Updated below with details from news conference.
Early results from a first-of-its-kind at-home testing program for COVID-19, analyzing more than 4,000 nasal samples sent in from across the Seattle region, have found the proportion of the disease to be lower than in medical testing programs, but still enough to potentially “represent thousands of unrecognized infections in the community.”
The approach, known as “swab and send,” is being held up as a potential model to help enable the widespread testing programs that health officials say will be critical as the country seeks to reopen businesses and start to return to some semblance of normal life in the months ahead.
Researchers from the greater Seattle Coronavirus Assessment Network (SCAN) said Friday afternoon that their best estimate from the first 18 days of results puts the community prevalence of COVID-19 at 24 cases per 10,000 people, or 0.24%, for the initial study period between March 23 and April 9. The prevalence declined toward the end of the testing period.
Among participants who reported symptoms consistent with COVID-19, 1.6% were found to have the disease.
They cautioned, however, that the sample size isn’t statistically significant, and it’s too early to draw any definitive conclusions from the ongoing study.
The approach being piloted by SCAN could be a model for testing elsewhere, at a larger scale, but it would need to be streamlined and made more cost-efficient, said Dr. Jay Shendure, scientific director of the Brotman Baty Institute and SCAN’s lead partner. For example, he said, the process of associating a sample with a specific person would need to shift from paper to digital input.
“From my perspective, it is a system that could be expanded,” Shendure said, discussing the initial findings on a conference call with reporters Friday afternoon.
SCAN partnered with groups including Amazon Care, a healthcare program for Amazon employees in the Seattle area, on infrastructure and logistics for the at-home tests.
“One of the things that we’re hoping to achieve through SCAN, and in particular by partnering with logistical entities that have a broader reach than just Seattle, is to work out some of those kinks, so the problems don’t need to get resolved in every nook and cranny of the country if efforts like this are sprung up elsewhere,” Shendure said.
SCAN was formed as a partnership between Public Health — Seattle & King County and the researchers behind the Seattle Flu Study, funded by Bill Gates’ private office, Gates Ventures, with technical support from the Bill & Melinda Gates Foundation and the Centers for Disease Control and Prevention. The study is implemented by the Brotman Baty Institute, which is a collaboration among UW Medicine, Fred Hutchinson Cancer Research Center, and Seattle Children’s, with data modeling support from the Institute for Disease Modeling.
Seattle Flu Study researchers were the first to detect COVID-19 in the US in January, before hitting regulatory hurdles that temporarily thwarted their work.
Here are more details from today’s announcement.
In its first 18 days, SCAN tested 4,092 samples. Nearly two-thirds of those were returned by individuals who in the seven days prior to enrollment had reported COVID-like illness (fever, cough, or shortness of breath). However, more than three-quarters of these individuals indicated they had not yet sought medical care. SCAN is the first COVID-19 surveillance program in the U.S. to use “swab-and-send” test kits that allow individuals to collect their own nasal sample and return it to a lab for testing without leaving home—while observing physical distancing guidance and reducing exposure to others.
SCAN testing among those reporting COVID-like illness (CLI) returned 44 (1.6%) positive results for COVID-19—a proportion lower than that being returned through testing within the medical system, but one that may still represent thousands of unrecognized infections in the community.
Cases were identified in all parts of King County. Based on this participant and county census data, SCAN modeling estimates that community prevalence of COVID-19 for the period is between 5 and 75 out of 10,000 people (0.05% to 0.75%), with best estimate of 24 per 10,000 (0.24%).
The initiative is designed as an infection surveillance program and is not meant to be a clinical service, researchers caution. SCAN leaders say they are working to expand the program and expect to provide another update in two weeks, on Friday, May 1.
