Two recently published studies indicate that COVID-19 infections were already circulating in the United States in December 2019. The question is whether these methodologies that could be applied to track the next pandemic.
One study evaluating blood donations found antibodies on the West coast as early as Dec. 13, 2019, and in blood donated on the East Coast by early January 2020 (). Both preceded the first documented COVID-19 infection in the United States, which has been widely reported as occurring on Jan. 19, 2020, in a traveler returning from China.
The other study, utilizing electronic medical record (EMR) analytics, demonstrated a spike in visits or hospitalizations for cough, a trend that persisted from Dec. 22, 2019, onward, exceeding norms for seasonal flu (). This spike was interpreted as evidence that the SARS-CoV-2 pandemic was already underway before the first case was established.
While the ongoing serologic testing of blood donations for viral antibodies “will advance understanding of the epidemiology” for SARS-CoV-2 and “inform allocation of resources and public health prevention interventions to mitigate morbidity and mortality,” it might also be a strategy for disease surveillance in the next pandemic, according to a team led by investigators at the Centers for Disease Control and Prevention.
Blood donation surveillance is not now used routinely to monitor for population-based health threats, but it is not a new idea, according to the lead author of the study,, of Emory University and director of the CDC’s Office of Blood, Organ, and Other Tissue Safety, Atlanta, and his coinvestigators. Most recently, blood donation surveillance was used in the United States to track the penetration of the Zika virus.
For early detection of respiratory infections, blood donations might have unique advantages over alternatives, such as surveillance of respiratory specimens from symptomatic patients. Not least, blood donation surveillance captures individuals who are not seeking medical care, according to the investigators.
EMR surveillance might also have unique advantages for population-based monitoring of health threats. For one, aggregate data from large EMR systems have the potential to reveal symptom patterns before they become apparent at level of clinical care, according to a team of collaborating investigators from the University of California, Los Angeles, and the University of Washington, Seattle.
Emphasizing an urgent need for “agile healthcare analytics” to enable “disease surveillance in real time,” the first author of the EMR study,, professor in the department of health policy and management at the University of California, Los Angeles, expressed the hope that the approach will “lead to better preparation and the ability to quickly provide warnings and track the next pandemic.”
In the blood donation surveillance study, the goal was simply to determine whether SARS-CoV-2 reactive antibodies could be found in blood donations before the first case was identified. Of the 7,389 archived blood samples tested between Dec. 13, 2019, and Jan. 17, 2020, 106 (1.4%) were reactive.
These were not true positives, acknowledged the investigators. True positives would require reactive antibodies in the context of a positive molecular diagnostic test or paired acute convalescent sera with rising titers. The investigators also cautioned that false positives could not be completely ruled out, particularly in light of cross-reactivity that has been reported with other human coronaviruses.
Nevertheless, the monitoring of blood donations offers substantial promise for “understanding the dynamics of SARS-CoV-2 pandemic from early introduction,” and the CDC is now collaborating on ongoing surveillance with the goal of contributing information that could be applied “to mitigate morbidity and mortality.”
Lessons learned from this pandemic are potentially relevant to the next.
The EMR study simply looked at whether the word “cough” was included more often in the notes from visits or hospitalizations between December 2019 and February 2020 relative to the preceding 5 years. The investigators drew on data from three hospitals and more than 180 clinics.
From Dec. 22, 2019, onward, cough was noted above the 95% prediction interval for all 10 weeks of the study. The excess was seen in the outpatient setting and among hospitalized patients. There was also significant excess in the number of patients hospitalized with acute respiratory failure during the study period.
“Our approach to analyzing electronic records could be helpful in the future as we included consideration of data from the outpatient clinics in addition to the emergency departments and inpatient settings,” Dr. Elmore reported.
Surveillance of influenza and influenza-like infections has been undertaken in the United States for more than 20 years, but Dr. Elmore contends that EMR data, particularly data from outpatient clinics are “usually a harbinger of what is to come” for emergency department visits and, ultimately, hospitalizations. She thinks that this is a resource not yet fully exploited.
“There are always opportunities to better harness EMR data,” Dr. Elmore said.
These are intriguing studies and “useful” for reconsidering when SARS-CoV-2 was introduced in the United States, according to, a professor of epidemiology and the associate dean of the University of Washington School of Public Health, Seattle. However, she noted that the task of translating data like these into actionable public health strategies has proven difficult in the past.
Symptom-based surveillance systems “have mostly served as situational awareness rather than early detection tools,” Dr. Baseman said. The problem is timely interpretation of a given signal.
Not that she doubts such tools “would be an incredible resource for humanity” if the current limitations can be resolved or that technological advances will lead to better methods of detecting and monitoring pandemics “at some point.” Rather, “we’re just not there yet,” she said.
SOURCE: ); Elmore JG et al. ).