues raising funding for scale. We had a low-cost, well-designed, and well-evaluated program with a mandate from the Ministry of Health to scale, and yet it was not easy to sell donors on the scale up. Donors want to be perceived as innovators, and the NGOs they fund often compete for funding by differentiating their approaches. We were asking them to fund collaboration and mainstreaming rather than developing their own separate mHealth systems. Only after Johnson & Johnson stepped in with some seed funding did the programme start to scale. The initial funding issues delayed the scale-up by close to a year, and affected the momentum and continuity of the project.

Now with the scale up in full swing, Programme Mwana shows how designing mHealth solutions with scale in the mind from the beginning can help them escape the intense gravity of pilotitis.