Drosten, of the Charité University Hospital in Berlin, is one of many calling for renewed vigilance, and he and others are urging a new control strategy that trades blanket lockdowns for measures specifically targeting clusters of cases, which play a key role in spreading the coronavirus. “We successfully aborted the [first] wave and now we should make sure that no new wave builds,” epidemiologist Christian Althaus of the University of Bern says.Putting more effort into finding clusters should also help epidemiologists understand where and how they emerge, says Hitoshi Oshitani of Tohoku University in Japan—which may have changed since the spring. “We've seen a massive change in the social structure and interactions of populations … from the start of the pandemic,” Kucharski says. The conditions that spread the virus then “won't necessarily be the same ones that are creating the risk now.” In Germany, for instance, many large outbreaks early in the pandemic occurred in long-term care facilities. Now, clusters are increasingly reported from workplaces.Instead, public health experts increasingly argue for targeting clusters of cases and superspreading events. Some studies estimate that 10% of patients cause 80% of all infections, whereas most don't infect anybody at all (Science, 22 May, p. 808). Drosten has urged that contact tracers spend more time finding the source of a new case—along with that person's contacts—than the new case's contacts; after all, the patient may not infect anybody else, but is likely to have caught the virus as part of a cluster, Drosten says.Adam Kucharski, a disease modeler at the London School of Hygiene & Tropical Medicine, agrees. “Looking backwards can actually give you a disproportionate benefit in terms of identifying infections,” he says. In a recent preprint, Kucharski and his colleagues estimated that “backward contact tracing” could prevent twice as many infections as tracing contacts forward alone. Experience in South Korea, where clusters at churches drove the epidemic early on, confirmed the value of this approach, says University of Florida biostatistician Natalie Dean.More-targeted measures probably won't be enough to keep the virus from resurging, Althaus says. “A point will be reached again where stricter measures have to be taken,” he says. But rather than complete lockdowns, he assumes they will be more like the lighter version applied in Sweden, which encouraged people to work from home and banned large gatherings while keeping shops and restaurants open.