The Mind & Life Institute: Celebrating 35 Years
Insights Journey into the heart of contemplative science
Journey into the heart of contemplative science

The Science of Compassion What we know—and don’t know—about compassion has bold implications for how we respond to the needs of our time. By Jennifer Mascaro

Illustration by Sirin Thada

Who am I to study compassion? I am a compassion scientist, which feels a little like choosing to ingest a tiny bit of poison and its antidote every workday. When I stare at a blank page to write about the science of compassion, I feel paralyzed by the presumptuousness of the endeavor and the reminder that I am a compassion researcher who struggles with compassion.

After four decades of a life full of failures of compassion—from near misses rooted in good intentions but lacking in wisdom and skillfulness to behavior that was outright antithetical to compassion—how dare I write with any expertise? This daily reminder of my compassion failures is a tiny taste of venom that feels embarrassing and humbling. But therein also lies a partial antidote, for compassion science also provides the evidence-based wisdom for improvement. These failures of compassion motivate my career as a compassion scientist; they also make me an apt scholar.

Although the scientific study of compassion was relatively new when I entered graduate school in 2005, it has received increasing attention and rigor from many corners of the research world. And with good reason: there is growing evidence that compassion is the bedrock to addressing many of the critical problems of our time. For example, both receiving and giving compassion are core ingredients for emotional well-being among children and adolescents,1 and appear to lay the foundation for health and psychological well-being beyond.2 Compassionate approaches to medicine and public health influence vaccine hesitancy,3 reduce health disparities,4 improve patient-provider trust and patient adherence,5 reduce provider burnout,6 and are a key component to addressing chronic pain and addiction.7 Compassionate structures, organizations, and communities will be required to bolster resilience in the face of ongoing crises such as climate change8 and the aftermath of pandemics.9 As carbon is the essential framework for life on earth, compassion is the emotional framework for addressing life’s suffering.

What is compassion?

In any research endeavor, one must always begin with a careful definition of the object of empirical interest. However, in the scientific record, we see that definitions of compassion have been inconsistent across time and research fields. More challenging, compassion has often been conflated with very similar terms such as empathy, empathic care, sympathy, and altruism. Stephen Trzeciak and Anthony Mazzarelli, the authors of Compassionomics, describe their attempt to aggregate and synthesize research on compassion as a “Sherlock Holmes” approach that often involved ignoring terms altogether since one researcher’s empathy is another’s compassion. Here is the definition I use10 as a compassion scientist. Compassion involves: 1) an awareness of another’s suffering, 2) a benevolent emotional response to this suffering, and 3) a desire or motivation to help relieve that suffering.

Compassion involves: 1) an awareness of another’s suffering, 2) a benevolent emotional response to this suffering, and 3) a desire or motivation to help relieve that suffering.

Defined this way, compassion appears to be baked into us by selective pressures that have arisen throughout evolutionary history. The first evolutionary pressures arrived with the mammalian order when the survival of vulnerable offspring depended on whether their mother was motivated and capable of care, something biologists refer to with a delightful phrase: “the appetitive drive to nurture”.11 And so, brain and body systems were shaped by evolution to support and promote caring for distressed offspring, for example, retrieving, sheltering, and feeding a crying infant. This evolutionary seed of maternal compassion was then honed by other evolutionary pressures that arose during human evolution, such as our cooperative child-rearing strategy12 that relies on allomaternal (“other” than mother) caregiving, and our complex and long-lasting social groups in which group members survive and thrive on reciprocal altruism (costly behavior that benefits non-related others).

However, evolution also baked in another feature: subjectivity. Our propensity to experience compassion is influenced by a multitude of factors, such as our context, our experiences (both recent and life history), our physiological and psychological state, and interpersonal factors that shape whether we think someone is worthy of our compassion. With this more complex, multifaceted, and subjective nature of compassion, scholars have begun to identify other relevant dimensions beyond our simple definition above. For example, Roshi Joan Halifax13 has described important non-compassion qualities and skills that must be present for compassion to arise, such as executive or attentional control, equanimity, interoceptive awareness, ethical intent, insight, emotion regulation, and distress tolerance. Others point to factors such as a sense of common humanity that makes another’s suffering relevant and overcomes the numbing (or outright contempt) that occurs if we think of others as not me or mine. Scholars such as Paul Gilbert14 point to the importance of factors that help us overcome compassion’s many and potent inhibitors. His work emphasizes both the strength of our self-interest and the costliness of compassion, especially when it is effortful, risky, or socially fraught. This theoretical approach highlights crucial psychological resources for overcoming fears and inhibitors of compassion, such as self-efficacy, confidence, and optimism that one can reduce the suffering they encounter.

It is the multidimensional and context-bound nature of compassion that I find so scientifically enticing. It also has what scientists call face validity—it just feels right to me. The times when my compassion has not failed, when instead I was jolted from my hard-wired me-centered orientation into an orientation that prioritized another’s emotional and cognitive point of view, and was able to improve their stead, those times feel like an emergent and gestalt event. They feel “under the skin” and below the level of consciousness and almost as though from a muse. They also feel emotionally complex—deeply sad, scary, and at the same time intensely rewarding and joyful. Research on the experience of compassion15 aligns with my personal experience, and several studies highlight that compassion is often tinged with both positive and negative emotions,16 making it distinct from near cousins such as empathy or kindness.

Understanding and Measuring Compassion

The curious scientist wonders what happens in the brain and body during such an emergent and complex interpersonal event, how the social ecosystem shaped it, and how history and experience potentiated or primed it to arise. There is also a real-world reason that studying the multidimensional, dynamic, and context-dependent nature of compassion is scientifically important. Understanding such a complex interpersonal event can help us understand how compassion may be cultivated. Given the importance of compassion in so many domains, we need to be able to measure it. However, many factors make compassion tricky to assess.

Given the importance of compassion in so many domains, we need to be able to measure it. However, many factors make compassion tricky to assess.

Ideally, research designs measure the insides and outsides of compassion, linking the affective, cognitive, and motivational components of compassion (the insides) with compassionate behavior (the outsides). Additionally, methods would ideally assess compassion from multiple perspectives to mitigate the thorny problem of biases (e.g., social desirability bias) and errors (e.g., misremembering past emotions and behaviors, or errors in predicting how we will feel about future events). Assessing compassion from multiple perspectives also has the potential to reveal essential mismatches between perspectives. However, measuring all facets of compassion from multiple perspectives can be close to impossible.

Understandably, research methods in compassion science are often shaped most by expedience and pragmatism, and the most common method for assessing compassion is using self-report measures. Shane Sinclair has conducted ground-breaking research17 using both quantitative and qualitative self-report to describe and assess physician compassion from the patient’s perspective.18 In our work studying compassion in healthcare, we have used self-report measures to identify the blocks and inhibitors to compassion, such as having a hectic and chaotic atmosphere or experiencing incivility from patients, which lead to burnout and compassion fatigue.19

Other methods have been used to characterize and quantify physiological states associated with compassion. Neuroimaging and psychophysiological measurements have been vital methodological tools, as they can reveal subtleties of which many of us are not conscious. For example, neuroimaging studies have been used to investigate differences20 between compassion and its close cousin, empathy. Other research paradigms have used heart rate variability measures to characterize how the vagus nerve affects cardiac rhythms. This body of research has led to the notion that the experience and behavioral manifestation of compassion appear to rely on the parasympathetic nervous system to modulate the emotional response to suffering.21

Another set of methods for investigating compassion prioritizes ecological validity, or the similarity between what is being measured and what occurs in real life. For example, helping behavior can be assessed while participants are not aware that they are being observed. In our current research, we use ambulatory audio recording and linguistic assessments to identify the quantity and quality of language by which compassion can be conveyed. For example, in a study of hospital chaplains, we found that chaplains’ self-reported compassion was associated with an inclusive and other-oriented linguistic style (e.g., more social words, more we language, and less I language). What’s more, when chaplains used more of this type of language, patients reported lower levels of depression after the chaplain visited with them, even when we controlled for the patient’s level of distress before the chaplain’s visit. The natural follow-up question that we are now pursuing is, what factors cultivate this skillful compassion?

Jenny further explores what compassion is, and shares her research providing compassion training to hospital chaplains at Mind & Life’s 2020 Summer Research Institute.

Growing Compassion

First, I want to distinguish between state- versus trait-based compassion. Some factors, and some interventions like compassion training programs, grow compassionate states, a relatively short-lived momentary experience of compassion. At the same time, other factors/interventions grow compassion as a psychological trait, a relatively stable readiness or tendency to experience and act compassionately across time and in a broad range of circumstances. This is not to say that compassionate states and traits are unrelated. In fact, many approaches to cultivating compassion presume that repeatedly engendering compassionate states will gradually strengthen the corresponding trait.

Compassion can be induced or potentiated with primes that essentially nudge us into a state that is conducive to compassion, making compassion more likely to arise in the moment. Such state compassion is more likely when we are primed with a feeling of secure attachment, usually by reflecting on a relationship or a time when we felt safe and held in care. Similarly, inducing feelings of gratitude and appreciation, interdependence, and a sense of common humanity will make compassionate states more likely to arise in response to suffering. In contrast, priming self-interest (e.g., economic self-interest or efficiency) makes compassion less likely.22 To paint with a broad brush, augmenting any of the constituent conditions detailed above—mindful attention, executive control, emotion regulation, ethical intent, and self-efficacy—will make it more likely that state compassion will arise.

On the other hand, one of the most potent ways to grow trait compassion is to become a primary caregiver of a young child. Primary caregivers experience hormonal and neurobiological changes that promote attentional, emotional, and behavioral responsivity to offspring. These brain and body changes occur across the broad spectrum of possible caregivers, from biologically related mothers and fathers,23 to adoptive parents,24 to doting grandparents.25 Research studies have directly tied these brain and body changes with a wide array of compassionate responses, including compassionate language, affectionate touch, and a type of interpersonally oriented vocalization often called ‘motherese.’ Together, these findings support the evolutionary origins of compassion I mentioned earlier—becoming a caregiver changes our brain and body to facilitate trait compassion.

But do these brain and body changes that support parental compassion make us more compassionate overall? Research by Carsten De Dreu suggests that the answer is no. Oxytocin does not promote indiscriminate compassion and appears to have the opposite effect,26 increasing disparagement of and conflict with individuals we perceive as being in an out-group.27 Behavioral research is consistent, showing that parents exhibit enhanced bias and aggression28 when their parental caregiving motivations are primed. This set of findings makes sense from an evolutionary perspective—what better way for a mammalian mother to help her offspring thrive than to care for them at the expense of unrelated others? However, it has profoundly important implications for today’s society, and raises the question: How can we grow compassion in both breadth and depth, rather than only in depth for our in-group, even at the expense of others?

How can we grow compassion in both breadth and depth, rather than only in depth for our own in-group, even at the expense of others?

That is the question that many research teams are attempting to address using interventions designed to cultivate sustainable compassion. For more than 15 years, we have been studying a program called CBCT® (Cognitively-Based Compassion Training), a system of contemplative exercises and reflective practices designed to develop and expand compassion. CBCT begins with a foundational moment of nurturance practice to engender feelings of interpersonal safety and secure attachment. Next, the content builds iteratively, starting with basic attention-building and stabilizing practices (often referred to as mindfulness) and progressing to more analytical approaches in which CBCT participants contemplate the ways they interact with self and others. CBCT acts in part on the dual premises29 mentioned earlier—that humans are hard-wired with an ability to experience compassion, and that our compassion is highly context-dependent and contingent.

Ultimately, CBCT is designed to increase sustainable compassion to all, not just to those for whom or instances when compassion comes easily. For example, we have found that medical students who completed CBCT reported increases in compassion and reduced depression and loneliness.30 Importantly, the greatest increases in compassion were among those students who came into the study with high levels of depression, suggesting that compassion training may be of most benefit to people who are struggling.

Using the Science of Compassion to Respond to Real-world Challenges

Jennifer Goetz, Dacher Keltner, and Emiliana Simon-Thomas published their widely-influential scientific review of compassion over ten years ago, which began with the statement: “Compassion is controversial.”31 Their paper has now been referenced more than 2,000 times, the most recent citing papers on topics as diverse as climate change responses,32 corporate irresponsibility,33 and grappling with the conceptualization of self-compassion.34 If compassion is still controversial, the importance of it to scientific and scholarly inquiry has only increased. Here are some important areas of curiosity that continue to inspire and motivate our research.

While compassion training appears to be beneficial, many of our studies also identified barriers to engaging with CBCT. For example, we recruited healthy adults to go through a 12-week CBCT course. Notably, before training, we scanned their brains using functional MRI while completing a “Pain for Self and Others” task that alternated between administering somewhat painful electric shocks to their wrist and showing them video clips of other people receiving shocks. This approach gave us an understanding of participants’ baseline brain function before embarking on CBCT. Next, we looked at whether baseline brain responses to watching others in pain would predict the amount of time that participants later spent engaging with the compassion meditation practices. We found that activation of the anterior insula,35 an area of the brain thought to be important for empathy, in response to the Other pain task (but not to receiving pain themselves) was predictive of compassion meditation practice time. The tentative interpretation is that people who entered the study with a less compassionate orientation were less likely to practice compassion meditation. We found something very similar in our study with medical students—self-reported compassion at the outset of the study was associated with ensuing CBCT practice time.

The above findings remind us that no intervention or training, however efficacious, will be successful if people are not able or willing to engage with it. It isn’t surprising that there are barriers to engaging with CBCT, nor is such a challenge unique to compassion training. One widely cited study found that it takes, on average, 17 years36 for an evidence-based intervention or practice to be successfully incorporated into the general practices of a health care setting. Our most recent research is aimed at identifying and examining new points of “entry” for compassion and compassion training that overcome barriers to access. With this line of research, we ask, how can we get compassion to people when they are most in need, and can we do it in feasible and acceptable ways? Paul Condon and John Makransky37 recently wrote a compelling argument that there is a rich historical precedent to skillfully adapt compassion training to “effectively [meet] the hearts and minds of people of new times and places.”

We are attempting this work via a partnership between contemplative science and spiritual health called Compassion-Centered Spiritual Health (CCSH™). CCSH interventions are delivered by healthcare chaplains, embedded and highly skilled members of the interprofessional healthcare team. To evaluate CCSH, we use implementation science,38 a trans-disciplinary form of scientific study that applies rigorous and consistent methods for assessing the implementation of a practice or intervention in a new setting. As we strive to integrate compassion into our societal structures, the careful work of implementation science will be critical. Using these methods, we can identify the barriers that may obstruct or hinder compassion training. We can also identify individual, interpersonal, systemic, and cultural factors that facilitate compassion.

The scientific methods for studying compassion—from self-report questionnaires, to fMRI, to heart rate variability, to behavioral observation—have advanced the field and generated new knowledge about the nature of compassion. However, these methods do not always lend themselves to characterizing wise and skillful compassion. As we integrate complexity into our understanding of compassion, we see more gaps in our existing scientific knowledge. In fact, we understand relatively little about the dynamic and relational emotional, cognitive, and embodied phenomena that unfold when we encounter suffering in our daily lives, outside the contrived parameters of a laboratory. To put it another way, observing brain responses to an emotional picture using fMRI may bear little resemblance to observing how an emergency nurse responds to an angry and fearful patient. And neither of these characterizations of compassion may bear a resemblance to observing how a compassionate community supports well-being or engages in restorative justice.

Alternatively, perhaps scientists are describing and quantifying a similar construct of compassion—from the fMRI to the community—but at differing units of analysis and filtered through different organizing principles. If so, then the knowledge we have gained “in the laboratory” about the neural, psychological, and behavioral inhibitors and facilitators of compassion can help us move toward solving the problems that exacerbate life’s deepest suffering. Ultimately, we will need to continue to refine and innovate our scientific methods to characterize this complexity and to leverage the science of compassion to respond to the needs of our time.

…when compassion becomes a social norm, the result is an upward spiral of interpersonal trust and cooperation.

In closing, I find that the scientific study of compassion is a spectacular excuse to contemplate compassionate exemplars—real-life bodhisattvas of compassion whose stories replenish the compassion well. And the recently released World Happiness Report tells us that this well is more, not less, full. In 2021, compared to the several years prior, the worldwide prevalence of charitable donations, volunteering, and rates of helping strangers increased. Instances of prosociality increased in every global region and were an astonishing 25% higher than pre-pandemic levels. This report is heartening beyond just its immediate implication because longitudinal research tells us that when compassion becomes a social norm, the result is an upward spiral of interpersonal trust and cooperation. Compassion begets trust, which begets more compassion. Bringing the science of compassion into conversation with wise leadership and healing-centered efforts can help facilitate such an upward spiral.

Have something to share? Email us at
Support our work with a gift to Mind & Life. 


  1. 1.

    Bach, J. M., & Guse, T. (2015). The effect of contemplation and meditation on ‘great compassion’ on the psychological well-being of adolescents. The Journal of Positive Psychology, 10(4), 359-369.

  2. 2.

    Roeser, R. W., & Eccles, J. S. (2015). Mindfulness and compassion in human development: Introduction to the special section. Developmental Psychology, 51(1), 1.

  3. 3.

    Mical, R., Martin-Velez, J., Blackstone, T., & Derouin, A. (2021). Vaccine hesitancy in rural pediatric primary care. Journal of Pediatric Health Care, 35(1), 16-22. [Full article]

  4. 4.

    Bruce, M. A., Skrine Jeffers, K., King Robinson, J., & Norris, K. C. (2018). Contemplative practices: A strategy to improve health and reduce disparities. International Journal of Environmental Research and Public Health, 15(10), 2253. [Free full text]

  5. 5.

    Post, S. G. (2011). Compassionate care enhancement: Benefits and outcomes. International Journal of Person Centered Medicine, 1(4), 808-813.

  6. 6.

    Zhang, Y. Y., Zhang, C., Han, X. R., Li, W., & Wang, Y. L. (2018). Determinants of compassion satisfaction, compassion fatigue and burn out in nursing: A correlative meta-analysis. Medicine, 97(26).

  7. 7.

    Vitzthum, L. K., Riviere, P., & Murphy, J. D. (2020). Managing cancer pain during the opioid epidemic—balancing caution and compassion. JAMA Oncology, 6(7), 1103-1104.

  8. 8.

    Gostin, L. O. (2017). Hunger, Health, and Compassion. JAMA, 317(19), 1939-1940.

  9. 9.

    Slavich, G. M., Roos, L. G., & Zaki, J. (2022). Social belonging, compassion, and kindness: Key ingredients for fostering resilience, recovery, and growth from the COVID-19 pandemic. Anxiety, Stress, & Coping, 35(1), 1-8.

  10. 10.

    Strauss, C., Taylor, B. L., Gu, J., Kuyken, W., Baer, R., Jones, F., & Cavanagh, K. (2016). What is compassion and how can we measure it? A review of definitions and measures. Clinical Psychology Review, 47, 15-27.

  11. 11.

    Numan, M., & Young, L. J. (2016). Neural mechanisms of mother–infant bonding and pair bonding: Similarities, differences, and broader implications. Hormones and Behavior, 77, 98-112. [Full article]

  12. 12.

    Hrdy, S. B. (2009). Mothers and Others: The Evolutionary Origins of Mutual Understanding. Harvard University Press.

  13. 13.

    Halifax, J. (2012). A heuristic model of enactive compassion. Current Opinion in Supportive and Palliative Care, 6(2), 228-235.

  14. 14.

    Gilbert, P. (2021). Creating a compassionate world: Addressing the conflicts between sharing and caring versus controlling and holding evolved strategies. Frontiers in Psychology, 11, 582090.

  15. 15.

    Condon, P., & Feldman Barrett, L. (2013). Conceptualizing and experiencing compassion. Emotion, 13(5), 817. [Full article]

  16. 16.

    Ashar, Y. K., Andrews-Hanna, J. R., Dimidjian, S., & Wager, T. D. (2017). Empathic care and distress: Predictive brain markers and dissociable brain systems. Neuron, 94(6), 1263-1273.

  17. 17.

    Sinclair, S., Hack, T. F., McClement, S., Raffin-Bouchal, S., Chochinov, H. M., & Hagen, N. A. (2020). Healthcare providers perspectives on compassion training: A grounded theory study. BMC Medical Education, 20(1), 1-13.

  18. 18.

    Sinclair, S., Beamer, K., Hack, T. F., McClement, S., Raffin Bouchal, S., Chochinov, H. M., & Hagen, N. A. (2017). Sympathy, empathy, and compassion: A grounded theory study of palliative care patients’ understandings, experiences, and preferences. Palliative Medicine, 31(5), 437-447.

  19. 19.

    Ash, M. J., Baer, W., Peacock, C., Haardörfer, R., & Mascaro, J. S. (2022). Advancing measurement of the sources and consequences of burnout in a comprehensive cancer center: A structural equation modeling analysis. American Journal of Medical Quality, 37(2), 95-102.

  20. 20.

    Singer T, Klimecki OM. Empathy and compassion. Current Biology. 2014;24(18):R875-R878.

  21. 21.

    Di Bello, M., Carnevali, L., Petrocchi, N., Thayer, J. F., Gilbert, P., & Ottaviani, C. (2020). The compassionate vagus: A meta-analysis on the connection between compassion and heart rate variability. Neuroscience & Biobehavioral Reviews, 116, 21-30.

  22. 22.

    Di Bello, M., Carnevali, L., Petrocchi, N., Thayer, J. F., Gilbert, P., & Ottaviani, C. (2020). The compassionate vagus: A meta-analysis on the connection between compassion and heart rate variability. Neuroscience & Biobehavioral Reviews, 116, 21-30.

  23. 23.

    Abraham, E., & Feldman, R. (2022). The neural basis of human fatherhood: A unique biocultural perspective on plasticity of brain and behavior. Clinical Child and Family Psychology Review, 1-17.

  24. 24.

    Glasper, E. R., Kenkel, W. M., Bick, J., & Rilling, J. K. (2019). More than just mothers: The neurobiological and neuroendocrine underpinnings of allomaternal caregiving. Frontiers in Neuroendocrinology, 53, 100741.

  25. 25.

    Rilling, J. K., Gonzalez, A., & Lee, M. (2021). The neural correlates of grandmaternal caregiving. Proceedings of the Royal Society B, 288(1963), 20211997.

  26. 26.

    De Dreu, C. K., Greer, L. L., Handgraaf, M. J., Shalvi, S., Van Kleef, G. A., Baas, M., … & Feith, S. W. (2010). The neuropeptide oxytocin regulates parochial altruism in intergroup conflict among humans. Science, 328(5984), 1408-1411.

  27. 27.

    Triki, Z., Daughters, K., & De Dreu, C. K. (2022). Oxytocin has ‘tend-and-defend’ functionality in group conflict across social vertebrates. Philosophical Transactions of the Royal Society B, 377(1851), 20210137.

  28. 28.

    Gilead, M., & Liberman, N. (2014). We take care of our own: Caregiving salience increases out-group bias in response to out-group threat. Psychological Science, 25(7), 1380-1387.

  29. 29.

    Ash, M., Harrison, T., Pinto, M., DiClemente, R., & Negi, L. T. (2021). A model for cognitively-based compassion training: Theoretical underpinnings and proposed mechanisms. Social Theory & Health, 19(1), 43-67.

  30. 30.

    Mascaro, J. S., Kelley, S., Darcher, A., Negi, L. T., Worthman, C., Miller, A., & Raison, C. (2018). Meditation buffers medical student compassion from the deleterious effects of depression. The Journal of Positive Psychology, 13(2), 133-142.

  31. 31.

    Goetz, J. L., Keltner, D., & Simon-Thomas, E. (2010). Compassion: An evolutionary analysis and empirical review. Psychological Bulletin, 136(3), 351.

  32. 32.

    Davidson, D. J., & Kecinski, M. (2022). Emotional pathways to climate change responses. Wiley Interdisciplinary Reviews: Climate Change, 13(2), e751.

  33. 33.

    Valor, C., Antonetti, P., & Zasuwa, G. (2022). Corporate social irresponsibility and consumer punishment: A systematic review and research agenda. Journal of Business Research, 144, 1218-1233.

  34. 34.

    Dunne, J. D., & Manheim, J. (2022). Compassion, self-compassion, and skill in means: A Mahāyāna perspective. Mindfulness, 1-9.

  35. 35.

    Mascaro, J. S., Rilling, J. K., Negi, L. T., & Raison, C. L. (2013). Pre-existing brain function predicts subsequent practice of mindfulness and compassion meditation. NeuroImage, 69, 35-42. [Full article]

  36. 36.

    Morris, Z. S., Wooding, S., & Grant, J. (2011). The answer is 17 years, what is the question: Understanding time lags in translational research. Journal of the Royal Society of Medicine, 104(12), 510-520.

  37. 37.

    Condon, P., & Makransky, J. (2022). Compassion and skillful means: Cultural adaptation, psychological science, and creative responsiveness. Mindfulness, 1-11. [Free preprint]

  38. 38.

    Bauer, M. S., Damschroder, L., Hagedorn, H., Smith, J., & Kilbourne, A. M. (2015). An introduction to implementation science for the non-specialist. BMC Psychology, 3(1), 1-12.