COVID-19 Data Dispatch

Tag: Biobot

  • Interpreting COVID-19 data as the CDC goes weekly and a fall surge approaches

    Interpreting COVID-19 data as the CDC goes weekly and a fall surge approaches

    As of this week, the CDC has switched to reporting COVID-19 cases and deaths on a weekly basis.

    As of this Thursday, the CDC is updating COVID-19 case and death data every week instead of every day. Here are some thoughts on interpreting COVID-19 data in the wake of this change, citing an article I recently wrote for The Atlantic.

    To me, the CDC’s shift to weekly updates feels like the end of an era for tracking COVID-19. While I understand the change, considering both our less-complete case information and other data analysis needs for the agency, I can’t help but wish we had a national public health agency with enough resources to continue providing us with frequent, reliable information on this ongoing pandemic. After all, shouldn’t that be the CDC’s job?

    The CDC has clearly deprioritized two major metrics (cases and deaths) that used to be the first places people looked to see the pandemic’s impact on their communities. Instead, the agency now points us to hospitalization metrics, variant surveillance, and wastewater—all metrics that are certainly useful, but may be harder for the average user to interpret.

    And even the case data we do have are quite unreliable at this point, as PCR tests become less and less accessible compared to rapid tests. Case numbers may be underreported by twenty times or more; it’s difficult to even get a good estimate of how far off the numbers are. Public communications like the CDC’s “Weekly Review” report fail to acknowledge this problem, and the agency does not appear to be making any effort to determine the true infection rates right now. 

    Through its current data communication choices, the CDC seems to be saying, “If you still care about keeping track of COVID-19, you’re on your own.” Even though we are likely heading for a fall surge and many people need to keep track of this disease in order to keep their communities safe.

    In absence of useful information from our public health leaders, it falls on us to survey the best available data sources and help others interpret them. My article in The Atlantic takes on this question, focusing on wastewater surveillance and population surveys as particularly useful sources we should consider right now.

    Interpreting wastewater data

    Wastewater data, unlike case data, don’t require people to actively go out and get tested: if their public sewer system is getting sampled for COVID-19, they will automatically be included in the data. You can look for wastewater surveillance in your area on Biobot’s dashboard, the CDC’s dashboard, or other state and local dashboards, depending on where you live. (I have a list of state dashboards here.)

    But interpreting wastewater data can be pretty different from interpreting case data. Here are a couple of key tips for approaching this source, based on my interview with Biobot president and cofounder Newsha Ghaeli:

    • Look at “directionality” and “magnitude.” “Directionality” means whether viral levels are going up or down, and “magnitude” means how they compare with earlier points in the pandemic. To quote from the story: “A 10 percent uptick when levels are low is less concerning than a 10 percent uptick when the virus is already spreading widely.”
    • If you do not have public wastewater data for your county, data from a neighboring county still provides useful info. When we talked, Ghaeli gave the example of a New York City resident looking at data from New Jersey or Connecticut counties neighboring the city: as people from these areas commute into NYC, a surge in one place could quickly drive a surge in the other.
    • Wastewater data are not a perfect proxy for infections. Scientists are still learning about how to best use this newer surveillance tool. Unlike clinical metrics (like cases), wastewater data can differ based on local environmental factors, and it often takes a long time for researchers to build useful interpretations of their communities. (See my past FiveThirtyEight story for more detail on this.)
    • These data can’t tell you who is getting sick. To comprehensively answer demographic questions, we need to actively survey people in our communities and ask them about their experiences with COVID-19. (See the story for more about how this works.)

    Other interpretation tips

    Beyond looking at wastewater data, here are a couple of tips I received from experts for readers seeking to watch their local COVID-19 numbers this fall:

    • “Look as local as you can,” said Pandemic Prevention Initiative expert Sam Scarpino. In other words, if you can find data for your individual county or even ZIP code, go there. 
    • Check multiple sources, and try to “triangulate” between them, said City University of New York epidemiologist Denis Nash. (I’ve provided similar advice in past posts like this one.)
    • Consider local events and behavior, Scarpino said. Quoting from the story: “If a popular community event or holiday happened recently, low case numbers might need to be taken with a grain of salt.”

    I also wanted to give a quick shout-out here to the People’s CDC, a volunteer science communication and advocacy organization. If you’re looking for more thoughtful analysis of national COVID-19 data, their weekly “weather reports” are a really helpful and accessible source. The organization also provides resources to help people push for more COVID-19 safety measures in their communities. 

    “People do want layers of protection, they do want to keep themselves in each other safe from COVID,” said Mary Jirmanus Saba, a geographer and volunteer with the People’s CDC whom I interviewed (with a couple of other volunteers) for my story. The weather reports and other similar initiatives help the organization’s followers “see that we really are there for each other,” she said.

    More wastewater reporting

  • Sources and updates, October 2

    • Johns Hopkins dashboard creator wins public service award: Lauren Gardner, an engineering professor at Johns Hopkins University, was recently awarded the 2022 Lasker-Bloomberg Public Service Award (a major prize in biomedical research) in recognition of her work on JHU’s global COVID-19 dashboard. This dashboard was one of the world’s first and most popular sources for tracking how the pandemic spread. Unlike many other projects, it has continued fairly consistently since early 2020, and continues to be a great resource for national and international data. Congratulations to Gardner and the other folks at JHU!
    • CDC releases updated chronic disease and risk factor data: This week, the CDC published a new iteration of its Behavioral Risk Factor Surveillance System (BRFSS), a major data source providing information on chronic conditions, health behaviors, access to healthcare, and more. The surveillance system uses surveys of over 400,000 American adults, conducted annually in all 50 states and several territories. While these aren’t COVID-specific data, the datasets can be a really helpful source for examining populations more vulnerable to COVID-19 in different parts of the country.
    • Increased respiratory illnesses in children: Another CDC update: researchers from the agency published a new study in the Morbidity and Mortality Weekly Report reporting increased cases of respiratory illness in kids this past summer. Strains of rhinovirus and enterovirus that haven’t circulated much in the last two years are back in 2022 and could cause problems this fall—especially as schools continue to operate in-person with relatively few public health measures—the CDC report suggests. For more context, see this recent newsletter by Katelyn Jetelina and Caitlin Rivers.
    • Biobot and CDC expand wastewater tracking to monkeypox: Biobot, the leading COVID-19 wastewater surveillance company, is expanding its work with the CDC to include monkeypox surveillance. As part of the CDC’s National Wastewater Surveillance System (NWSS), Biobot will coordinate data collection and analysis for both COVID-19 and monkeypox through at least January 2023. “We hope this can demonstrate the flexibility and versatility of this technology for governments across the country,” Biobot president and cofounder Newsha Ghaeli said in a press release.
    • Launch of the Data Liberation Project: This is not COVID- or even health-specific, but I wanted to give a quick shout-out to the Data Liberation Project, a new effort by Jeremy Singer-Vine (widely known in data journalism circles as the author of the Data Is Plural newsletter). The new project is “an initiative to identify, obtain, reformat, clean, document, publish, and disseminate government datasets of public interest.” I hope to see some COVID-19 datasets liberated through this project!

    All featured sources
  • National numbers, October 2

    National numbers, October 2

    Coronavirus levels in Boston, Mass. wastewater spiked intensely last week. Chart via MWRA/Biobot.

    In the past week (September 24 through 30), the U.S. reported about 330,000 new COVID-19 cases, according to the CDC. This amounts to:

    • An average of 47,000 new cases each day
    • 100 total new cases for every 100,000 Americans
    • 13% fewer new cases than last week (September 17-23)

    In the past week, the U.S. also reported about 26,000 new COVID-19 patients admitted to hospitals. This amounts to:

    • An average of 3,800 new admissions each day
    • 8.0 total admissions for every 100,000 Americans
    • 6% fewer new admissions than last week

    Additionally, the U.S. reported:

    • 2,400 new COVID-19 deaths (350 per day)
    • 81% of new cases are caused by Omicron BA.5; 13% by BA.4.6; 3% by BF.7;  1% by BA.2.75 (as of October 1)
    • An average of 400,000 vaccinations per day

    Official COVID-19 numbers continue to drop nationwide, with case counts down 13% and new hospital admissions down 6% this week compared to the prior week. Still, signals from wastewater suggest this is no time to let our guard down, especially if you live in the Northeast.

    Biobot’s wastewater surveillance dashboard is back this week (after a one-week hiatus). Nationally, this surveillance suggests coronavirus transmission is at a high plateau close to what we saw during early fall of last year, before Omicron hit. The Northeast is driving that recent trend, with an overall coronavirus concentration twice as high as the concentrations reported in other regions.

    This region remains a hotspot for Omicron BF.7, the subvariant of BA.5 that could be the U.S.’s next dominant lineage. Nationwide, BF.7 is slowly competing with BA.5: it’s grown from causing about 1% of new cases to 3.4% over the last month, according to CDC estimates. BA.4.6 also continues to grow, while BA.2.75 has remained relatively constant.

    Within the Northeast region, Boston stands out: the city’s wastewater surveillance program (run by Biobot) reported a major spike last week. And by major spike, I mean an increase of more than 100% week-over-week, according to the city’s public health department. “This spike in our wastewater concentration is of great concern and another reminder that the pandemic is far from over,” said Dr. Bisola Ojikutu, Boston’s public health commissioner, in a press release.

    Wastewater spikes typically precede case spikes by a couple of weeks, though we’ll have to see whether the significant drop in PCR testing in recent months changes this pattern. Either way, this is a good time to get a booster shot (more on boosters below), stock up on masks and rapid tests, and start planning safety measures for the holidays.

    Boston is running vaccine clinics in response to this potential new surge, and the city public health department “recommends” masking indoors. But any further mitigations are likely out of the question, even though they could have a huge impact.

    more national numbers

  • COVID source callout: Wastewater data on a break

    If you went to check Biobot’s COVID-19 wastewater dashboard this week, you may have noticed that the company hasn’t updated its data since September 14.

    Biobot’s website doesn’t provide an explanation for the lack of updates, but you can find one on another source: the wastewater data page run by the Massachusetts Water Resources Authority (MWRA), covering the greater Boston area. As Biobot was founded in Boston (by scientists at the Massachusetts Institute of Technology), the city is one of Biobot’s longest-running COVID-19 sampling sites.

    As of September 25, a note on the MWRA page reads: “Biobot will be closing their lab next week for a company retreat. We don’t expect any data updates between 9/18/2022 and 9/25/2022. They expect to catch up to the normal posting schedule by 9/27/2022.”

    It seems logical to assume this lab closure also applies to the company’s national dashboard. 

    Now, to be clear, I have nothing against Biobot for taking a company retreat; anyone working in public health deserves a break sometimes, and I hope the retreat is productive for the company’s staff. But when a private company represents the best source for a crucial pandemic-tracking metric—as Biobot is able to provide unified, national and regional estimates better than the CDC can—even a week-long break can be a problem for people who rely on these data.

    Ideally, there should be enough public health funding going around that everyone working in wastewater surveillance can take all the breaks they need without disrupting data systems. This is just a small example of a much bigger issue.

    more source spotlights
  • Wastewater data gap follow-up: an update from Biobot

    Wastewater data gap follow-up: an update from Biobot

    Quite a few sites on the CDC NWSS dashboard are still not reporting recent data, but Biobot is working to bring them online.

    Last week, I pointed out a data gap on the CDC’s National Wastewater Surveillance System (NWSS) dashboard: hundreds of sewershed sites on the dashboard have not been updated with recent data in weeks.

    In this post, I hypothesized that the gap was likely a result of a shift for NWSS, as the CDC has switched from a contract with testing company LuminUltra to a contract with Biobot. In both cases, the outside company had been hired to conduct wastewater sampling and analysis for sites that don’t have capacity to do this themselves; as NWSS transitioned between testing providers, some sites were left without recent data.

    This week, I’m excited to share an update on the situation, courtesy of Becca Malizia, Biobot’s science communication manager, who reached out after seeing last week’s post. Below, you’ll find more details on the recent data gaps and Biobot’s new relationship with the CDC.

    Malizia confirmed that the transition between LuminUltra and Biobot has led to delays in wastewater data availability for some sites in the NWSS network. She pointed me to a footnote on the CDC dashboard, labeled “May 2022 Coverage Limitations”:

    Beginning April 15, 2022, approximately 150 wastewater sites in 29 states began transitioning to a new wastewater testing provider. During this transition, these sites will not have recent data displayed and will be colored gray on the map. It will take several weeks for enough data to be collected to calculate the metrics displayed on this page. Results for these sites are expected to be available again between mid-May and June 2022.

    Now, if the CDC was going for full transparency and ease of dashboard interpretation here, the agency should have placed this important note somewhere more obvious to the average user—not buried at the bottom of the page. But I’m glad to see this public information, including the estimate of when results for the transitioning sites will be available.

    Biobot also acknowledged the data gap in a Tweet on May 19, explaining that the company has experienced issues in distributing testing kits to participating sites:

    Further contributing to data delays, the sewershed sites for which Biobot is now in charge of sampling require a “minimum number of samples” before key metrics on the CDC NWSS dashboard can be calculated, Malizia said. The calculations for these metrics change from one lab to another, so Biobot needs to use data from its protocols rather than data from the prior LuminUltra contract.

    “Sites in the process of switching over from the previous contract may have some lag until there is enough data to do the calculations for the CDC metrics,” Malizia wrote. She also pointed to several other reasons why a sewershed site might collect wastewater data, but not have its data appear on the CDC dashboard, such as: a sewershed serving under 3,000 people, a sewershed serving a specific institution (like a college campus), and a sewershed where local leaders have elected not to send data to the CDC.

    Biobot has already onboarded more than 200 wastewater treatment plants, Malizia said; this includes sites that were previously included in the NWSS contract with LuminUltra, though a full list of those sites is not publicly available. By mid-July, Biobot aims to have 500 sites participating in its CDC program. The company works with state and local health departments to select wastewater sampling sites and coordinate with treatment plants.

    Now, it’s important to note that, outside of its CDC NWSS contract, Biobot coordinates wastewater testing and analysis for hundreds of sewershed sites through a program called the Biobot Network. This program is a public service offering from Biobot: the company does testing and analysis at no cost to sewersheds. But Biobot also does not send individual, sewershed-level data back to the participants. “Rather, the data is aggregated at the county level on our public dashboard (biobot.io/data) for the benefit of policymakers and the general public,” Malizia explained.

    (The free Biobot Network includes two sampling sites in Hillsborough, Florida which used to be paid Biobot customers in 2021, Malizia said, in comments responding to the Tampa Bay Times article I quoted last week. The final reports those sites received when they were paying customers were in August and October of last year. “Individual wastewater treatment plants can choose whether or not to share these reports with local government agencies,” Malizia said.) 

    Before it became a CDC contractor, Biobot was not able to submit wastewater data to the agency. Only state government agencies have access to the CDC NWSS platform used for data reporting, Malizia said; the CDC has made an exception for Biobot under the new contract. To me, this helps explain why there’s not a lot of overlap between Biobot Network sites and CDC NWSS sites, as well as why some other wastewater sampling (done by universities, research institutions, etc.) does not appear on the CDC dashboard.

    But, now that Biobot is a CDC contractor, will the company provide Biobot Network data to the agency? I asked Malizia this question, to which she responded:

    The Biobot Network will remain separate from the NWSS, however sites enrolled for the CDC NWSS Program will also be given the option to opt into the Biobot dashboard.

    In summary: Biobot is working hard to restore data from sites already in the CDC NWSS network and expand that network to more sites that don’t have capacity for wastewater sampling on their own. However, thanks to a combination of CDC bureaucracy and complex public and private data systems, it seems unlikely that we will get a singular dashboard including all wastewater testing sites in the country anytime soon.

    Also, Biobot’s current contract is only for nine months. Are we going to see another round of data gaps next winter, if the CDC decides to switch wastewater testing companies again?