Rapid advancements in digital health — spanning technologies like telemedicine, personalized digital nudges, and large language models — present new possibilities for addressing persistent problems in healthcare access, cost, and quality. Yet there is an urgent need for effective strategies to ensure digital health innovations improve healthcare delivery and health outcomes benefiting underserved populations especially in low- and middle-income countries (LMICs).

A new viewpoint paper published in the Journal of Medical Internet Research presents an analytic framework for evaluating the foundations of digital health innovations. Focusing on the digital health landscape in LMICs in South and Southeast Asia, the paper’s co-authors provide actionable recommendations for tailoring and implementing digital health solutions across diverse global health settings.

The co-authors include Siyan Yi, an associate professor at the National University of Singapore’s School of Public Health and a former postdoctoral fellow with the Asia Health Policy Program (AHPP) at Shorenstein APARC; Stanford health economist Karen Eggleston, the director of AHPP; Kochukoshy Cheruvettolil, former senior strategy officer with the Bill & Melinda Gates Foundation’s Digital Health and Artificial Intelligence division; Kiran Gopal Vaska of the National Health Authority, Ministry of Health and Family Welfare in New Delhi, India; and several collaborators from Stanford’s School of Medicine, among others.

Citing the advantages of a “human-in-the-loop” approach to Artificial intelligence (AI) — a collaborative approach that integrates human input and expertise into the lifecycle of machine learning and AI systems — the co-authors emphasize the potential benefits of a “researcher-in-the-loop” approach to digital health innovations in LMICs. Their collaboration for this paper was conceived during a panel discussion on digital health innovation featuring Cheruvettolil, Vaska, and Yi. Hosted by AHPP in October 2023, this event was part of APARC’s Autumn 2023 seminar series, Exploring APEC’s Role in Facilitating Regional Cooperation.

Sign up for APARC’s newsletters to receive our scholars’ research updates > 

Building Social Value in Digital Health Adoption

While the promise of digital health is universal, its implementation must address local needs and be tailored to promote adoption, recognizing that underserved populations lack adequate infrastructure and equitable access to healthcare services and digital innovations. Yet it is crucial to balance the social benefits of digital health innovations with the social costs of adoption. It is also imperative to gather evidence and insights on successes and failures when adopting such innovations in diverse local contexts.

To address these needs, the paper's coauthors advocate for a dual approach that combines a top-down, evidence-based strategy focused on designing and scaling national digital health infrastructures with the bottom-up creation of a vibrant ecosystem of local experiments generating shared experience.

“This two-pronged approach of top-down design and bottom-up experimentation, enabled through careful evaluation of effectiveness and impact, allows economies with different starting conditions to seize opportunities to ‘leapfrog’ towards more robust, resilient health systems fitting their contexts rather than imitate the development path of any given current high-income country or region,” the coauthors argue.

This theoretical framework recognizes that the social net benefit of digital innovations is heavily influenced by the initial conditions and the intended outcomes shaped by the local context. For example, in LMICs, persistent disparities in digital access and digital literacy determine the unfolding of the introduction, adoption, and integration of digital health interventions into the broader health system and compound existing inequities in access to healthcare.

Still, the co-authors emphasize that digital technologies can potentially deliver significant, even transformative benefits in LMICs that surpass the impacts seen in high-income health systems — provided that robust oversight and stewardship uphold the “First, do no harm” principle.

Integrated Digital Health Solutions

One of the coauthors’ key recommendations is to create a robust digital health infrastructure with adaptable data architectures, interoperability, and stringent privacy standards. They underscore the necessity of creating an integrated data ecosystem, as exemplified by India's Ayushman Bharat Digital Mission (ABDM), which facilitates personalized healthcare and evidence-based policymaking through effective data management.

Yet varying data standards across countries can pose challenges to comparability and effectiveness, in addition to ethical and legal dilemmas around sensitive health data. Singapore's Personal Data Protection Act and India's Digital Personal Data Protection Act are examples of regulatory frameworks protecting sensitive health information. These regulations are crucial for building trust among patients and healthcare providers — an essential prerequisite for successfully adopting digital health solutions.

The paper also highlights the potential of mobile health (mHealth) interventions and AI applications in revolutionizing healthcare delivery by bridging geographical barriers, expanding access to underserved populations, and shifting care from hospitals to communities. For instance, mHealth initiatives in India and Cambodia leveraging the widespread adoption of low-cost smartphones have shown promise in improving maternal and child health outcomes in low-income settings. Social media platforms further amplify the reach of public health campaigns and community-based interventions by utilizing trusted local voices.

AI is a new transformative force in digital health, enabling advancements in precision health, from diagnosing complex conditions to supporting clinical decisions. For example, in Thailand, a deep learning algorithm integrated into the national diabetic retinopathy screening program provides real-time, specialist-level diagnostics in community settings. AI solutions like this hold particular promise for addressing the challenges of low-resource health systems, including shortages of trained staff and limited equipment.

On the other hand, the coauthors caution against the risks associated with digital health technologies, including misinformation and ethical concerns. Clear data collection, storage, and sharing guidelines are critical, along with measures to safeguard sensitive health information.

Social Legitimacy and Local Partnerships

The paper emphasizes a call to action for researchers, policymakers, healthcare professionals, and civic organizations to invest in digital health infrastructure and foster interdisciplinary, local partnerships to promote need-driven, culturally sensitive, and equitable interventions. By prioritizing community engagement, the co-authors argue, digital health innovations can bridge geographical gaps, enhance healthcare accessibility, and ultimately improve health outcomes for underserved populations.

For example, the NGO Noora Health has successfully leveraged digital health across India, Bangladesh, and Indonesia to improve access to caregiver training. Successful features of their programs include mobile chat services and digital curricula integrated into comprehensive care delivery models that utilize local healthcare systems through partnerships with local governments, policymakers, and community health workers.

Ultimately, the co-authors note that the key to success lies in effectively stewarding health sectors to address priority population health needs and improve equity. “Policies should support the generation of evidence assessing digital health applications so that patients, providers, and policymakers can ask and answer the right questions in a suitable timeframe to enable a virtuous cycle of learning and improvement.” In resource-constrained LMICs, it is also essential to consider the financing levers in the public and private sectors to understand the feasibility of adopting and scaling up. In these contexts, private philanthropy can secure resources, mitigate risks, and generate evidence about best practices and scalability.

In a landscape where technology and healthcare intersect, the path forward must include prioritization of local contexts, engaging communities, and building a robust digital health infrastructure that serves the needs of all. The future of healthcare may depend on it.

On October 18, as part of its autumn 2023 seminar series on APEC in advance of the organization's meeting in San Francisco in November, Shorenstein APARC and its Asia Health Policy Program (AHPP) presented the series' second event, Asia-Pacific Digital Health Innovation: Technology, Trust, and the Role of APEC. The featured panelists were Kiran Gopal Vaska, Director of the National Health Authority of India, and CK Cheruvettolil, the Senior Strategy Officer, Digital Health and Artificial Intelligence, at the Bill & Melinda Gates Foundation. Siyan Yi, the Director of the Integrated Research Program at the National University of Singapore and a former AHPP fellow, moderated the conversation.

While India is not an APEC member, Indian initiatives are examples of leveraging technology to better the health of the most vulnerable citizens in low- and middle-income countries (LMICs). Kiran Gopal Vaska gave an overview of the Ayushman Bharat Digital Mission (ABDM), India's latest health initiative that focuses on the interoperability of health records, services, and health claims. He stressed that ABDM was built on previous digital infrastructure, like Aadhaar, the national digital identity system, and Digilocker, a digital storage scheme for citizens' health and other records.

The approach India has taken is for the government to build the rails—the infrastructure of the system—and create a space where the private sector can develop applications integrated with that space through application programming interfaces (APIs), avoiding the siloing that can hamper the interoperability of data.

Regarding health data, privacy is a crucial concern at the patient level. ABDM addresses this concern through the use of a consent artifact. Individuals decide whether hospitals or other medical service providers have access to their data, and this access has levels of granularity: you can share specific portions of 7 different data types, like immunizations or prescriptions. You can limit that sharing to a particular period, like one day.

Also participating on the panel was CK Cheruvettolil, who discussed strategies by the Bill & Melinda Gates Foundation in leveraging the power of mobile phones to augment the work of Accredited Social Health Activists (ASHAs), the more than one million female frontline health workers in India. ASHAs can use mobile phone cameras, sensors, and streaming data to better care for low-birth-weight babies and other patients. 

He explained the critical role of taking local context into account when developing software by using the example of pregnant Indian women in their third trimesters. The custom for Indian mothers, especially in rural areas, is for the child to be born in the maternal grandparents' home. If software were to store only the mother's address, healthcare workers in the grandparents' jurisdiction would not know that a pregnant woman in the critical third trimester would soon be giving birth at a local address.

Kiran Gopal Vaska noted that India had solved the technological issues, and now the task was to push for adoption. He emphasized that the technologies underlying India's digital health stack were created as public goods for the world, and for LMICs to support each other in advancing digital health technologies, the key was interoperability, "using standards that are accessible and acceptable worldwide."

How do civilians make their way through complex, violent environments? How do people form judgments and make decisions about their survival, or other goals? These are some of the questions APARC’s 2022-23 Shorenstein Postdoctoral Fellow on Contemporary Asia Aidan Milliff has sought to answer in his research, which employs different tools to study political violence, ranging from interviews and oral history archives to decision-making experiments.

APARC awards the Shorenstein postdoctoral fellowship annually to support recent PhDs who research contemporary political, economic, or social change in the Asia-Pacific region, or topics in international relations and international political economy in the region. Milliff earned a PhD in Political Science from the Massachusetts Institute of Technology, where he was affiliated with the MIT Security Studies Program and Harvard’s Lakshmi Mittal South Asia Institute. He is a former James C. Gaither Junior Fellow in the South Asia Program at the Carnegie Endowment for International Peace.

For Milliff, the Shorenstein postdoctoral fellowship afforded the opportunity to advance his research and refine a forthcoming manuscript project. He recently presented his work in a South Asia Initiative seminar entitled “How Indians See China.” Examining some 60 years of data on Indian public attitudes towards China, Milliff’s research shows clear historical trends in Indian opinion towards China. In recent years, Indian views of China had been souring well before the border crisis of 2020, and before government policy began to harden. Using a rich body of new polling data, Milliff examines how the government is constrained by and seeks to shape the public’s opinion towards China. 

We caught up with Aidan to discuss his research and experience at Stanford this academic year. The conversation has been slightly edited for length and clarity.

First off, congratulations on receiving the 2022 Best Paper Award from the American Political Science Association Conflict Processes Section for your job market paper! How did you develop this paper project?

Thank you! This paper is kind of the article-length version of a book manuscript that I have been working on during my time at APARC. The book is trying to tackle a big, amorphous question: how do ordinary people make decisions about their safety and survival when they are threatened by political violence?

The paper focuses a little more narrowly on how targeted civilians behaved in one important historical episode of violence, a short and very deadly pogrom in India in 1984. It is organized around an empirical oddity: Why did we see in 1984 that very similar people who were facing the same threat sometimes made quite different choices about how to behave in order to survive?

I argue that it's because choices about survival, especially during sudden, high-intensity political violence, depend a lot on how people interpret their environments —and reasonable, well-informed people can really disagree about these interpretations.

How has your time at APARC as the Shorenstein Postdoctoral Fellow aided your research?

Being a fellow at APARC has been amazing in two ways. First, the fellowship is a big gift of time. Having an entire academic year to focus on producing research, especially on moving this book forward, is really beneficial, and I'm so grateful to APARC for giving us postdoctoral fellows time to get big things done.

Second, the community at APARC and Stanford has been outstanding. Coming out of graduate studies, where most people are in an environment defined by a disciplinary identity, it's really exciting to be part of a group that's focused on an area. I've learned a lot from exchanges with sociologists, economists, anthropologists, and policy practitioners who all share an interest in the Asia-Pacific.

What other aspects of your time at APARC have you benefited from?

APARC has been an outstanding home base for participating in the broader Stanford community. Stanford is kind of a dream for a political scientist doing a postdoc because there is such a big community of scholars across various departments, centers, and schools who share an interest in social science problems. It's been great to interact with people all across campus, but you hardly have to leave Encina Hall for this kind of cross-pollination. It's been great to learn from colleagues downstairs at CDDRL and CISAC, and down the hall in the Political Science Department.

Are there any people at APARC that you particularly benefited from working with?

I'm very lucky to be at APARC with a small cluster of other people focused on South Asia. Working with Research Scholar Arzan Tarapore and visiting fellow Nirvikar Singh, both of whom are experts on very different aspects of South Asia, has been great.

What is on the horizon for you? What's next?

This summer, I'll be joining the Florida State University faculty as an assistant professor of political science.

Any advice for students interested in your field?

It's an exciting time to be studying the politics of South Asia, and there are some very important questions that still need answers (or need better answers). Don't be alarmed or deterred if the most urgent and intellectually compelling questions are hard to fit into the disciplinary categories you are familiar with. 

This commentary was originally published by The Lowy Institute.

The latest AUKUS announcement ignited a furious debate. How should we weigh the risks and rewards of the plan for Australia to acquire nuclear-powered submarines? Critics, from former prime ministers to sceptical analysts, questioned the plan’s strategic efficacy, its opportunity cost, its feasibility, and its impact on Australian sovereignty. Well may these debates continue, but the intent is plain: to change the military balance of power in the region.

With that, AUKUS stands in bold contrast to another grouping of partners, the Quad, comprising Australia, India, Japan and the United States. It may now be time for the Quad to take some lessons from AUKUS.

In purpose and composition, AUKUS and the bilateral partnerships that underpin it can scarcely be compared to the Quad. The Australia-US alliance is a treaty-bound security pact resting on decades of trust and institutional connections. The Quad has no formal mandate, no military role, and brings together major powers that have only recently begun to consult on regional matters. The Quad cannot be expected to function like an alliance.

But if the Quad’s members decide to assume a security role in the region, they could learn something from the Australia-US alliance.

As part of a research project last year, my collaborators and I invited Australian government officials to study and role-play whether the Quad could deter Chinese aggression. In the “matrix game” we designed, the officials tried to dissuade Beijing mostly by signalling their displeasure in a range of ways. They issued statements clarifying their position, sent ships on patrol, and offered more statements. Conspicuously, they rarely sought to change the material balance of power to make aggression more difficult or costly. China, in that exercise, was undeterred and achieved its objectives.

By contrast, the notable feature of the Australia-US alliance in recent years is its attempt to change the material balance of power in the region. Since US Marines began rotational deployments to Darwin a decade ago, the force posture initiatives by the United States and Australia plan to position fighter and bomber aircraft, ships and soldiers in Australia. Critics can and should debate the merits of each of these steps, but there is no denying they are changing facts on the ground – and, with AUKUS, under the water.

These material changes in military power are the types of action that would be necessary to deter a potential adversary from aggression. If they can reduce the chances that the aggressor could successfully pull off an attack at acceptable cost, they may force it to think twice. China, for example, will only be dissuaded from attacking Taiwan if it fears the military odds are stacked against it, or the overwhelming international response would scupper its “national rejuvenation”. Sternly worded letters will not do.

Sending signals designed to show resolve is an integral part of effective deterrence, but alone they are not enough. In the absence of material effects, “cheap talk” may even send the opposite message. Signals without action lack credibility. As in our matrix game simulation, they may convince the adversary it faces no real consequences, inadvertently green-lighting its plans for aggression.

If the Quad chooses to take on a deterrence mission, it could use a range of material actions, drawing inspiration from the Australia-US alliance. Over the long term, and most ambitiously, changing the material balance of power would involve its members building new military capabilities. AUKUS sets a high bar, but Australia’s recent Defence Strategic Review is expected to have likely recommended a host of other, more achievable modernisation initiatives.

Quad members could also shift the material balance quickly and cheaply by repositioning existing military forces. Quad members have valuable real estate that, with new access and basing arrangements, could significantly disrupt Chinese military planning. Australia’s Cocos Island, and India’s Andaman and Nicobar Islands, for example, both lie within tantalisingly close range of key chokepoints and China’s military facilities in the South China Sea.

This is a risky business. The Quad taking on a new military role would trigger political sensitivities, among its members and especially in the region. So delicate messaging would be of utmost importance. Indeed, new military activities undertaken by Quad members need not even be branded as Quad initiatives. Observers also fear that new capabilities or operational ties may constrain Australia’s future political choices. But the lack of those capabilities or integration would be even more constraining – lacking the military means to act would pre-emptively deny Australia the option to act if its political leaders so wished.

The Quad’s military preparations may also provide opportunities for China. Beijing’s intent to gain control over Taiwan, by force if necessary, is clear, abiding and immutable. But it could still exploit others’ military preparations as a pretext to make material preparations of its own. These are serious risks that demand consideration. But without risk, there can be no deterrence.