On Ignorance and Public Health.

In a recent opinion piece, Jamie Holmes argued that we should be teaching ignorance. The study of ignorance, agnotology, is growing, and Routledge recently published an agnotology handbook, though books in the area have been available for decades. At core, the study of ignorance rests at the edges of knowledge. Professor Michael Smithson, who has long written on the topic, uses the very helpful metaphor of a “knowledge island.” The more we know, the bigger the island, and the bigger the island, the longer the island’s shores—i.e the more ignorance we have, the less we know. We are most comfortable at the center of the island of knowledge, where we know what we know. But how do we grapple with the uncomfortable fact that we may be spending much of our time on the island’s beaches, where the limits of our knowledge lie?

I found this thinking helpful to a consideration of how we, in population health scholarship, go about doing our work. We aspire to improve the health of populations and to produce science and scholarship that help us understand how we may do so ever better. This rests on the production of knowledge—on getting ourselves to the center of the island so that we may then think of actions we can take to promote population health. Hence, a large body of scholarship convinces us that smoking is associated with a variety of pathologies, and we then agitate for the introduction of measures to curb smoking, hence improving population health. Population health science evidence demonstrates that driving while under the influence of alcohol is associated with substantial vehicular morbidity and mortality, and this results in the development of a plan of action and the subsequent implementation of that plan, resulting in a decrease in drunk driving and its consequences across the country.

But what about cases where the causes are not so clear—where we are shrouded in ignorance, walking on the shores of the islands of knowledge? What do we do when we do not know what the right answer is? This is far from uncommon and, in fact, I would suggest is probably the case much more in our area of scholarship than we are willing to admit.

Let us take one particular contentious issue at the moment: e-cigarettes. The advent of electronic cigarettes has occasioned engaged and increasing discussion in the public health literature for nearly a decade, including a body of work by Michael Siegel and other faculty at the School. The central question about the issue is whether electronic cigarettes—well accepted to be less harmful than regular cigarettes—are a “slippery slope,” becoming a “gateway” to the use of regular cigarettes, or re-normalizing cigarettes and potentially chipping away at decades of hard fought gains against smoking rates in general. The other side of the story may be that e-cigarettes may be a useful harm reduction approach, providing a new tool in the public health armamentarium to decrease cigarette smoking in populations. Over the past months, two reports have been published, with conflicting findings. A study from Los Angeles found that adolescents who used e-cigarettes were more likely than non-users to report subsequent initiation of combustible tobacco. By contrast, a UK government independent review found no evidence that e-cigarettes represent a path to smoking for non-smokers. Several commentaries have dissected the evidence, some concluding that the threats of e-cigarettes are real and have urged action; others have been more nuanced.

It is not my purpose here to review the e-cigarette literature, leaving that to others who are experts in the field, but I raise it as a particularly interesting area of contention in public health at the moment. Clearly the e-cigarette industry is big business, and this introduces a new element into an area we thought we had a good grip on. But, just as clearly, the data are not at all conclusive, and relatively sober analysts are, on reviewing the data, coming to different conclusions. I have published on another such area of public health contention—the reduction of salt in populations—and have commented, in a previous Dean’s Note, that our uncritical acceptance of hypotheses that may seem compelling, but that dwell on contentious evidence, stands us poorly in public health.

So, coming back to this central topic, what are we to do with ignorance? What are we to do with uncertainty and ambiguity? At a core level, we have, as members of an academic community, the extraordinary privilege of luxuriating in the expansion of the shores of the knowledge island. The fact that there may be ambiguity about e-cigarettes or population salt reduction can be cause for celebration. These are interesting questions to address, interesting studies to conduct, interesting answers to discover. What could be more engaging for those of us charged with generating knowledge as our core, foundational identity? But, it is the second part of our self-definition that compels us to engage these questions with more than academic interest. We wish to understand the causes of population health—to generate knowledge so that we can transmit that knowledge to those who are in a position to improve the health of populations. And ignorance and ambiguity may be far less useful to transmit to those who indeed control the levers that produce health. To add further tension to this, our cause is so pressing—the health of many—that questions cry for rapid answers. If e-cigarettes are emerging quickly, surely we should know quickly what effect they might have so that we may act with similar alacrity to save lives?

But should we?

I would suggest five thoughts that might serve us well to deal with ignorance and ambiguity in population health scholarship.

First, we simply need to recognize that ignorance is an inevitable part of what we do. Not knowing is not anomalous; it is, in many ways, normative. An embrace of ignorance can help recalibrate our expectations. We should expect not to know if e-cigarettes can help or hurt our progress against combustible tobacco. We should accept that a substantial proportion of cancers are stochastically determined. And we should recognize that this ambiguity is inevitable when dealing with complex, multi-scale processes that evolve over time. This approach frees us to be clear when we do not know, and to articulate more clearly when it is that we might want to act. It also suggests that we should be preparing our students to deal with ambiguity, to be comfortable in the grays, and to understand that public health remains a pragmatic discipline even in the face of such uncertainty.

Second, stemming from the first, we need to recognize that public health action does not necessarily need to follow causal certainty, and that we can, and not infrequently should, act even when we do not know all the answers. Action can be predicated on values, on analogy, on anticipation of future trends, or on causal links built on a hard-won understanding of the likely consequences of omission of action. However, if we do act, we should not do so on false data pretexts, invoking knowledge and clarity when in fact ignorance of causal effect and ambiguity are the order of the day. This motivates us to find greater clarity about when and why we should act, and to do so with transparent honesty about the attendant uncertainties, allowing us the option of changing course when some of our ignorance falls away.

Third, we must recognize that false certitude is not without consequence. The obvious unintended outcome of false certitude is taking action that turns out to be inadvertently harmful. The use of hormone replacement therapies by women across different age groups and their potential role in increasing heart disease is a classic example of policy recommendation built on an imperfect understanding of causal processes. But other costs of false certitude are, even if less direct, just as important, including loss of trust in the profession and misdirection of resources from areas of scholarship and action that could be promising, as we put all our eggs in the wrong basket.

Fourth, we must carefully separate genuine ignorance from manufactured ambiguity. There is ample precedent of particular special interests working to generate doubt about the state of the science when, in fact, the scientific facts were in and clear. The classic example of this involves industry efforts to repudiate findings about lead poisoning. Sometimes we do know the answer. We do know that combustible tobacco is deadly. We do know that seat belts save lives. In these cases, the role of public health is easier. Here we must muster as much creativity as possible to make sure conditions are such—no tobacco smoking, universal use of seatbelts—that they promote the health of the public, even if special interests (perhaps with commercial motivations) are aligned against these efforts.

Fifth, we need to remember that our capacity for individual prediction remains tremendously limited. Most of our data that aims to understand the production of health comes from population-level data, making fraught individual prediction extensions from these data. I have written on this in a previous Dean’s Note and in peer-reviewed publications, but note it here again as an important element of the limitations of our knowledge, with particular relevance for clinicians and for the utility of individualized genomic approaches to health prediction.

These five thoughts embed many others, some which I will engage in future Dean’s Notes. Some ideas that emerge from them, however, remain only sketchily formed in my mind, trained as I am to try to look for clarity and certitude through my own scholarship. Perhaps a training of our emerging scholars that emphasizes the expectation that they will face ever expanding shores of ignorance, and that much of their careers will involve action in the face of uncertainty, will help a new generation see the role of public health differently, pushing us to grapple much more clearly with the role, constraints, and possibilities of public health action in the face of ambiguity.

I hope everyone has a terrific week. Until next week.

Warm regards,

Sandro

Sandro Galea, MD, DrPH
Dean and Professor, Boston University School of Public Health
Twitter: @sandrogalea

Acknowledgement: I am grateful to professor Leonard Glantz for alerting me to my ignorance about agnotology.

Previous Dean’s Notes are archived at: https://www.bu.edu/sph/category/news/deans-notes/