NYTimes | Who should get vaccine booster shots and when? Can vaccinated people with a breakthrough infection transmit the virus as easily as unvaccinated people? How many people with breakthrough infections die or get seriously ill, broken down by age and underlying health conditions?
Confused? It’s not you. It’s the fog of pandemic, in which inadequate data hinders a clear understanding of how to fight a stealthy enemy.
To overcome the fog of war, the Prussian general and military theorist Carl von Clausewitz called for “a sensitive and discriminating judgment” as well as “skilled intelligence to scent out the truth.” He knew that since decisions will have to be made with whatever information is available in the face of an immediate threat, it’s crucial to acquire as much systematic evidence as possible, as soon as possible.
In the current crisis, that has often been difficult.
These days, some experts grapple for answers on Twitter. They might be trying to figure out the effect of a vaccine booster shot by reverse engineering a bar chart in a screenshot from Israel’s Ministry of Health, or arguing with one another about confounding factors or statistical paradoxes.
Why this stumbling in the fog? It may seem like we’re drowning in data: Dashboards and charts are everywhere. However, not all data is equal in its power to illuminate, and worse, sometimes it can even be misleading.
Few things have been as lacking in clarity as the risks for children. Testing in schools is haphazard, follow-up reporting is poor and data on hospitalization of children appears to be unreliable, even if those cases are rare. The Food and Drug Administration has asked that vaccine trials for children aged 5 to 11 be expanded, which is wise, but why weren’t they bigger to begin with?
While the pandemic has produced many fine examples of research and meticulous data collection, we are still lacking in detailed and systematic data on cases, contact tracing, breakthrough infections and vaccine efficacy over time, as well as randomized trials of interventions like boosters. This has left us playing catch-up with emerging threats like the Delta variant and has left policymakers struggling to make timely decisions in a manner that inspires confidence.
To see the dangers of insufficient data and the powers of appropriate data, consider the case of dexamethasone, an inexpensive generic corticosteroid drug.
In the early days of the pandemic, doctors were warned against using it to treat Covid patients. The limited literature from SARS and MERS — illnesses related to Covid — suggested that steroids, which suppress the immune system, would harm rather than help Covid patients.
That assessment changed on June 16, 2020, when the results of a large-scale randomized clinical trial from Britain, one of all too few such efforts during the pandemic, demonstrated that dexamethasone was able to reduce deaths by one-fifth among patients needing supplemental oxygen and an astonishing one-third among those on ventilators.
The study also explained the earlier findings: Given too early, before patients needed supplemental oxygen, steroids could harm patients. But comprehensive data from the randomized trial showed that when given later, as the disease progressed in severity, dexamethasone was immensely helpful.
Dexamethasone has since become a workhorse of Covid treatment, saving perhaps millions of lives at little cost or fanfare. Without that trial, though, it might never have been noticed because of a problem called confounding: when causal effects of different elements can’t be considered separately. If doctors give multiple drugs to patients at the same time, who knows which drug works and which one does not? Or, if they choose which drug to give to whom, those more ill may be getting effective drugs, but the severity of their illness could end up masking the positive effect of the drug. Trials allow us to sort through all of this.
Randomized trials are not the only source of useful data. For example, it would have been difficult to quickly determine how transmissible the Delta variant is — a crucial question — without the data collected from close and systematic observation.
If a variant is spreading quickly somewhere, it might be more transmissible, or it could have simply arrived in that area early and gotten a head start. Or it might have just hit a few superspreader events. We’ve had variants appear, generating alarming headlines, that were later shown to be no more threatening than previous ones.
0 comments:
Post a Comment