Can there ever be scientific certainty during a pandemic caused by a novel infectious disease? Ann C. Keller, associate professor at Berkeley Public Health and a political scientist, says no—and argues that the Centers for Disease Control and Prevention would be wise to remember that.
In an essay published online ahead of print in the Journal of Health Politics, Policy and Law, Keller takes stock of the CDC’s post-COVID reforms. Much-debated at the agency and in the press, these reforms-in-progress aim to improve the agency’s data collection and analysis, and seek to reduce error and controversy in pandemic response. She also offers advice for how the agency could improve communication with the public during the next major infectious disease outbreak.
Keller, who has been studying CDC’s pandemic response since 2002, contends that when addressing a pandemic driven by a novel pathogen, error is unavoidable; and that strong policies achieve public acceptance by surviving controversy rather than trying to avoid it.
“I was surprised by criticism of public health agencies that seemed to stem from post hoc analysis,” she said in an interview. “A lot of critics, for example, were essentially arguing, ‘The CDC should have written guidance at time T with data that was available at time T+n [n =weeks, months, years; meaning some time after time T]. Asking civil servants to know a thing before it is knowable is not exactly a strategy that a manager can implement in a future pandemic.
“In response, I wanted to address what I thought was a very complex problem. How do you hold experts and policy makers accountable when you know they are likely to make mistakes? What does it mean to ‘learn in plain sight,’ especially in a massively contested political environment?”
Keller’s article is part of a book-in-progress that develops the idea of what she calls “accountable learning” in crisis response. It grew out of her years studying the politics of science, and the effects of hyperpolarization on executive branch agencies and management challenges of responding to infectious disease outbreaks.
Keller reviews four cases in which CDC guidance during the pandemic became the subject of intense debate among experts with opposing views. These apparent guidance failures contributed to both the public and health officials losing trust in the CDC.
The four cases in the paper are:
CDC’s changing and confusing recommendations on using cloth face coverings or other masks in the early days of the pandemic;
political censorship, including that of CDC official Nancy Messonnier, after she gave a press conference warning the public to prepare for ‘significant disruption’. Soon after her remark, she was barred from talking to the press and no longer appeared in White House briefings. Press calls to her office were intercepted by others.
Changing conditions that sparked changing guidelines on vaccine shortages and the possibility of breakthrough infections for those who were fully vaccinated;
and COVID-19 isolation guidelines that emphasized the importance of isolation early in the course of infection.
“Experts and policymakers are not omniscient,” Keller said. “During a crisis, they have to make decisions with too little information and too little insight into what might happen next.”
Her concept of “accountable learning” offers a solution that requires that those making policy commitments explain their decisions and articulate how they plan to test their assumptions over time.
“Transparency around this learning process can create accountability around a realistic performance standard,” Keller said.
Keller hopes that the paper will spur a discussion over what the public can realistically expect from public health agencies during a novel infectious disease outbreak.
“I ask myself, ‘Why do we want experts to lead during a pandemic?’ It’s not because their expertise gives them perfect insight. It’s because we think experts can learn more quickly and with less error than non-experts,” she said. “What I’d like to see is more willingness of the public to be open to the real dynamics of learning in this very challenging environment.”
In her article, Keller also examines the reforms that CDC proposed in 2024, including: increasing efficiency and scope of data collected through electronic reporting; improving the development, production, and distribution of reliable tests; and increasing analytic capacity through improved modeling, which the CDC has said it hopes will generate better and faster insights about future outbreaks.
But Keller doesn’t believe that faster data collection would translate into faster insight. Instead, she suggests that CDC consider how to engage the public after they have issued erroneous guidance—a problem that is nearly impossible to avoid.
She acknowledges that her strategy would only work if elected officials allow for open, unadulterated communications between the CDC and the public. She notes past censorship by the Trump administration, and Trump’s vow to severely cut civil service jobs during his second presidential term.
“Many Americans may believe that federal experts operate outside of the control of elected officials, but the reality is that federal agencies are subject to an array of political controls exercised by Congress, the courts, and, most directly, by the White House,” Keller said.
“In 2020, Trump was able to prevent experts at the CDC from communicating their understanding of the best ways to prevent COVID-19 transmission. He plans to go further during his second term in that he has promised to end the civil service protections that have been in place in the U.S. since the 1880s,” said Keller.