[ad_1]
The last 20 months have turned every dog into an amateur epidemiologist and statistician. Meanwhile, a group of conscientious epidemiologists and statisticians believe that pandemic problems can be solved more effectively by embracing engineer thinking: that is, focusing on pragmatic problem solving with an iterative, adaptive strategy to make things work.
In a recent essay, Reporting Uncertainty in a Pandemic, researchers reflect on their role during a public health emergency and how they could be better prepared for the next crisis. The answer, they write, may lie in rethinking epidemiology from a more engineering point of view and a less “purely scientific” point of view.
Epidemiological research informs public health policy and its inherent mandate for prevention and protection. But the right balance between clean research results and pragmatic solutions proved alarmingly elusive during the pandemic.
We need to make practical decisions, so how much uncertainty really matters?
Seth Gikema
“I’ve always imagined that in this kind of emergency, epidemiologists would be useful people,” said John Zellner, co-author of the essay. “But our role is more complex and less defined than I expected at the beginning of the pandemic.” A model of infectious diseases and a social epidemiologist at the University of Michigan, Zellner witnessed a “crazy spread” of scientific articles, “many of which have given very little thought to what this really means in terms of positive impact.”
“There were a number of missed opportunities,” says Zellner, caused by a lack of connections between the ideas and tools proposed by epidemiologists and the world they need to help.
Denial of security
Co-author Andrew Gelman, a statistician and political scientist at Columbia University, presented the “bigger picture” in the introduction to the essay. He compares the outbreak of the pandemic by amateur epidemiologists to the way war turns every citizen into an amateur geographer and tactician: “Instead of maps with colored pins, we have exposure charts and numbers of deaths; people on the streets argue about infection mortality and herd immunity in a way they could discuss military strategies and alliances in the past. “
And with all the data and public discourse – are masks still needed? How long will vaccine protection last? – came the series of uncertainty.
Trying to figure out what just happened and what went wrong, researchers (including Ruth Ezioni of the University of Washington and Julien Rio of the University of Bern) did something like a reconstruction. They looked at the tools used to address challenges, such as assessing the rate of human-to-human transmission and the number of cases circulating among the population at any given time. They evaluate everything from data collection (data quality and interpretation are perhaps the biggest challenges of the pandemic) to model design to statistical analysis, as well as communication, decision-making and trust. “Uncertainty is present at every step,” they wrote.
Still, Gelman says, the analysis still “does not sufficiently express the confusion I went through in those early months.”
One tactic against all uncertainty is statistics. Gelman thinks of statistics as “mathematical engineering,” methods and instruments that relate to measurement as well as discovery. The statistical sciences are trying to shed light on what is happening in the world with variation and uncertainty. When new evidence arrives, it must generate an iterative process that gradually refines prior knowledge and thins security.
Good science is modest and able to improve in the face of uncertainty.
Mark Lipsich
Susan Holmes, a Stanford statistician who was not involved in the study, also sees parallels with engineering thinking. “An engineer always updates his picture,” she says, reviewing with the advent of new data and tools. When solving a problem, an engineer suggests a first-order approximation (blurred), then a second-order approximation (more focused), and so on.
However, Gelman had previously warned that statistical science could be used as a machine to “wash away uncertainty” – intentionally or not, stupid (uncertain) data coming together and looking convincing (certain). Anti-uncertainty statistics are “too often sold as something like alchemy that will turn these uncertainties into certainty.”
We witnessed this during the pandemic. Drowning in turmoil and uncertainty, epidemiologists and statisticians – both amateurs and experts – clung to something solid, trying to stay afloat. But as Gelman points out, demanding security during a pandemic is inappropriate and unrealistic. “Premature security is part of the challenge of pandemic solutions,” he said. “This jump between uncertainty and security has created a lot of problems.”
Rejecting the desire for security can be liberating, he says. And partly here comes the engineering perspective.
Amazing thinking
For Seth Guickema, co-director of the Center for Risk Analysis and Informed Decision Making at the University of Michigan (and Zelner’s collaborator on other projects), a key aspect of the engineering approach is to immerse yourself in uncertainty by analyzing the mess, and then take a step backwards with the perspective: “Do we have to make practical decisions, so how much uncertainty really matters?” Guikema. “But if that doesn’t really affect my best decisions, then it’s less critical.”
For example, increasing the coverage of SARS-CoV-2 vaccination among the population is a scenario in which even if there is some uncertainty about exactly how many cases or deaths vaccination will be prevented, the fact that it is very likely to reduce and both, with little adverse effects, is the motivation sufficient to decide that a large-scale vaccination program is a good idea.
An engineer always updates his photo.
Susan Holmes
Engineers, Holmes said, are also very good at breaking down problems into critical pieces, applying carefully selected tools and optimizing for constraint solutions. With a team of engineers building a bridge, there is a cement specialist and a steel, wind and civil engineer specialist. “All the different majors work together,” she says.
For Zellner, the idea of epidemiology as an engineering discipline is something he took from his father, a mechanical engineer who set up his own healthcare design company. Using a childhood full of building and fixing things, his engineering thinking involves driving – perfecting a transmission model, for example, in response to a moving target.
“Often these problems require iterative solutions in which you make changes in response to what works or doesn’t work,” he says. “You keep updating what you do when more data comes in and you see the successes and failures of your approach. For me, this is very different — and more appropriate for the complex, non-stationary problems that determine public health — than the kind of static, monotonous image that many people have of academic science, where you have a great idea, test it, and the result is preserved. in amber for all time. ”
[ad_2]
Source link