Last updated 4/12/2022

Intro

Much interest has recently been given to the BA.2 variant of Omicron. This variant appears to be substantially more transmissible than earlier variants of Covid, including earlier variants of Omicron. This transmissability is leading to surges of Covid cases in many countries across the world.

Despite this transmissibility, the United States is yet to see a BA.2 surge, and the country is enjoying its lowest case rate since early July, 2021.

One easy explanation for this lack of surge is that the United States is yet to see substantial BA.2 cases. This, however, is challenged by CDC data that estimates that roughly half of cases for the week ending 3-26 were BA.2.

Another explanation is that earlier Omicron infection provides significant protection against BA.2, and the United State’s earlier Omicron surge is therefore protecting against BA.2. France, however, does not line up with this hypothesis. France had a larger Omicron surge than did the United States, and they are now experiencing a large BA.2 surge.

With the easy explanations ruled out, I aimed to plot the data to see if I could shine light on what’s happening.

Setting up the analysis

This analysis uses the tidyverse family of packages.

library(tidyverse)
## -- Attaching packages --------------------------------------- tidyverse 1.3.1 --
## v ggplot2 3.3.5     v purrr   0.3.4
## v tibble  3.1.5     v dplyr   1.0.7
## v tidyr   1.1.4     v stringr 1.4.0
## v readr   2.0.2     v forcats 0.5.1
## -- Conflicts ------------------------------------------ tidyverse_conflicts() --
## x dplyr::filter() masks stats::filter()
## x dplyr::lag()    masks stats::lag()

Relevant data comes from two datasets. First, I use the Our World in Data covid dataset and focus on American cases.

USA_cases <- read_csv("https://covid.ourworldindata.org/data/owid-covid-data.csv") %>% 
  filter(iso_code == "USA")

Next I use the CDC’s estimates of variant prevalence. I filter for nation-wide estimates, and make sure dates are in a proper date format.

USA_variant <- read_csv("RegionsDashboard.csv") %>% 
  janitor::clean_names() %>% 
  filter(usa_or_hhsregion=="USA") %>% 
  mutate(date = lubridate::mdy(day_of_week_ending))

I combine the variant and case datasets. The case count for each variant is inferred by multiplying the CDC’s point estimate of variant share with the total cases in the country.

USA_variant <- USA_variant %>% 
  left_join(USA_cases, by = c("date")) %>% 
  mutate(variant_count = share*new_cases_smoothed)

Plot

Here, I plot the frequency of each variant, along with the total national case count (in black).

plot <- ggplot(USA_variant, aes(x = date, y = variant_count, color = variant, group = variant)) +
  geom_line() +
  geom_point(size = 0.5) +
  geom_line(aes(y=new_cases_smoothed), color = "black") + 
  geom_point(aes(y=new_cases_smoothed), color = "black", size = 0.5) +
  labs(title = "USA total covid cases and variant cases across time") +
  xlab("Week ending...") +
  ylab("Case Count") +
  scale_color_discrete(name = "Variant") +
  theme_minimal()

And we can see an interactive version of this plot:

plotly::ggplotly(plot)

Discussion

It is hard to see from the full plot, but zooming in on the end of the plot reveals something interesting. While the overall case rate is currently flat, this result is a function of two phenomena occurring at approximately the same rate: 1) the decline of the BA.1.1 Omicron variant and 2) The increase of the BA.2 Omicron variant.

As BA.1.1 is quickly approaching 0, it is likely that the growth of BA.2 will soon be the predominant force in explaining US case rates. Thus, I expect to see at least a slight increase in US covid cases in the coming weeks.

Update 4/5/2022

As of April 5, 2022, the CDC has made data available up through the week ending 4/2. The basic trend has continued: BA.2 cases are rising and Omicron cases are declining at similar rates. This can be seen in the following plot:

However, BA.2 growth rate has not intensified. If anything, it has slowed. This may suggest that the BA.2 “surge” is nearing its peak.

If, indeed, BA.2 is nearing its peak, the US may avoid another substantial Covid-19 case wave at this time.

Update 4/12/2022

Signs continue to point towards this being a strange moment of the pandemic in the United States. Per OWID data, the country recently reached a new low in hospital admissions, albeit with a small, recent uptick.

However, BA.2 cases continue to rise. Whereas last week there was some indication that BA.2 may have been plateauing, I do not see evidence of that in this week’s data. At this stage, the CDC estimates over 85% of cases are BA.2, so the case trajectory will be almost fully explained by BA.2.

This slight uptick in cases is further corroborated by recent increases in waste water virus load found in many regions of the country. Total cases will more than likely rise over the next week.