Former FDA chiefs Scott Gottlieb and Mark McClellan are pushing for the government to implement a national surveillance system for COVID-19. Gottlieb has been working with the White House’s Coronavirus Task Force on how to implement a sentinel surveillance system, and a spokesperson for the Centers for Disease Control and Prevention confirmed to Inside Health Policy that there have been discussions and some work is being done.
Currently, the best system the United States has for tracking COVID-19 is counting the number of people who have died from the disease, Nirav Shah, adjunct professor of medicine at Stanford University and senior scholar for the university’s Clinical Excellence Research Center, told IHP.
Though the death rate is viewed as being fairly accurate, most of those people likely got infected three weeks earlier. Hospitalization rates are a much more representative number and come closer to showing the infections happening in real time, but only about half of the states are reporting hospitalizations for COVID-19, Shah says.
In an American Enterprise Institute’s coronavirus response roadmap, published March 30, Gottlieb, McClellan, former FDA Chief of Staff Lauren Silvis and other public health experts say a high-performing disease surveillance system is needed to progress through phase I of pandemic response -- moving toward less-restrictive physical distancing guidelines while at the same time not sparking an acceleration in COVID-19 case counts.
A high-performing surveillance system should leverage widespread and rapid testing at the point-of-care and serological testing to gauge background rates or exposure and immunity. It should also have support from local public health systems and health care providers, the report says.
The objective of a COVID-19 surveillance system, the World Health Organization wrote in a March 20 guidance, would be to: monitor trends in COVID-19 disease at national and global levels; rapidly detect new cases in countries where the virus is not circulating and monitor cases in countries where the virus has started to circulate; provide epidemiological information to conduct risk assessments at the national, regional and global levels; and provide epidemiological information to guide preparedness and response measures.
Gottlieb and McClellan point to CDC’s Influenza-Like Illness Surveillance Network (ILINet) as a potential model for a severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) surveillance system. SARS-CoV-2 is the virus that causes COVID-19.
They add that SARS-CoV-2 surveillance will benefit from data sharing and coordination with health care providers and payers.
To develop and support such a system, CDC needs to convene an intergovernmental task force with outside experts as needed and input from states and the health care community, they say.
Once a robust surveillance sentinel system is in place and is coupled with widespread point-of-care testing and a robust ability to implement tracing, isolation and quarantine, the United States can enter phase III of pandemic response, which involves lifting all restrictions.
But there are weaknesses in CDC’s system, according to Martin, Blanck & Associates Senior Partner Joel Greenspan, who also formerly served as CDC assistant director for planning, communications and external relations.
Among those weaknesses: duplicated costs of provider recruitment and data management, low practice coverage, duplicated efforts in provider practices, inconsistent weekly provider compliance, slow data turn-around, lack of publicly available metropolitan statistical area influenza-like illness data, and lack of forecasting capability during the current epidemiological week, Greenspan wrote in a February 2015 article in the Online Journal of Public Health Informatics.
The current version of ILINet, which was established in the 1980s, relies on more than 2,000 volunteer providers recruited by states to report syndromic influenza-like illness, something Greenspan refers to as an ongoing hunter-gatherer approach to influenza-like illness outpatient surveillance.
Much has changed in the U.S. health care system since the 1980s, he notes. For example, e-commerce standards emerged in the 1990s, creating ubiquitous amounts of easily accessible electronic health care administrative data. And since 2001, new public health surveillance approaches and investments have emerged including methods for syndromic surveillance.
In his paper, Greenspan says big e-health data can be harvested immediately to begin developing an ILINet 2.0 and for faster and more granular influenza-like illness surveillance for the U.S. public health and national security communities.
He adds that an alternative ILINet 2.0 model shows that electronic health records can generate timely outpatient influenza-like illness signals without recruiting providers directly. Additionally, tracking anti-influenza prescription drugs provides a comparable signal to provider-office influenza-like infection signals.
As the United States explores what is needed to set up a national sentinel surveillance system, Shah points out that health technology company Kinsa Inc., which makes Bluetooth-connected thermometers, and Unacast, a data insight organization, already have made data available to the public and policymakers that can help make decisions based on real-time information.
Kinsa, for example, takes data from its network of smart thermometers and associated mobile apps to provide a map of influenza-like illness linked with fever.
The company developed a method for identifying anomalous influenza-like illness incidence outbreaks in real-time using illness signals developed from its geospatial thermometer data and 12-week illness forecasts. Kinsa flags anomalously high incidence data by comparing real-time influenza-like illness incidence to expected seasonal influenza trends, the company explains on its website, HealthWeather.us.
The method is being applied to aid in early detection of potential COVID-19 outbreaks.
At HealthWeather.us, the company forecasts illness trends for every county in the continental United States from March 1, before widespread COVID-19 infections were observed, and compares its real-time data to those expectations. For example, in Brooklyn, NY, the company began to see anomalous events into the second week of March, Kinsa’s website says.
“This provides us with guidance of where potential COVID-19 outbreaks may be occurring. This method holds promise for real-time illness anomaly detection efforts used to identify emerging pandemics and severe flu outbreaks,” the company says.
Shah, who is an advisor to Kinsa, says the company has deployed a million smart thermometers across the country.
“The beauty of this is we know how to predict what the flu looks like based on years of data. CDC currently predicts the flu for three weeks out based on their data. This company has cracked the code on predicting flu 20 or more weeks away in the future,” Shah says.
Because health experts know what the flu is and when it occurs, Kinsa, through its thermometer data, can subtract out what it knows is the flu, and anything that shows up on the map that’s not the flu could either be a resurgence of influenza or considered a presumptive COVID-19 case.
“This isn’t saying it’s definitively COVID-19. It’s saying it’s abnormal, go check it out. In many cases, it has turned out to be COVID-19,” he says.
Unacast also has made publicly available the data it has collected to grade social distancing within communities.
The organization has applied its Real World Graph data engine to create a Social Distancing Scoreboard using anonymized data from mobile phones and their interactions with each other.
The data capture how people have adapted their everyday behavior, including: working from home instead of the office, avoiding non-essential trips to entertainment or recreational facilities, and canceling vacations. Those results have been extrapolated to the population level to show how each state and county are adhering to social distancing guidelines.
Unacast’s map, available at Unacast.com, notes where there have been changes in non-essential visits, changes in average mobility based on distance traveled and newly reported COVID-19 cases.
The organization says the map can help reinforce the importance of social distancing, but it also can help unearth trends that will help project scenarios in the short- and mid-term future.
Data from private companies like Unacast and Kinsa could work for short-term decision-making in the health care community, but also eventually be built into a bigger national surveillance system, Shah says.
“This is repurposing existing data for the public good. Private companies doing things for the public good,” Shah says.
He adds, “We need to bridge the artificial divide between the rigor and trusted information sources of the past and the creative approaches available today that leverage big data, connected devices and machine learning. Today’s epidemic requires today’s tools.” -- Beth Wang (bwang@iwpnews.com)
