The Role of Natural Experiments in Advancing Public Health Science and Practice

Public health has expansive ambitions. From the preamble to the World Health Organization’s constitution, which defines health as a “state of complete physical, mental and social well-being and not merely the absence of disease or infirmity,”¹ to recent discussions about the aspirations of public health that grapple with the social and economic forces that shape health,^2,3 public health aspires to create a world that generates health and the opportunity for all to flourish and reach their potential. This is an important vision of what public health can do. It also, however, poses operational challenges. Although, for example, a health in all policies approach has long been at the core of public health thinking,⁴ it remains far from the daily reality for most public health practice.

IMPLEMENTING THE VISION OF PUBLIC HEALTH

There are many reasons that implementing efforts that align with this vision of public health is challenging. Although we can envision that getting to a better, healthier world will require orienting our systems—for example, transportation, finance, and urban planning—to generate health, it is often harder to fully grasp how to get there. The vision is ambitious, and it is much easier to succeed in a narrowly prescribed set of actions than in a set of actions that requires whole system change. We see the challenges as twofold: (1) we have a paucity of evidence on which systemic interventions will help generate health, and (2) there is a lack of evidence for how to implement and encourage the widespread adoption of such interventions. Although there is a growing body of evidence for such efforts,⁵ it remains far narrower than, for example, the literature on behavioral interventions that public health has deployed much longer and more liberally.

The evidence base on how to implement transformative change that aligns social and economic systems to generate health is more limited for many reasons. For one, generating this evidence requires mechanisms that reward and fund such scholarship, and it has been amply demonstrated that funding for such approaches lags behind funding for biomedically or individually oriented approaches.⁶ However, it is also true that it is harder to launch efforts that experimentally test whether social and environmental changes improve health than it is to test more focused, individually centered interventions. Changing whole systems takes longer, is more expensive, may be unable to control for all possible alternative hypotheses, and may introduce more ethical challenges than do efforts in a narrowly defined scope. It is not surprising, then, that we have less evidence than we should to implement the broader agenda of public health. This lack of evidence, in somewhat of a vicious circle, results in less investment in the studies that could inform an ambitious and more robust public health practice agenda.

THE ROLE OF NATURAL EXPERIMENTS

This is why natural experiments have tremendous potential to inform public health scholarship, insight into health equity, and public health practice. Natural experiments—studies that take advantage of conditions that are imposed on populations outside the investigators’ control—create opportunities for us to study what happens when particular large-scale changes happen. Natural experiments can help us build an evidence base that provides the impetus for implementing systemic change to create health equity and improve health. Two studies in this month’s issue of AJPH illustrate the potential of natural experiments well.

First, Komisarow and Pakhtigian (p. 881) wanted to investigate whether closing coal-fired power plants reduces asthma-related conditions among children. There is good evidence that children exposed to air pollution have worse respiratory systems.⁷ However, that evidence, borne out through a range of observational studies, has limited potential to influence policy that is specifically about coal-fired power plants related to health improvement.

Komisarow and Pakhtigian took advantage of the 2012 closure of three large coal-fired power plants near Chicago, Illinois. Using wind data, population data, particulate matter 2.5 definitions, and zip code–level rates of emergency department asthma visits for children younger than four years between 2009 and 2017, and applying difference-in-difference methods—a useful analytic approach for such studies—the authors showed that asthma hospitalizations decreased by 12% in zip codes close to the plants that were closed compared with zip codes that were farther away. This is an elegant illustration of the potential of systemic intervention. We know that air pollution is associated with worse respiratory symptoms, but now we also know that specific action to close the sources of air pollution—in this case coal-fired power plants—can improve health. Public health can use these important data to contribute to broader discussions that weigh the costs and benefits of such closures.

The second article in this issue of AJPH of note on this topic is that of Kessler et al. (p. 927), who capitalized on the development of a family health strategy in Bagé City, Brazil, that aimed to reorganize and restructure the health system and, in doing so, strengthen primary care. Although this was a health intervention, the authors used the intervention to assess an outcome that was well beyond the initial intent—socioeconomic gaps in health. The authors correctly note that it is not yet known whether health programs and policies on reducing health inequalities are effective. While seeking to clarify this question, they found that family health strategy implementation reduced social inequalities in both all-cause and avoidable mortality among older adults, suggesting that family health strategy implementation may help achieve better health for all in a rapidly growing economy marked by substantial inequality, such as Brazil.

LOOKING AHEAD

What brings the Komisarow and Pakhtigian and the Kessler et al. articles together is their building on an opportunity presented by a change that was out of their control—a natural experiment—to assess what import this change had for a desirable public health outcome. In showing that the changes studied did indeed promote health and health equity, the authors point the way to advocacy efforts that can extend their work, consistent with the mission of public health. Other articles in this issue of AJPH employ empirical methods to study the natural experiment that has perhaps occupied most of our attention over the past year: COVID-19. We hope these studies will begin to illuminate how policies have affected transmission rates and vaccine rollout. We look forward to discussing the evolving insights that will emerge from the study of COVID-19.