At a compassion education conference in 2018, Buddhist scholar Thupten Jinpa described new scientific evidence on the effectiveness and benefits of compassion training. Why, he asked, despite the growing popularity of this training, were we not seeing much of an impact in the world? Compassion did not seem to be “scaling up.” “If we want to make a real change,” he said, “we have to change institutions . . . . We need to seek expertise and advice from people outside our community, who understand systems.” 

Scaling up large health programs, working through complex systems, is what we do in global health. Guided by the science of epidemiology, we study how and why disease is distributed. We ask, why does disease “cluster,” affecting certain people, at certain times, in certain places? Understanding these patterns can provide the information needed to reduce the risk of disease and promote human health. So naturally, while listening to Jinpa, we wondered: could the statistical tools and methods of epidemiology also be used to “scale up” compassion? 

Epidemiology is a latecomer to compassion science. Psychologists and neuroscientists have studied compassion for decades. Spiritual traditions have taught compassion for centuries. Collectively, they have focused on cultivating the capacity for compassion in individuals. “Scaling up” requires broadening the field to include institutions, systems, and populations—which is precisely the strength of epidemiology. 

The novel idea that epidemiology might contribute to compassion science faced headwinds. Epidemiologists considered compassion to be too “soft,” too difficult to measure. Epidemiology is a quantitative science after all, for which precise measurement is crucial. On the other hand, compassion scholars and teachers warned that attempts to measure compassion would ruin it. Finding a “middle way” required innovative dialogue and interdisciplinary collaboration. We started with two exploratory meetings at the Task Force for Global Health, a review of the compassion science literature through an epidemiologic lens, and, in March 2026, a special issue of the International Journal of Wellbeing, “Towards an Epidemiology of Compassion”. Our intent was to encourage scholars and thinkers from other disciplines to contribute their insights into how an epidemiologic approach could help to build a more compassionate world.

Challenging questions

While epidemiology has proven its value in public health, we found that applying it to compassion requires wrestling with several fundamental questions. First, do we have the right conceptual framing for compassion at a population level—are we thinking about it in the right way? Compassion educators who regard compassion as an individual human skill are doubtful. Only individuals, they say—not organizations or communities—have the capacity for warm-heartedness, which is necessary for compassion. However, they concede that organizations differ in their capacity to acknowledge suffering and respond with compassion, just as individuals do: Some are more compassionate than others. A reliable framework for understanding why compassion clusters in some organizations or systems more than others could provide clues for nurturing compassion in all organizations.

A second question is whether current epidemiologic and statistical methods are up to the task of analyzing the complexity of factors, interacting at multiple levels, that determine the extent to which compassion is present and how it is expressed. These factors include individual intent; physical, relational, historical, and emotional influences; and environmental and cultural determinants, among others.

Perhaps the most critical question for an epidemiology of compassion is that of measurement: What should be counted? Even at the individual level, measures of compassion rarely capture its essence. A recent scientific review by psychologist Cassandra Vieten and her colleagues found 503 measures of compassion and empathy—evidence that consensus is lacking. At the population level, outside the controlled environment of a laboratory, the challenge is even more complex. Measures of intention, empathic resonance, brain activity, or physiologic changes—which are useful in experimental contexts—are neither feasible nor reliable for studying compassion in populations. Brendan Ozawa-de Silva and Jennifer Mascaro have argued that, at the macro level, compassion can be measured by how well it is institutionalized through policies and actions. But are such policies and actions the same as compassion? Not according to those who insist that compassion is a skill or trait only of individuals, or that, for action to be considered compassionate, we must know that it is motivated by a positive intent. 

Gaining momentum

What began years ago as a vaguely defined notion among a small group of enthusiasts is gaining momentum. As an example, “Towards an Epidemiology of Compassion” featured 23 articles by more than 90 authors from a broad range of disciplines. Significantly, the articles provided much-needed perspective on conceptual frameworks, epidemiologic methods, and measurement. 

Conceptual frameworks. Working independently, several groups have proposed new conceptual frameworks to guide an epidemiology of compassion. Each of these frameworks or conceptual models emphasizes the need to understand how compassion manifests at multiple levels, including the individual, collective, and societal levels. Some of the frameworks describe how compassion “flows” across these levels, while others focus on determinants of compassion at different stages of life. Taken together, these conceptual models highlight the emerging understanding of compassion at the collective level and open new lines of inquiry into collective expressions of compassion—actions that are possible only through the shared intention and coordinated work of many people.

We have argued elsewhere, for example, that the field of global health, which seeks to alleviate and prevent suffering at a global level, is a manifestation of compassion. Exciting recent work on organizational compassion by Ace Simpson, Monica Worline, Jane Dutton, and others goes beyond these frameworks to describe the processes and features that characterize compassionate organizations. These include, for example, compassionate leadership; communication networks that alert staff to the presence of pain or suffering; roles, routines, and social networks that facilitate alleviation of suffering; and the use of physical space to increase staff accessibility and interaction.

Methods. The convenings at the Task Force for Global Health confirmed that epidemiologic methods are well-suited to studying compassion at the institutional and population levels. Methods used in chronic disease epidemiology, in particular, are designed to handle the extraordinary complexity of multiple “risk factors” interacting across different levels and time points. Other epidemiologic methods, such as geospatial analysis, can be used to map where and when specific forms of suffering (such as hunger) occur and whether resources (such as food pantries) have been put in place to address that suffering.

Measures. Although a single ideal measure of compassion remains elusive, significant progress has been made in refining and triangulating measures of compassion, particularly in health care systems. For example, Shane Sinclair and his colleagues have developed companion measures of compassion that take the perspectives of both patients and health care providers. Taking a systems approach, Anupama Jha and her colleagues developed a suite of measures to evaluate the impact of compassion training and health systems transformation in district hospitals in Bihar, India. Beverley Fehr and Susan Sprecher recently applied an epidemiologic lens to synthesize two decades of psychological research that used the same existing measure—the Compassionate Love Scale—and provided fresh insight into the clustering of compassion by time, place, and person.

Looking to the future

Since Jinpa’s keynote address in 2018, we’ve seen a global outpouring of interest in compassion and how it moves through human systems. Along with this interest has come a desire for a scientific approach that can guide research and inform practical efforts to nurture compassion in institutions and systems. Epidemiology contributes a population-level lens to compassion science that can help reveal and quantitatively describe the interplay of factors that enable compassion to arise in specific settings. So, what is needed now for compassion science to realize its potential through epidemiology?

Broaden the conversation. As Jinpa noted, we must reach beyond our own community to broaden the conversation. We must include voices from other disciplines—as well as voices of those who do not align themselves with any academic discipline. Progress on the central challenge of measurement can come only from a thorough understanding of compassion in different cultures and settings, incorporating perspectives across the breadth of human experience. We must learn particularly from Indigenous communities and colleagues in the global South. We need the perspectives of evolutionary anthropologists, sociologists, primate biologists, artists, peacebuilders, spiritual leaders, and theologians. And the conversation should extend more deeply within epidemiology itself, for example, to explore how the tools of infectious disease modeling and social epidemiology can help us understand how compassion is transmitted or spread across populations. 

Conceptual frameworks. There is much to be learned from in-depth exploration and comparison of the new conceptual models of collective compassion. Broadening and deepening the conversation to refine, apply, align, and learn from the new conceptual models can lead to an integrated framework, a conceptual scaffolding for further research and testing of ideas. 

Data. Epidemiology is a data-driven science. Advances in conceptual frameworks inevitably lead to new questions and stimulate the search for data to answer those questions. There is much to be gained from epidemiologic analysis of existing data, as shown by Fehr and Sprecher, as well as from including compassion-related questions in new research on human flourishing, happiness, and well-being. For example, analysis of compassion-related questions in the Global Flourishing Study by Julia Nakamura and her colleagues provides intriguing insight into the spatial distribution of compassion at a global level.

Systems perspective. Compassion science has largely focused on the giving of compassion by individuals. The recent increase in scholarship and interest in Self-compassion and receiving compassion from others has opened a rich inquiry into how compassion “flows” across these dimensions, as well as the implications of such flow for well-being and Mental health. In unpacking their multi-level framework, James Kirby and his colleagues argue that the systems in which humans live and work not only facilitate or constrain the flow of compassion among individuals, but these systems also serve as “active agents that can give and receive compassion.” While they may not be capable of warm-heartedness, organizations “possess the primary driver of compassion: intention,” and they can be sensitive to suffering and committed to its alleviation. This radical approach opens a new landscape for compassion science, for which epidemiology, with its systems perspective, is well-suited. 

Influence on epidemiology. Historically, epidemiology has overwhelmingly focused on disease and threats to human health. Tyler VanderWeele and his colleagues recently argued that the tools of epidemiology also should be applied to promote human flourishing. An epidemiology of compassion opens the field to a broader “positive epidemiology,” similar to developments in positive psychology. By inviting narratives and first-person accounts of the experience of compassion, epidemiology will be stretched in new ways—not unlike the field of contemplative neuroscience.

Compassion and social justice. The systems perspective of epidemiology, with its attention to populations, leads to the unavoidable question, rarely explored in individual compassion training, whose suffering is to be alleviated? In prompting this question and providing an analytic framework to address and monitor it, epidemiology brings compassion science into much-needed conversation with ethics and social justice. 

We live in a world that seems increasingly polarized. The need for compassion at the institutional and societal levels has never been greater. Applying the epidemiologic toolkit and its systems perspective to compassion science will not be a panacea, but we and others believe that it can make important contributions, just as it has in global health. We look forward to working with others who share this view—and to engaging in conversation with those who don’t. And we are eager to see more writing, thinking, and noticing of the ways that compassion manifests in communities and organizations. These real-world observations and reflections, combined with robust epidemiologic methods, hold the potential to bring us to the next frontier in compassion science—and to answering Jinpa’s question about how to “scale up” compassion.