While hospitals across the country are busy developing contingency plans for the Covid-19 patient surge, the actual timing of the surge in any given community remains something of a mystery. Existing epidemiological models offer predictions about when local cases will peak, but the data they use is often too incomplete to provide anything more than a general timeline. Testing can confirm existing cases, but testing capacity remains low and rarely captures patients experiencing early symptoms.
This is where the new Stanford Medicine National Daily Health Survey comes in. The survey is designed to serve as a Covid-19 early warning system, identifying new infection hot spots across the country days before patients start appearing in the hospital.
I recently spoke with Lawrence "Rusty" Hofmann—chief of Interventional Radiology and Medical Director of Digital Health at Stanford Hospital, who worked with epidemiologists and data scientists to develop the tool—and Leslie Haas, digital health strategy manager at Stanford Health Care, to learn more about the tool.
The goal, Hofmann said, was to develop better intelligence about the spread of the disease in the absence of comprehensive testing. "So far we've tested just over a 2 million Americans, which sounds like a lot but is still less than one percent of the population," Hofmann said. "We need tools to help us better understand the spread of Covid-19 sooner."
How does the tool work?
The principle behind the survey is simple. Since the lag between the onset of first symptoms and hospitalization is around 10 days, and since the average hospitalization rate has been around 20%, knowing how many people in a given ZIP code have recently developed symptoms should give us a glimpse into future hospital demand.
Using a simple web interface supported by a chatbot, the survey asks respondents a few basic questions to collect demographic information, the ZIP code in which they live, any symptoms they may be exhibiting on that day, and any previous Covid-19 exposures they may have had. The entire survey, which is free and available to the public, takes less than two minutes to fill out initially, and less than 10 seconds each subsequent day.
How will the tool be used?
Once the data has been collected, de-identified, and analyzed, Hofmann said his team plans to share it widely with public health officials, health systems, and health care professionals to assist them in coordinating their response to surging demand.
In that 10-day window between the first onset of symptoms and seeking medical attention, hospitals can take steps to ensure they have the right inpatient staff mix in place, or enough ventilators and PPE available to meet the estimated need. Similarly, local public health officials can use that time to determine when shelter-in place orders could be lifted—or in some cases re-instituted—and ramp up targeted testing in coordination with hospitals in the community. With widespread adoption, once these stay-at-home orders are lifted, the tool can be used to monitor if there is a resurgence of Covid-like symptoms to further guide health officials on next steps.
The tool's limitations
Of course, the effectiveness of the survey depends on the number of people participating. The more responses, the more accurate the picture of where we are and where we are going. It also depends on how often people participate. Daily participation is essential to capture how Covid-19 may be spreading through a community. So far, the voluntary survey has enrolled more than 250,000 members or daily responses since it was first launched on April 2. According to Haas, the ultimate vision is to not only increase use across the United States, but also to expand international to aid global Covid-19 control efforts.
How can providers aid in the fight to identify hot spots before they happen? According to Hofmann and Haas, they can start by encouraging all providers, patients, community partners, and members of the public to download the survey and fill it out daily. The greater the number of participants in any given community, the more reliable the data will be for hospitals serving that community.
Other Covid-19 symptom tracking tools
Stanford is not alone in creating new Covid-19 symptom-tracking tools. In the absence of a simple, effective mechanism for identifying patients, a range of other health systems including Mount Sinai Health System, Massachusetts General Hospital, and Intermountain Healthcare have developed their own tools.
Existing tools serve a range of functions, from gathering information about the local spread of the virus to directing patients with Covid-19 symptoms to appropriate health care resources. Other players are also getting into the game. On Friday, Google and Apple announced a joint project for contact tracing that will send smartphone notifications to people who have come in contact with someone with the virus. The widespread adoption of these symptom-tracking tools is essential to developing an effective track-and-trace approach that can be applied nationally. And public health experts agree that a national strategy is needed to effectively curb the spread of the new coronavirus—and to detect and prevent any future outbreaks from reaching epidemic levels.