The world this week News in focus ROBERT NICKELSBERG/GETTY ROBERT A health-care worker checks data on a computer before treating a person with COVID-19 in New York City. WHY THE UNITED STATES IS HAVING A COVID-19 DATA CRISIS Political meddling and years of neglect for public-health data management are to blame.

By Amy Maxmen Prevention has updated its website with frustrating data researchers, who want to help near-real-time information on local outbreaks. authorities make decisions that can save lives. outh Korea’s grip on the coronavirus The site also reports several COVID-19 statis- “We shouldn’t be flying blind at this point,” faltered this month when a large tics for every region of the country. says , a biostatistician at the Uni- church in Seoul had an outbreak — Data dashboards in Singapore and New versity of Florida in Gainesville. “We shouldn’t involving 915 cases as of 25 August. Zealand offer similar windows into how the have to speculate.” The government has reinstated coronavirus is spreading within their borders. Experts told Nature that political meddling, Srestrictions in the city to prevent a surge, but This helps policymakers and citizens deter- privacy concerns and years of neglect of it’s also reporting details of the outbreak pub- mine how to go about daily life, while reduc- public-health surveillance systems are among licly. For instance, it has shared that 120 people ing risks — and provides researchers with a the reasons for the dearth of information in infected at the church have spread the corona- wealth of data. By contrast, the United States the United States. virus to people at 22 venues, including 4 call offers vanishingly few details on how the dis- Although information isn’t the only tool centres and 3 hospitals in Seoul. ease is spreading, even as people increasingly that can be used against a pandemic, South Almost every day for the past seven months, socialize and travel, and authorities reopen Korea’s attention to data correlates with its the Korea Centers for Disease Control and schools and businesses. This state of affairs is overall success at controlling the outbreak:

Nature | Vol 585 | 3 September 2020 | 13 ©2020 Spri nger Nature Li mited. All rights reserved.

News in focus the country has had about 3.5 cases per The CDC was further sidelined in July, health records voluntarily, but he’s not sure 10,000 people overall, and there have been scientists say, when the Trump administra- whether the state or local health department around 2 COVID-19 deaths per week over the tion announced that data on COVID-19 cases use the information, which troubles him. He past month. By contrast, the United States has and hospitalizations would be diverted away worked on the response in Guinea during had 175 cases per 10,000 people overall, and from the agency, and handled instead by a new the outbreak in 2015, and says authorities there about 7,000 people have died of the disease system launched at the HHS, the head of which collected data on how people were infected every week for the past month. reports directly to the president. So far, the with the Ebola virus, to curb its spread. “This South Korea owes its detailed intel to a HHS’s dashboard has been up to a week behind crucial information would help us drill down coordinated network of public-health centres in reporting data, and it includes information transmission further” for COVID-19, he says. in 250 districts that send information rapidly only on case counts and hospital capacity, “It’s crazy we aren’t collecting it.” to the Korea Centers for Disease Control and rather than details such as the location of dis- Prevention. Sung-il Cho, an epidemiologist ease clusters. An HHS spokesperson says that Access denied at Seoul National University, attributes the the new system streamlines reporting from Epidemiologists at universities would like system’s success to this centralization of 6,000 hospitals in the country. to help overburdened state and local health power — along with rapid hiring of ‘tempo- But Georges Benjamin, the executive departments to analyse whatever data they rary epidemiologists’ to meet demand dur- director of the American Asso- have so they can help officials to direct an ing the COVID-19 pandemic. These scientists ciation (APHA) in Washington DC, says that efficient coronavirus response. Typically, have helped to lead contact-tracing investi- rather than streamlining data collection, the health departments share disease-surveil- gations that produce succinct, anonymized change has confounded information gather- lance data with researchers on request. But details such as those listed about the outbreak ing further. Some hospital administrators are during the COVID-19 pandemic, epidemiolo- at Sarang Jeil Church in Seoul. now confused about which agencies to report gists have been denied access. For instance, The United States isn’t pursuing contact to, he says. He’s also frustrated because the Steffanie Strathdee, an HIV epidemiologist tracing for COVID-19 to nearly the same extent US$10 million spent on the HHS’s system at the University of California, San Diego, as South Korea, but its disease surveillance also might have been put to better use improving often requests case data broken down by flows from local health departments to the fed- parameters such as locality, race and prob- eral level. For years, the US Centers for Disease “We’ve begged for able exposure route, such as intravenous Control and Prevention (CDC) has used this sys- drug use. “Surveillance is done and it’s made tem to follow the spread of outbreaks such as a money over the years transparent, and this is what we have been current surge in infections with the bacterium to build a solid accustomed to,” she says. “Only for this epi- Salmonella, and to trace them back to their information highway.” demic, things have changed.” sources. However, the surveillance system has This year, Strathdee and other epidemiol- run into problems at several levels during the ogists asked for anonymized COVID-19 data pandemic. The result is that a lot of data, such public-health data management at the CDC from the California department of health. as information on where people were exposed and at the health departments it collaborates E-mails seen by Nature show that the research- to the coronavirus, are missing. And the data with across the country. This outdated system ers’ requests considered the need for individ- that do exist are made public only slowly. has buckled under the pressure of 5.7 million ual privacy with measures such as asking the The CDC and four US health departments COVID-19 cases in the United States. state for age ranges instead of exact ages. But declined to speak to Nature about how they’re Benjamin notes that many health depart- the request was denied — something that had managing COVID-19 data. However, former ments still share data by fax, which is more never happened to Strathdee before. The scientific employees and researchers who time-consuming than digital methods. A department’s director, Mark Ghaly, explained collaborate with them offered suggestions lack of funds also means that overburdened in a 3 July e-mail that releasing information about why data are lagging and lacking in the staff don’t have enough time to analyse the from people’s records would “necessitate a United States. data they have. The APHA and other sci- careful and time-consuming analysis of each entific organizations have long called for record to determine what data fields may be Increased scrutiny resources to improve data surveillance in the made public from each individual report”. Some speculate that because the pandemic US public-health system. Without up-to-date, reliable informa- is politically charged, data describing the sit- “We’ve begged for money over the years to tion on who is infected, why and where, uation are guarded closely by officials in the build a solid information highway so that we US scientists, policymakers and the public administration of President . can collect data rapidly and share it with the must instead rely on media reports and inde- Researchers say that investigations published people that need it in a timely way,” says Benja- pendent efforts to consolidate data, such as in the CDC’s journal Morbidity and Mortality min. “But we’ve never gotten what we needed.” the COVID Tracking Project from magazine Weekly Reports have been thorough, but are This long-standing neglect has been exac- The Atlantic and the COVID-19 Dashboard posted online long after they might influ- erbated by the lack of national leadership compiled by researchers at Johns Hopkins ence outcomes. For example, on 31 July, the during the pandemic, say the researchers University in Baltimore, Maryland. CDC reported that 260 staff members and who spoke to Nature. There are no nation- But data from news outlets aren’t necessar- children at an overnight camp in Georgia wide requirements for the information that ily comprehensive or universally trusted, and had been infected more than a month earlier hospitals and testing laboratories must report the dashboard doesn’t detail where transmis- (C. M. Szablewski et al. Morb. Mortal. Weekly to health departments. Ranu Dhillon, an epi- sion is happening. Rep. 69, 1023–1025; 2020). Samuel Groseclose, demic-response specialist at Harvard Medical There is an urgent need for such informa- a public-health specialist who retired from School in Boston, Massachusetts, who is cur- tion, says Caitlin Rivers, an epidemiologist at the CDC in 2018, suggests that the reports rently treating people with COVID-19 in Vallejo, Johns Hopkins, because people are returning are undergoing an unusual amount of review California, says that neither the local health to work, to socializing and to schools. That within the agency, and perhaps within its par- department nor the CDC tells him to report means that precisely tailored interventions ent agency, the US Department for Health and where patients might have been exposed to are more important than ever. “It’s not Human Services (HHS). the coronavirus. He records these data in their enough to just tell people to be cautious.”

14 | Nature | Vol 585 | 3 September 2020 ©2020 Spri nger Nature Li mited. All rights reserved.