# How would you properly interpret this graph? [closed]

I recently came across this article and I am having a hard time trying to properly interpret this graph. The graph is pictured below. Below that is the actual link to the article. I appreciate any help on this matter.

https://www.medrxiv.org/content/10.1101/2022.12.17.22283625v1.full.pdf

• I haven't read the article so it's not obvious to me what it is about the graph you don't understand. It would be helpful if you explained what's not apparent to you. Commented Jan 21, 2023 at 4:18
• @CareyGregory I am confused on to properly interpret it. The graph shows an increase in incidence of covid that is directly proportional to the amount of vaccines someone takes. Commented Jan 21, 2023 at 4:28
• In addition to @CareyGregory's comment, here's a reminder (or introduction) to scientific papers format: figures and tables are always referenced at least twice in any given paper - once in the text itself, and once in the text directly accompanying them (tables have titles, figures have captions). These in-paper references are usually very helpful in understanding the context and meaning of the tables/figures. Therefore, it would be helpful for all readers of the question if you: 1) add these references to the question, 2) explain what you don't understand about them. Commented Jan 21, 2023 at 19:33
• @Don_S I appreciate the advice on the format. New to stackexchange. I'll do it in the proper format next time. Commented Jan 22, 2023 at 1:51
• This question has been edited twice now to replace the graph, leaving me unsure what the graph in question actually is. OP, you need to verify that the graph in this question is the one you're asking about, and you need to heed @Don_S 's comment about including explanatory text. So I'm closing this question until you edit it to provide more details and ensure that the right graph is in the question. Commented Jan 22, 2023 at 4:10

On the Y-axis, you can see this is labeled as a cumulative incidence; cumulative means you're adding up over time. So, all the curves have to go up as you go to the right, because over time you accumulate more infections. The slope of the graph would be the rate of infections; a straight line would mean you have a constant rate of infection. If there were zero new infections past some point, then the line would level off at horizontal, but it would never go down because it's a cumulative plot: it's showing all infections that occurred on a given day or sooner.

The colors and associated labels on the right indicate that the different lines are different groups of people, grouped based on the most recent time they had a COVID infection before the study started. So, if at the start of the study someone never had COVID, they'd be in the black group. If they were last infected during the Omicron BA1/BA2 phase, they'd be in the blue group. Note that the people running the study probably don't know what actual strain any individual was infected with, but they're relying on systematic differences in strains caught at different times to show up as systematic differences in these groupings.

I assume the reason they're looking at these data is to try to make inferences about whether prior infections are protective in the future. Because the black line is above all the others, we are meant to assume that having some previous COVID infection provides some level of immunity. If not, you'd expect people not previously infected would be just as likely to be infected in the future. About 6% of people in the study who hadn't previously had COVID were infected over the 90 days of the study.

The trend is that the other lines closest to black are infections from earlier in the pandemic. That means those people have only a little more protection than the naive group, either because the currently circulating strains are different than the ones they built immunity to or because immunity has waned over time.

The people least likely to be infected are those who have been infected recently by a recent strain from the Omicron lineage, less than 2% over 90 days.

There are no groupings by vaccine in this graph, so you can't use it to say anything about vaccines, and anyone that does use it to say something about vaccines is a) trying to mislead you, b) incompetent, or c) there's some misunderstanding of what they are actually saying.

• I mistakenly put the wrong graph in the original question. I apologize. I just updated it Commented Jan 22, 2023 at 2:01
• @BishopD It is not allowed, and extremely rude, to make such a substantial change to a question. Please be respectful of the time of people who might answer your question and take better care to know what you're posting in the future. Commented Jan 22, 2023 at 2:11