What accounts for the recent drop in COVID-19 cases in the US?

Data and Metrics Infection and Spread

Q: It looks like Covid-19 infection rates have dropped significantly in the US since the vaccine became available. Is the vaccine already having an impact? What else would account for the drop?”

A: Of all the possible explanations — vaccination, immunity from prior infection, behavior change, seasonality, and less testing — the best answer seems to be: Likely a little bit of everything! None of those factors could, on its own, produce the steep drop in cases that we’ve observed.

[Updated Feb 18 pm: Nerdy Girl Dr. Jenn Dowd just published a longer piece in Slate on this same question – link below!]

First, let’s do the numbers. How fast are new cases dropping in the US? Pretty darn fast: 7-day average new cases as of February 17 was 77,567, down from the peak 7-day average on January 11 of 248,725 — a 69% drop in a little over 5 weeks! We haven’t had a 7-day average new case count this low since October 28.

Let’s run through possible explanations, starting with Eileen’s initial question:

1️⃣💉Vaccination: While over 56 million doses have been administered as of today, back on January 11 (when cases started to drop) that number was just under 9 million — likely not widespread enough to bring down overall case numbers this quickly. In the country with the most vaccine doses given per 100 people — Israel — cases continued to climb until vaccination rates were much higher than they are in the US currently. That said, vaccinating health care workers and vulnerable older adults is likely helping depress transmission.

2️⃣ 💪 Immunity from prior infection: Is it possible that enough people have already had COVID-19 that there are fewer available “susceptibles” to get infected? Experts estimate that somewhere in the range of 20-40% of US residents have already had COVID-19, with some amount of protection offered against subsequent infection. If those who have already gotten COVID-19 were more at risk (through social behavior, occupational exposure, etc.) than those who haven’t gotten COVID-19 (due to staying at home, masking, etc.), that would be make it even harder for the virus to find new susceptibles over time, and each new case would infect fewer people.

3️⃣ 😷 Behavior change: No one was surprised to see cases peak about 2 weeks after the Christmas and New Year’s holidays. The drop since then may reflect people traveling less, socializing less, and staying at home with their pod. The very scary numbers in early January plus new guidance on masking and a new emphasis on masking from the Biden administration may have improved masking behavior.

4️⃣ ☀️❄️ Seasonality: While many viruses show strong seasonal patterns, it turns out that even the experts aren’t exactly sure what drives seasonality. It’s likely the product of both temperature and humidity’s effects on the virus itself, as well temperature and humidity’s effects on the social interactions that drive transmission (like, we all go inside when it’s cold.) So far, there’s not strong evidence for seasonality being the main driver of our drop in cases, but it might be contributing.

5️⃣ 👩‍🔬Less testing: Uh oh. Are we only observing fewer cases because we’re testing less? It’s true that tests per day are down about 25% from a peak of over 2 million in mid-January. But what’s chicken and what’s egg here? Fewer tests because fewer cases? Or fewer (observed) cases because fewer tests? There are two clues that the observed drop in cases is not because we’re testing less: test positivity (the percent of all tests done that come back with a positive result) is not going up (if it were, that would indicate we’re missing a lot of positive cases), and hospitalizations are also going down, fast (good news!).

Are we starting to see the impact of vaccination? Maybe a smidgen, with more to come soon, we hope. We also think prior disease is leaving fewer susceptibles available to be infected, and people’s behavior (supported by new policies) is keeping the virus in check. Seasonality and testing rates are less likely to be responsible here.

Stay tuned, and keep using those SMARTS!

Helpful Links:

Nerdy Girl Dr. Jenn Dowd answers this same question on Slate

Dr. Natalie Dean’s twitter thread

Derek Thompson in the Atlantic

Reis Thebault in the Washington Post

Our Dec 12 post on “When will the numbers start going down?”

Link to original FB post