scorecardresearch Skip to main content

How far has the coronavirus spread? The answer may be in the sewers

The Massachusetts Water Resources Authority's sewage treatment plant on Deer Island.David L. Ryan/Globe Staff/file

As more states look for ways to reopen, they’re hampered by the lack of sufficient testing for the coronavirus, making its prevalence impossible to gauge.

So public health officials are looking at another way to measure the spread of the contagion: by examining raw sewage.

A Somerville company working with Harvard University, MIT, and Brigham and Women’s Hospital recently conducted a study of sewage on Deer Island that suggested stool samples from waste-water treatment plants could provide a more accurate picture of the extent of the disease than clinical testing alone.

Results from the study, conducted in late March and recently published by the online archive medRxiv, found that the number of actual infections was probably at least five times higher than confirmed cases at the time.


“The ability to capture so much information on so many people, so quickly, is really valuable,” said Newsha Ghaeli, president of Biobot Analytics, which analyzed the sewage from Deer Island. “While we should be pushing clinical testing as we reopen, this allows us to see how the virus is trending, and it could provide a valuable early warning for the reemergence of the virus.”

The company, one of only a few testing waste water for the virus, is now collecting samples from hundreds of treatment plants around the country in hopes of identifying emerging outbreaks, especially in areas with less clinical testing.

Each treatment plant has been asked to collect samples of untreated sewage over a 24-hour period, and then send the samples to Biobot Analytics, which has been analyzing them for free.

Ghaeli said the company plans to release the results of its broader analysis soon, but preliminary results support the earlier findings from Deer Island.

“Where we’re seeing hot spots are hot spots that have been confirmed by cases,” she said. “But what’s interesting is that we’re able to see an order magnitude more than the confirmed clinical cases. Even in areas with few confirmed cases, we’re seeing more.”


For example, samples from a waste-water treatment plant in New Castle County in Delaware led the company to estimate that more than 15,000 people had likely been infected with the virus by mid-April, about 15 times the number of cases confirmed there at the time.

Similar results were found in Linn County in Oregon, where sewage samples suggested as many as 730 people had the virus late last month — 10 times the number of positive tests. At one point last month in Miami-Dade County in Florida, the company estimated there had been nearly eight times more infections than testing revealed.

While there are clear benefits to trying to measure the prevalence of the virus from sewage — it’s cheaper, faster, and easier than testing large numbers of people — there are significant uncertainties and potentially large margins of error.

Sewage analysis, for example, must estimate how much the average person defecates per day, and how much genetic material from the virus should be found in infected stool. The value of the analysis also hinges on whether the amount of virus per sample varies by how sick someone becomes and other unknown factors.

The company has been refining its assumptions after its March analysis of sewage from Deer Island estimated that anywhere from 2,300 to more than 115,000 people were infected. At the time, only 446 cases had been reported in the Boston area.


Dr. Eric Alm, a professor of biological engineering at MIT who cofounded Biobot Analytics, said the company now believes the number of infections from the Deer Island data was much closer to the lower side of the estimate.

“We’re still potentially estimating too many people, but we expect this to become more accurate,” he said.

Still, epidemiologists and others see promise in testing sewage to gauge whether infections are rising or falling, giving policy makers a clearer picture of when they can safely relax restrictions.

After convening a recent online summit of public health and water-treatment experts to discuss the value of such analysis, Peter Grevatt, CEO of the Water Research Foundation, a Denver group that studies water quality, called it “a very promising approach.”

While he acknowledged the uncertainties of the data, he cited findings in the Netherlands and elsewhere that suggest a strong correlation between the amount of the virus in sewage and the actual number of infections.

“This can be a very important tool,” he said.

Rolf Halden, who conducts similar research as director of the Center for Environmental Health Engineering at Arizona State University, called the analysis of waste water “a critically important tool” to managing the pandemic and reopening society.

Before the emergence of the coronavirus, his team had used similar technology to track opioid use in different communities.

“The cost savings of using this approach are huge,” Halden said. “Testing waste water can be up to a million times cheaper than screening individuals. It has to get done, and has to get done quickly.”


The fact that the virus can be tracked in sewage has raised concerns that untreated waste water could serve as a means of transmission, especially when it overflows from sewers during heavy storms, a perennial problem in urban areas.

After heavy rains, for example, clumps of fecal matter have often found their way into major waterways in the region, such as the Merrimack River, which provides drinking water to more than 600,000 people and is used for fishing, kayaking, and other recreation.

“We are extremely concerned that this is a potential public health threat,” said John Macone, codirector of the Merrimack River Watershed Council.

He noted that some cities in the watershed, such as Lawrence, Lowell, and Haverhill, have had relatively high infection rates, although there are other reasons for that. He and others worried that the risks from untreated sewage could rise as summer approaches.

“We are seeing considerably more human activity on the Merrimack, with docks being installed, boats being launched, and fishermen going after migrating fish,” he said.

If sewage can transmit the virus, older cities such as Boston would be particularly vulnerable.

“We have ancient, sometimes cracked, sometimes clay sewer pipes, and virtually every rainstorm . . . can lead to massive releases of raw sewage,” said Kyla Bennett, director of Public Employees for Environmental Responsibility in New England, an advocacy group.


Officials at the Centers for Disease Control and Prevention said it remains an open question whether untreated sewage could be spreading the virus.

But Kate Fowlie, a spokeswoman for the agency, said the risks of such transmission are “thought to be low.”

“There is no evidence to date that this has occurred,” she said.

Fowlie added that the agency is exploring the potential for waste-water testing but has so far deferred to private companies.

David Abel can be reached at Follow him @davabel.