Happy Nice Mall

The last 20 months have turned every dog ​​into an amateur epidemiologist and statistician. Meanwhile, a group of well-meaning epidemiologists and statisticians began to believe that pandemic problems could be solved more effectively by adopting an engineer’s mindset: that is, focusing on pragmatic problem solving with an iterative, adaptive strategy to get things done.

In a recent article, “Accounting for uncertainty during a pandemic“Researchers ponder their role during a public health emergency and how they can better prepare for the next crisis. Their answer may lie in reimagining epidemiology more from an engineering standpoint and less from a “pure science” standpoint.

Epidemiological research informs public health policy and its naturally practiced duty of prevention and protection. But the right balance between pure research results and pragmatic solutions has proven alarmingly difficult during the pandemic.

We must make practical decisions, so how important is uncertainty really?

Seth Guikema

“I always imagined epidemiologists would be helpful people in an emergency like this,” says Jon Zelner, co-author of the paper. “But our role is more complex and more ill-defined than I expected at the start of the pandemic.” Zelner, an infectious disease modeler and social epidemiologist at the University of Michigan, has witnessed a “frantic proliferation” of research papers, “many of whom had very little thought in terms of what it really means in terms of having a positive impact.”

There were “many missed opportunities,” says Zelner, because of the missing links between the ideas and tools epidemiologists propose and the world they were meant to help.

give up certainty

Co-author Andrew Gelman, a statistician and political scientist at Columbia University, put the “big picture” in the paper’s introduction. He likened the pandemic’s explosion of amateur epidemiologists to the way the war turned every citizen into an amateur geographer and tactician: “Instead of maps with colored pins, we have exposure and death count charts; “People on the street are arguing about infection mortality and herd immunity in a way they might have discussed war strategies and alliances in the past.”

And with all the data and public discourse—Are the masks still necessary? How long will vaccine protection last?—the rain of uncertainty has come.

Trying to figure out what just happened and what went wrong, the researchers (which included Ruth Etzioni of the University of Washington and Julien Riou of the University of Bern) performed a kind of reenactment. They examined tools used to overcome challenges such as estimating the person-to-person transmission rate and the number of cases circulating in a population at any given time. They evaluated everything from data collection (the quality and interpretation of data were arguably the pandemic’s biggest challenges), from model design to statistical analysis, as well as communication, decision-making, and trust. “There is uncertainty at every step,” they wrote.

Still, Gelman says, the analysis “doesn’t adequately express the confusion I experienced in those early months.”

A tactic against all uncertainty is statistics. Gelman considers statistics to be “mathematical engineering”—methods and tools that are as much about measurement as discovery. Statistical science tries to illuminate what is going on in the world with a spotlight on diversity and uncertainty. As new evidence arrives, it must establish an iterative process that gradually refines prior knowledge and sharpens certainty.

Good science is humble and has the ability to heal itself in the face of uncertainty.

Marc Lipsitch

Susan Holmes, a statistician at Stanford who was not involved in this research, also sees parallels with the engineering mindset. “An engineer always updates his picture,” he says, and revises it as new data and tools become available. When dealing with a problem, an engineer will use a first-order approach (fuzzy), followed by a second-order approach (more focused), etc. presents.

However, Gelman forewarned that statistical science can be deployed as a machine to “clear obscurity”—intentionally or not, crappy (vague) data is brought together and made convincingly (precise) visible. Statistics against uncertainties are “sold as a kind of alchemy to turn these uncertainties into certainty.”

We witnessed this during the pandemic process. Drowning in turmoil and unknowns, epidemiologists and statisticians – amateurs and experts alike – were on the hunt for something solid as they struggled to stay afloat. But as Gelman points out, asking for certainty during a pandemic is inappropriate and unrealistic. “Early certainty has been part of the challenge of decisions in the pandemic,” he says. “This jump between uncertainty and certainty has caused a lot of problems.”

Letting go of the desire for certainty can be liberating, she says. And this is, in part, where the engineering perspective comes into play.

a cunning mindset

For Seth Guikema, co-director of the Center for Risk Analysis and Informed Decision Engineering at the University of Michigan (and Zelner’s collaborator on other projects), an important aspect of the engineering approach is diving into uncertainty, analyzing complexity, and then, “We have to make practical decisions, then How important is uncertainty really?” taking a step back in perspective. Because if there’s a lot of uncertainty — and uncertainty changes what are optimal decisions, or even what are good decisions — then it’s important to know that, says Guikema. “But if it doesn’t really affect my best decisions, then it’s less critical.”

For example, increasing SARS-CoV-2 vaccination coverage across the population is a scenario where, while there is some uncertainty about exactly how many cases or deaths vaccination will prevent, it is highly likely to reduce both and with little downside. effects are sufficient motivation to decide that a large-scale vaccination program is a good idea.

An engineer always updates his picture.

Susan Holmes

Holmes points out that engineers are also very good at breaking down problems into critical parts, applying carefully selected tools, and optimizing solutions under constraints. With a team of engineers building a bridge, there is a cement specialist, a steel specialist, a wind engineer and a structural engineer. “All the different specialties work together,” he says.

For Zelner, the concept of epidemiology as an engineering discipline is something he inherited from his father, a mechanical engineer, who founded his own company designing healthcare facilities. Drawing from a childhood full of building and repairing things, the engineering mindset involves improving a transmission pattern, for example, in response to a moving target.

“Often these problems require iterative solutions where you make changes in response to what works and what doesn’t,” he says. “You keep updating what you’re doing as more data comes in and you see the successes and failures of your approach. To me, this is very different from the single, finished static image that many people have about academic science, where you have a big idea, and is more suited to complex, non-stationary problems that define public health. Test it and your result will always be preserved in amber.”