Applying the methods of chronic disease epidemiology to the study of compassion

Abstract

Now, more than ever, the world needs compassion—at the level of the individual, the community, the workplace, and society as a whole. As an abstract concept, the idea that increasing compassion will benefit society on many levels is simple. But as a target for public health intervention, our fundamental understanding of compassion is poor. How can we craft interventions to increase our willingness and capacity to help others through adversity and suffering when our understanding of the factors that increase compassion, the barriers to compassion, and their interplay, is limited? Work by a wide range of social and behavioral scientists has laid the foundations for compassion studies by establishing consensus definitions and identifying recurring themes in compassionate interactions. We believe that applying epidemiologic methods to study the determinants of compassion will provide a complementary, quantitative foundation for future initiatives, interventions, and benchmarking. In this paper, we explore how compassion has many of the same features as chronic diseases, including multiple dimensions, time scales, and degrees of modifiability, evolution over time, non-linear progression, and complex, interacting influences. We weigh the suitability of various epidemiologic study designs for studying compassion. We then propose a number of analytic approaches from chronic disease epidemiology that could be used to build quantitative, and potentially even causal, models of compassion. We examine the advantages that would be gained by using methods such as multi-level modeling, sequential conditional mean modeling, marginal structural modeling, and machine learning to study compassion and its influences, along with the potential limitations of each method. Finally, we round out the paper with a discussion of the challenges involved in using epidemiologic frameworks to study compassion.