Verified by 
I recently read with interest a piece by Michael Schulson in Undark which compared and contrasted the methodological norms of evidence-based medicine (EBM) with those of public health. The piece, which I encourage everyone to read in full, sketches areas of tension between some in public health and in the EBM space, focusing on how these tensions shaped the pandemic moment and our subsequent reflections on that time. The piece was interesting on several fronts, offering insights into the limitations of data science in guiding the work of public health.
One element I thought was implied, but not explicitly stated, in Schulson’s piece was the role that moral argumentation plays in advancing the work of public health. We aspire to create and implement effective public health interventions because we are motivated by a moral imperative to shape a healthier world. We get closer to such a world by generating data that help guide our actions and policies and then by making the moral argument in favor of these steps, toward building a mass movement for health. I would like to use today’s post, prodded by Schulson, to talk about evidence in public health—to discuss the data that are at the heart of all we do, without which moral argument can only ring hollow, and how we can continue to maintain the essential role of science in our work.
How best to think about the data we use in population health science, to the end of these data informing public health? I suppose throughout my career I have alighted on three dimensions that seem worth summarizing here (and, as I do, I find I am somewhat surprised that I have not summarized them in this way previously). They are: what questions should our science pursue, what characteristics should our science have, and what qualities should we look for in our data?
First, what questions should we pursue in our science? Science is, fundamentally, an arena for inquiry, and it takes discipline and focus to determine what we should ask, and what answers we should devote our time and resources to seeking. I have long believed we should be in the business of asking questions that matter about issues of consequence for health. I first argued for consequentialism as an animating principle for our work more than a decade ago, intending this to mean that given the foundational role that population health science plays in public health, we should tackle issues that matter. That means that we should address issues that shape the health of many populations, pursue science that is about ubiquitous forces, and tackle issues where we can provide the data to inform contemporary questions of consequence. Certainly, there is a place for scientific inquiry that is about a more free-form mode of exploration, with immediate real-world applicability a secondary concern. Yet in a world of limited time and resources, and with many pressing threats to health, it sems to me that our pursuit of health should aspire, always, to be maximally consequential.
This leads to the second question: what characteristics should our science have? Population health science should be rigorous—which is to say, it should be good science. This means science that produces quality data, and it means science that does good in the world. Sometimes, in this current moment of ready rancor and easy takedowns, we forget that at the heart of all we do should be science done with the capacious aspiration to be good. Even as we sometimes fall short of this goal, we should simply be able to recognize when we fall short and commit to doing better. For our science to do good in the world, it must rest on a foundation of solid methodology and an unflinching willingness to follow data wherever it leads. Sometimes our findings will confirm our preconceived views of the world and sometimes they will not. No matter what the data tell us, we should have the integrity to face facts and engage with what our science seems to be saying. This extends to the methodologies we adopt, such as, for example, randomized controlled trials. I have written recently with Michael Stein that we should not be in the habit of treating randomized clinical trials as the end-all-be-all of health sciences. We should respect what they can teach us, while maintaining the skeptical posture that is so necessary for a science that is able to grow and evolve in the face of changing circumstances. We should never adopt methodologies simply because they are easy or tell us what we want to hear, but rather we should focus on what is the best tool for the job in front of us, and then, having used that tool, engage honestly with the results.
Finally, what qualities should we look for in our data? First and foremost, our data need to be inquisitive. I mean here, to quote Merriam-Webster, “given to examination or investigation” and “inclined to ask questions,” and I acknowledge a debt of gratitude to an old statement of values from The Milbank Memorial Fund that surfaced this idea. We should not launch our inquiries with ready-made answers in mind, but, rather, with an openness to new information and, potentially, to unexpected conclusions. This means cultivating a vision of our science that engages not just with where we think health is today, but with where it might be tomorrow. We need to be asking questions in our science that may be different from what others are asking, challenging ourselves and our field to embrace a forward-looking vision for our science, and, through our science, our work as activists pursuing a healthier world. Because, of course, our work should be forward looking, given our interest in creating a healthier tomorrow.
Thinking “outside of the box” in this way can be, admittedly, difficult. Internally, there is the challenge within our field of groupthink, and the self-reinforcing tendency to stay within the lanes of what we regard as acceptable opinion. Externally, our science is subject to a range of outside influences—from funders to voices in the media—whose attitudes towards what we do are not easily dismissed. This makes it challenging to advance ideas within science which may be seen as idiosyncratic. Yet advancing such ideas is often precisely what makes for science that matters—that moves the ball forward on what is most consequential for health.
A version of this post also appears on Substack.
