The UN Foundation recently put out a fairly comprehensive report summarizing mobile health (mHealth) initiatives around the globe. Many of these have been anecdotally successful and give good reason for future optimism, but as Matt Over from the Center for Global Development so aptly pointed out, there have been very few rigorous evaluations of mHealth programs to date. At best, we have measures of process outcomes (time saved, number of reported health events, patients enrolled, money saved, etc.), which are fantastic data but not necessarily indicative of the outcomes that are most important. Whether the application is epidemiological surveillance, community health worker networking, or diagnostic treatment and support, an mHealth initiative fundamentally seeks to improve health- this is quantified by measures such as mortality and morbidity rates, cumulative disease incidence, etc. What if the ability to remotely monitor patients leads to a complacency that actually increases morbidity and mortality in a patient community? Suppose an epidemic tracking program that relies on remote forms receives 500 updates in a month, but there is a framing bias that confounds the results and ends up wasting resources? I am certainly playing devil’s advocate to some degree, but the point is simply this: process measures lead us to believe that mHealth interventions should be improving population health outcomes, but until we measure them directly we don’t actually know. It is vitally important to understand how a program with an accepted structure will predictably affect health outcomes. Furthermore, if initiatives hope to scale, their efficacy must be proved beyond a shadow of a doubt. Governments, partner organizations, donors, the medical community- these are just a sampling of the people that need to be convinced.
Given the recency of mHealth, it is understandable that process measures have been the data norm- they are focused on clear targets and can be acquired within a short period of time. A randomized controlled trial that seeks to measure changes in population health, on the other hand, is an undertaking that requires an added degree of planning, resources, and time. This is precisely what we at FrontlineSMS:Medic intend to do. Without divulging any specifics prematurely, we have been talking to several major healthcare delivery and surveillance organizations in Bangladesh and should be finalizing our partnerships shortly. The trial will be large in scale, randomized across several villages, and will take several years to complete. As of now, it will be the first of its kind to quantify changes in population health outcomes as a result of a CHW-empowering texting program. More to come.
