The COVID Variant Tracking ProjectPosted on
Just as we are developing vaccines for the novel Coronavirus that causes COVID-19, concern has rightfully exploded over multiple COVID Variants that have appeared on the screen. To track these Variants, in you might say, a COVID Variant Tracking Project or effort, we’ve started to compile data to recalculate the R value for individual variants across the country on a state-by-state level. One challenge is that some states release data daily while others do so on a weekly basis. Thus far data has been released for 3 days all within a week – 3/25, 3/28, and 3/30. From a quick glance, the B.1.1.7 (UK) variant data from California appears alarming: 471 cases on March 25th, 563 cases on the 28th, and 822 cases reported as of yesterday, March 30th. This represents rapid growth for a short time period and on track to double in less than a week, possibly in 5 days (pending reporting errors). The current R value for COVID in California is .94 and rising, comparing that to the (albeit preliminary) calculated R value for the B.1.1.7 variant of between 1.2 and 1.5. A value of 1 is the threshold value that suggests disease spread and growth (R>1), versus a disease in regression (R<1).
The concern is that California may be on its way to a Third Wave as the R value changes while the B.1.1.7 variant takes hold. We will look into what this may foretell for the rest of the country and other nations hit by this variant.
As data regarding the new variants continues to evolve, it’s interesting to see how the Center for Disease Control and Prevention (CDC) is releasing information. From a scan of the CDC’s web page titled: US Covid-19 Cases Caused by Variants, the CDC is doing two very interesting things:
- Releasing the data only 3 days a week – Sunday, Tuesday, and Thursday – is odd because we have become used to daily case reporting throughout the pandemic.
- The more important distinction is that these variant-case numbers are based off a sampling of cases from the reporting states, rather than a true representation of what’s going on.
Is it okay that the CDC is only taking a sampling of the case numbers, without disclosing sample size? Or whether the sample is a fair representation of the overall population? Does limiting the number of days cases are being reported help or hurt efforts to gain real visibility into the pandemic and ultimately defeat it?
Variant reporting is still not ubiquitous. As of this date (4/7/21) only 25 states and territories are reporting on the P.1 (Brazil) variant, for instance.
This is extremely disheartening as on 3/25 there were 79 reported* cases in the US of the P.1 variant, as of 4/7 there are 356 cases. So, of this sampling of unknown size the P.1 variant has grown 4.5x in under two weeks. While the current R value for COVID overall is low, these variants have the ability to raise the R value quickly as they take hold.
However, without further information on the sample size or process, we can’t immediately draw conclusions from these numbers.
Number of states reporting on COVID variants
At least the trend in the number of states reporting on variants is headed in the right direction.
As mentioned above, the CDC is releasing COVID variant data three times per week through a sampling process. With regards to the sampling size used to calculate the number of cases of variant strains, there is a finite number of serology tests that can be done as they also take time and resources. The CDC has the capacity to preform 750 serology test a week, while contracted commercial labs can perform around 6,000 tests a week. Having the ability to sample ~7,000 COVID cases weekly for variance is a solid start, but with the US averaging over 50,000 cases a day this sampling is close to only 2% of the total weekly cases. This could mean that the variants are the dominant strains in America right now, from 4/1/21 to 4/8/21 there were 7,049 cases of B.1.1.7 reported. Total cases found of variant strains from 4/1 to 4/8 were 7360 – meaning 95.78% of US COVID identified through serological testing were due to the UK variant. The balance were 210 cases of P.1 and 101 of B.1351.
If we assume these 7,09 to represent 2%, this means we may have over 350,000 possible cases of just the B.1.1.7 variant in the country if the genetic testing was possible for each case.
Going forward, especially as COVID may continue to mutate, it may be helpful for the CDC to test and report for the presence of more than the Big 3 variants. There are 19 named variants, and an additional 200 potential lineages under observation. The almost complete domination of COVID by the UK variant begs the question if Americans have developed some sort of immunity to the original COVID-19 strains? And would this be actually a good sign? Could we already be in the third wave and have no idea, as states are lifting mask mandates and social distancing? Could the rapid rise of the P.1 variant be attributed to the border crisis currently evolving?
Following up on our April 9th post, the CDC has now stopped reporting on variant cases specifically but rather has shifted to proportions of total cases in the states that are reporting. They [CDC] briefly explain how they came to these numbers. At the time of writing this the CDC has completed a total of 41,000 serology tests of Covid cases over a four-week period, in that same time the United States reported over 1.6 million new cases of Covid. This represents 2.6% of cases being run through serology tests. According to the CDC website (https://covid.cdc.gov/covid-data-tracker/#variant-proportions) , “Proportions in the table [presented] are only shown for states for which CDC has at least 300 sequences from specimens collected during this timeframe. Proportions are calculated using empirical (unweighted) data, which are subject to change over time and will be updated as more data become available. Proportions of variants do not represent the total number that may be circulating in the United States and may not match cases reported by states, territories, tribes, and local officials. For states and jurisdictions not listed, CDC has insufficient genomic surveillance data for the specified time period.” 
I think that what we are learning is that this pandemic is able to mutate (evolve) almost as fast – or in some cases faster – than we are able to develop countermeasures such as vaccines. Scientist try to keep up with new information, yet the information is not always disseminated in the best and most useful manner. As of the weekend, the CDC ceased reporting on the number of new variant cases, could this be a sign of inaccurate case reporting? At the time, 31 states were reporting on the P.1 and 36 on the B.1.351 variant. The CDC switched from reporting specific cases to a reporting of the proportions of new cases by strain of Covid. In this new reporting format, only 39 states are present and of those 39 not all of them are reporting on the 5 separate fields currently being tested (B.1.1.7, B.1.351, B.1.427/429, other lineages, and P.1). The variants have been discovered for some time now, but their tracking has started only more recently. In an ideal world, we would have every single covid cases run through genomic surveillance to know the true prevalence of the variants/strains currently in our country. We need transparency and the ability to process this information ourselves to better see what’s truly going on in the states – and at the local level.
As of Monday, April 12th, the CDC has updated their variant case percentages from 3/13 to 3/27, this update shows the massive amount of growth that is currently being experienced. As of 1/16/21 the B.1.1.7 represented 3.1% of cases being experienced, as of the most recent update on 3/27/21 the variant now represents 44.1% of total cases, albeit based off an admittedly small sample size of roughly 41,000 genomic tests. As these numbers continue to rise, it begs the question if we are entering a new wave? The seven-day moving average of COVID cases is currently rising once again, with some states reporting cases at the same level as during the second wave. This may well be indicative of the new strains due to their increased virality and spread as compared to the original strain we have seen for a little over a year. Are the variants less deadly? Or the vaccines are working? Or are our natural immune systems – which also evolve – building a natural immunity? COVID won’t be novel forever, after all. As more and more people are getting vaccinated this could be limiting the strain’s fatality, but not its transmissibility.
Check back for updates, and follow our LinkedIn page for the latest news on our COVID response efforts.
– by Scott I Gershman, Data Analyst, and Jay Etherton, Public Health Analyst