I Just Got My Vaccine. Am I Immune Yet? What Data from Vaccine Trials Suggests
Whether you’re wondering about your immunity in the first few days after your vaccine, if your past infection gives you immunity, or whether we’ll reach herd immunity, I put together 5 plots to give simple, visual, and memorable answers to these questions.
When am I immune?
Q: I just got my vaccine. Am I immune?
For Johnson and Johnson, here are the reported vaccine efficacy percentages over time since the dose (interpret these as the percent reduction in risk, from your baseline risk of getting COVID). [1]
The JnJ vaccine takes a couple of weeks before immunity really starts kicking in, although see the footnote explaining why this might be a conservative estimate.
For Pfizer, immunity seems to kick in much faster, getting to 50% after around a week:
Similarly for Moderna, the vaccine reaches 50% efficacy after around a week:
When am I infectious?
Q: I just got exposed to someone with COVID. Can I pass it on?
First of all, you should think about how high risk your exposure was. Were you indoors, close, talking or exercising, and together for a long time? Has this person infected other people? These questions are most important in understanding your risk of having contracted it.
If you did, here’s what your infectiousness will look like. For 48 hours, you can’t transmit it. The virus is still incubating. After that, you’ll probably quickly become your most infectious, peaking around 5 days after exposure. This is when you’re most likely to develop symptoms. From there, you’ll get less and less contagious until you’re healed. I interpreted this narrative with my plot. This is somewhat guesswork and there is a lot of variation between patients. The most important things to know are that the first two days after exposure are safe, and the four days bookending the start of symptoms are the most unsafe.
Here was my methodology: The first two days have to be zero [5]. Symptoms usually start around day 5 [5]. Viral load has to peak around then. A huge meta-analysis of viral load gives a pretty solid estimate of viral load from that point on. So I started at zero, filled in the initial spike to peak between days 3 and 5, and then copied the graph of viral load from the meta-analysis.
Here’s the ugly graph in R:
And a prettier graph I drew:
Also, around days 5 and 6 is when there becomes a greater than 50% chance that a PCR test confirms you have COVID [5]. This is the thing with exposures — you have to wait a while before a test will confirm that you have COVID. Even 5 or 6 days after exposure, your negative test has about a 50% chance of being false.
Where am I at risk?
Q: Should I wear my mask outside? Etc.
Indoors, low ventilation, unmasked, close to a known spreader who is still infectious, or to lots of people, for a long time. Say it again. Indoors, low venti — ok, point being there’s a HUGE RANGE of risk across different situations.
Here’s an example of the relative risk of getting COVID inside and outside.
This is from a theoretical model of airflow indoors and outdoors. The risk of getting COVID is orders of magnitude higher indoors versus outdoors, mostly because the spaces inside buildings are a lot smaller than the space COVID particles disperse into when people are breathing outside — on the order of 10x or even 100x when it’s windy.
Is herd immunity possible?
Q: Is herd immunity even possible?
A. The basic reproductive number(R0) for COVID is 2.5, before any immunity or public health measures. To achieve herd immunity, the actual reproductive number (R) needs to be less than 1. Mathematically, 60% of people need to be immune for this to happen. In actuality, public health measures like mask-wearing and social distancing can bring R down in tandem with immunity. Over the course of the pandemic, R seems to have hovered in the 0.8–1.5 range. [8]
By Feb 1, 20% of the population in Israel was 3 weeks out from being vaccinated, adding to the 10% with antibodies from previous infection. If Israel’s R was around 1, the 20% extra immunity would have dropped it to around 0.8 (slightly more since the vaccines aren’t 100% effective, but I’m just making estimates here). I plotted their COVID dashboard from that point on alongside an exponential decay line R = 0.8, every week. [9]
Every covid patient is infecting less than 1 other person. Not enough people are getting infected for the virus to keep up. Hence the exponential decay in COVID cases in Israel.
Are vaccine blood clots a risk?
Q: Will JnJ KILL ME????
It would be a big surprise, even as surprises go. I had to make this plot 3D to even show that there was any risk of JnJ blood clots compared to the risk of death from COVID. I even discounted COVID risks by 90% since you probably won’t get COVID, so that this represents your raw risk from either proceeding as usual unvaccinated and hoping you don’t get COVID, or getting JnJ. The JnJ blood clot rates are likely inflated for people not on hormonal birth control, while the COVID death rates are inflated for people without underlying conditions. These errors are on similar orders of magnitude so I ignored both together.
I already had COVID. Am I immune?
Q. How much immunity do I get from having already had COVID?
Comparable to the vaccines, unless you’re old. In Denmark, scientists followed up on over 10 million COVID PCR tests conducted on their population in 2020. They divided up the study into two groups. The first group was people who had tested positive for COVID during the first wave, and the second was people who hadn’t. Then they examined patients during the second wave to see if group 1 tested positive less frequently than group 2. This would suggest that past infection gives a degree of immunity.
They found that people with past infection were 80.5% less likely to test positive again. Among the subset of this population age 65+, their risk of reinfection was only reduced by 47.1%. This was equally true 3 months after getting COVID as 7+ months later. The chart below compares the immunity from past infection to the immunity acquired from vaccines:
I would think these numbers are even higher if we assume that people who contract COVID on average live risker lives, and on average took more risks after getting COVID under the reasoning that they were protected by some immunity.
I would also think protection against severe infection is comparable to the vaccines (95–100%), although the study did not address this question.
The study also references the two other studies of its kind; one in Qatar that followed up on 43,000 people and found 95% protection against reinfection, and another UK study that followed up on 20,000 health care workers and found an 83% reduction in risk of reinfection.
Disclaimer: I cut a few corners in my data collection. For the immunity plots, I only had charts to use, so my guesses for the exact values of points are probably +/- 10%. I was simply looking at plots and guessing where points intersected with the x and y axes. For the timeline of an infection, the plot is illustrative. There are no formal axes as a result.
Data: My GitHub repository is available below. [12]
Sources
[1] This line probably underestimated the vaccine efficacy because it includes people who may have had COVID at the start of the experiment. It shows lesser efficacy because vaccinated people may have simply already had COVID. This probably has the biggest distortion on the first week or two, when those trial participants were registering their positive results. After that, the graph is probably correct.
[2] https://www.cdc.gov/vaccines/acip/meetings/downloads/slides-2021-02/28-03-01/02-COVID-Douoguih.pdf. From the graph
[3] https://www.nejm.org/doi/10.1056/NEJMoa2034577
[4] https://www.nejm.org/doi/full/10.1056/nejmoa2035389
[5] https://medical.mit.edu/covid-19-updates/2020/10/exposed-to-covid-19-how-soon-contagious
[6] https://www.medrxiv.org/content/10.1101/2020.09.28.20202028v1
[7] https://www.medrxiv.org/content/10.1101/2020.12.30.20249058v2.full.pdf
[8] https://www.gov.uk/guidance/the-r-value-and-growth-rate.
[9] https://ourworldindata.org/covid-vaccinations. Also note: One week was my estimate of the step size of covid growth. R tells you what the line will look like, the growth rate tells you what the intervals on the x-axis will be. I guessed 1 week accounting for a few days of incubation, and the next week or two being the most contagious days, so 7 seemed like a good average.
[11] https://www.thelancet.com/journals/lancet/article/PIIS0140-6736(21)00575-4/fulltext
[12] My GitHub: https://github.com/wmcconnell17/I-Just-Got-My-Vaccine.-Am-I-Immune-Yet-What-Data-from-Vaccine-Trials-Suggests-5-other-COVID-Qs-
