The effort, led by Lawrence “Rusty” Hofmann, MD, professor and chief of interventional radiology, is a daily survey that tracks the occurrence of possible COVID-19 symptoms in communities. Called the Stanford Medicine National Daily Health Survey, it aims to provide data that could flag an uptick in COVID-19 cases before they reach the hospital, acting as a sort of early warning system for regions of the United States with large populations of undiagnosed individuals.

Published data shows that there is an average span of 10 days from the time people infected with the coronavirus that causes COVID-19 start feeling sick until the time they are hospitalized, assuming they progress to that level of severity. The way Hofmann sees it, that’s lost time, and the survey could help raise the alarm far sooner.

“I was sitting in my office 2 ½ weeks ago, frustrated that we’d only tested 35,000 Americans,” said Hofmann, who is also medical director of digital health. “Now we’ve tested just over a million Americans, which may sound like a lot, but that’s still less than a third of 1% of the U.S. population. It made me realize that what we’re doing right now just isn’t working.”

His team, which includes epidemiologists and data scientists, launched the survey on April 2. 

The success of the tool, Hofmann said, depends on one thing: You. And your neighbors, and your neighbors’ neighbors. “To work, this survey needs people,” he said. “My hope is that people see taking this survey as their civic duty and as a way to be involved in fighting COVID-19.” 

The data from the survey could help local and national officials predict surges of the disease and take action to mitigate the impact. That may include directing additional medical resources where surges are predicted or implementing new policies to prevent the virus from running rampant. The data, collected from participants who opt in to taking the survey, will be deidentified and shared broadly with public health officials and health care professionals to help inform decision-making and hospital preparedness.

Turning to data

Potential early warning signals from the survey are critical because they allow hospitals and health care workers to prepare, Hofmann said. “It’s a lot like our national air defense system. It picks up incoming threats thousands of miles out so that we can prepare our defenses,” he said. The same goes for hospitals. The extra time could be critical for finding bed spaces, organizing staff and securing equipment necessary to care for incoming patients.

The survey’s data, which Hofmann and his team will analyze as it comes in, will be coupled with local testing results to better predict the true prevalence and severity of COVID-19.  This will allow health officials to better respond, including by sending supplies and resources where surges will occur and developing public health policy, for example, implementing or lifting shelter-in-place orders.

The initial survey takes two minutes to complete, with daily surveys thereafter taking seconds. To broaden accessibility, Hofmann and his team are currently translating the survey into five languages and, depending on funding, hope to offer it via text message as well. “The survey was designed to be simple and streamlined to encourage participation,” Hofmann said. “We’re using a chatbot to help guide people through the process.”

After providing their ZIP code, users are asked about any recent COVID-19-like symptoms they’ve experienced and any contact with people who have COVID-19, among other basic details. Unless users opt out, they will receive a daily email reminding them to complete the survey.

Light lift, long-term commitment

“It’s crucial to continue to take the survey every day, especially during this time of uncertainty and while we shelter in place,” Hofmann said. “But even after this initial phase is over, there will be pockets of COVID-19 cases that crop up, which is why it’s important to make a habit of it and stay consistent.”

Hofmann also acknowledges that the survey will inherently capture people who might have other coronaviruses or flus, due to symptom overlap. But over the past decade, careful tracking of colds and flus have created a relatively predictable pattern of infection rate and timing for these viruses. In addition, survey data coupled with in-person swab data can be used to further refine predictions. Data that deviates from that pattern, he said, would help differentiate COVID-19 from other illnesses.

“This survey is crucial for monitoring symptomatology and understanding what’s happening on a continual basis, rather than a single point in time, like the swab test,” Hofmann said. “We don’t want to be caught flat-footed, because this fight isn’t over until we have a vaccine.”

You can participate by taking the survey here: http://stan.md/covid19. 

If you’re interested in partnering with Hofmann and his team, please contact Leslie Haas at lesliehaas@stanfordhealthcare.org.