Why is it taking so long to scale up digital health? Three years ago, Digital Square was launched as a partnership of 40 organizations working together to solve three well-known barriers to scale in digital health: (1) fragmented, uncoordinated investments, (2) lack of investment into interoperability and other scale-related requirements for health software, and (3) large capacity gaps in digital health. In response to each of these barriers, Digital Square has been building, experimenting and learning about solutions. To coordinate and align investments, Digital Square has been iterating and learning about governance structures, procurement mechanisms, knowledge management platforms and a series of financing mechanisms with the aim of making coordination easier for all involved. To improve the pathways to scale for global goods, Digital Square has been learning about marketplace structures, global good development needs, and required norms, standards, policies and guidelines to support scale. To strengthen capacity, Digital Square has been learning about governance and coordination strategies for country digital transformation efforts, and approaches to supporting digital health learning for country government leaders and technocrats. This TED Talk is designed to share what we’ve learned, provoke discussion on unanswered questions we are grappling with, and engage the community in shaping the future of Digital Square. We will share achievements, failures and lessons learned; where the initiative is heading; and why that matters for both scaling digital health global goods and accelerating country-led digital transformation of health systems. At the close of this TED Talk, we would like to launch version 0 of our strategy for community input, reflection and engagement. Our hope is that with this session, we provoke conversation about where and how we as a community can work together to help accelerate the scale of digital interventions in health systems.

The pace of digital innovation is accelerating. As a result of consistent investments, new capabilities, and improving systems, the volume of data collected through traditional and novel methods has rapidly ballooned in recent years. These data sets are in turn informing new emerging technologies, driven by machine learning systems that draw on existing data to tailor insights. However, these exponential technologies learn from the material they are provided by human beings – data that sometimes excludes women and girls, youth and adolescents, minorities, and other historically marginalized groups. If our technology learns from these incomplete data sets, its insights will serve only a fraction of the global population, and millions will be left behind in the next wave of innovation. This talk will address these challenges through research anecdotes and personal narrative, and will challenge the audience to think differently about how it innovates and the technology it promotes.

In a world where 400M people have no access to healthcare and 2B have poor access (unaffordable, distant), telemedicine is seen by many as one of the most promising routes to close these gaps and speed universal health coverage. Indeed, experience in many low-resource settings has shown that digital technologies can disproportionately help improve prevention, diagnosis, treatment and follow-up of patients. Telemedicine in particular is credited with reducing the cost and burden of travel to clinics for care, increasing access to doctors and improving patient outcomes. Among the more innovative evolutions of telemedicine for LMIC settings are small portable units, simple enough for frontline health workers to operate, that perform nearly all WHO essential diagnostics, while automatically digitizing and transmitting results so a remotely located doctor can interpret test results while meeting patient and HW via low bandwidth video consultation. By moving care to patients instead of the other way around, telemedicine contributes to a shift in the balance of power from the traditional vendors (doctors) to a new model where services are made more accessible and more flexible depending on the consumer's needs. Can telemedicine can find viable price points and workable business models? Is telemedicine a true disrupter, able to dramatically transform healthcare access for poor and remote populations? Or has the global health community pinned too many hopes on these promising technologies?

Over the past decade, we have made marked progress in formalizing the field of digital health. From a 2018 World Health Assembly Resolution to the 2019 launch of the Digital Health Guidelines, digital health has gained unprecedented centrality in the health systems strengthening discourse. A number of countries now boast digital health projects operational at subnational or national scale, providing information services to clients or supporting the delivery of care in a community or clinical setting. Technical investments in digital health "global goods: and in 'backbone' systems to facilitate interoperability have been made by major donors, while the Broadband Commission and ITU have advocated for improvements in the digital infrastructure which supports digital systems. There is, however, a few important "elephants" in the room that are seldom discussed in large fora -- related to whether the health systems in which digital innovations are being introduced are prepared to accept unprecedented levels of data clarity and transparency. In many LMIC settings, decades of unintentional and intentional dysfunction have been masked in the unreliability, obscurity and sluggishness of paper-based systems. This talk will focus on bringing to the forefront a number of critical threats to the widespread adoption and use of digital systems -- important active and passive obstacles to changing status quo. Digital data systems risk uncovering a range of inconvenient, uncomfortable or challenging data which, given its temporal and factual accuracy, challenge well-established practices and sometimes even successes. At the macro-level, systems may begin to report lower coverage than previously reported with paper records and multiple levels of 'editing'. Client denominators previously cobbled together manually may shrink or grow with the use of unique digital identifiers. Performance metrics like attendance and hourly performance can be estimated by mining metadata, revealing patterns of inefficiency easily hidden in paper ledgers.

When your music app asks you to log in to go online, then tells you you need to be in online mode to log in? That’s #offlineworst. In the last 5 years the digital tools that back development have made huge strides in their maturity. Many teams have had the time to invest in modern UI’s, making it feel like software in our market is “catching up.” It can be easy to forget that underneath the hood these apps are solving technical challenges like asynchronous offline transaction management that are seriously hard. So hard, in fact, that large and well financed consumer software teams mostly avoid even attempting to address them. When they do the results are often equal parts frustrating and comical. In this light hearted talk we will review the at-times kafkaesque outcomes that have arisen from popular consumer software’s attempts to implement requirements which are foundational to the most basic software tools used in our field. We use this backdrop to celebrate the continuing success of systems in digital development to reach increasing scales, but also as a sobering reminder of the invisible and essential ‘weight’ that these systems carry with them; and the cost that comes with it.

Two years ago the opening keynote of the Forum asked panelists and participants to place their bets on the future of digital health. One of the most common (and, we're happy to report, winning!) bets was on increased donor coordination. USAID created Digital Square to enable that co-investment, and as of early 2019 an initial $22M of investment by USAID has been matched by a further $14M in co-investments by major digital health donors including BMGF, OGAC, and GIZ, to name a few. This could be an Ignite-style session or a panel looking at what's working, what have we learned, what do we want to replicate in the future, and what change is needed to get there?