“We are ready to believe numbers, not people”

What caused the questions of scientists The authors of the letter asked their Russian colleagues to provide initial data, because now the results of the study are supported only by illustrations. These graphs show the number of antibodies or lymphocytes that appeared in the body of volunteers after vaccination. Values ​​are inexplicably repeated for different measurements. For example, nine volunteers had no antibodies at all on the first day, they appeared on the 14th day, and then for the results for the 21st and 28th days, the graph shows identical results. Such repetitions were found, among other things, between groups: in different people who were injected with different components of the vaccine, the distribution of antibody levels turned out to be the same.

Tell us what kind of research you do.

- I am engaged in the analysis of biomedical data, in particular, checking their reliability. You can find in Nature and other journals my publications with investigations based on this analysis - they made a lot of noise in the scientific community.

On Friday, I read with interest an article in The Lancet on the Sputnik V vaccine. I was surprised by some apparent inconsistencies. I contacted several colleagues to analyze the data together. As a result, we decided to write an open letter and point out the inconsistencies found. Just this. We can't check anything since we don't have "raw" input data. While we are just concerned that there may be errors, there may be repetitions in the results. But specifically, without the original data and FCS files, we cannot assess their quality. So far, we only have statistical data - the mean and the standard deviation. But there are no detailed data on the control group of volunteers, on those who were vaccinated. Not enough information to judge: this is a very, very unlikely coincidence or publication error.

The lead author of the article [Denis Logunov] has already said that he will not answer us directly. He is ready to answer only to the editors of The Lancet if they question the results of his work. Actually, it's weird. Such an important publication in such a reputable journal. You expect that all the data will be immediately provided.

You mentioned that one of the authors of the article, Denis Logunov, has already answered you in absentia. He states that "the data on credits is exactly [as they are presented in the illustrations in the article]."

- This, of course, is not enough. We are ready to trust numbers, not people.

Let's assume that the illustrations reflect the actual results of the study. Can this be explained somehow? For example, the fact that very few volunteers took part in the study.

- Hardly. Take, for example, the third illustration [which shows a population of T-cells that fight the virus], which comes in our second letter. We need to explain how you can first get some value in a certain group of patients for one attribute, and then exactly the same value, but for another attribute. First, the result of the change in the number of CD4+ T-lymphocytes for nine patients is given there. And then exactly the same result for other cells — CD8+ T-lymphocytes. How is this possible? It is precisely such coincidences that must be justified, because this is very, very strange and unlikely.

This may be due to the limitations of measuring instruments? That is, the results fit only within certain values.

— But then it would be necessary to admit that these are not specific values, but qualitative studies. There are also drawn two graphs with a logarithmic scale with values. On the first for nine patients, as stated in the description, there are values ​​of the population of CD4+ cells. And on the second - a population of CD8 + cells. This is not a quantitative study, but the values ​​obtained by the method described in the work.

Either the studies weren't done as described - but then we have to know how - or we have to conclude that a very, very, very unlikely situation has occurred. And then you have to explain why it happened.

These values ​​- for CD4+ and CD8+ lymphocytes - are in no way related to each other. They cannot exactly match in nine people.

When it comes to research as important as testing a vaccine during a pandemic, the only way out is to immediately sound the alarm. Simply because magazines are in a hurry to publish

And this can not be explained in any way by insufficiently detailed illustrations?

— That's why we ask for the initial data in the FCS format. This situation just shows why you can not publish only illustrations. If there were data, there would be no questions.

You are in the business of exposing falsification in research. What should be the procedure after such claims have been made?

- There is a different approach. But when it comes to research as important as testing a vaccine during a pandemic, the only way out is to immediately sound the alarm. Simply because magazines are in a hurry to publish. You may have noticed that in the study, the last blood sample was taken on August 11th. We add a week for writing an article and about ten more days for reviewing before publication. This is very fast, even too fast for such a topic.

But since they publish quickly, it is necessary to respond quickly. Faster than under normal conditions.

This is exactly what we did with the hydroxychloroquine article in The Lancet . A letter was written to the editor. We are doing the same now. You may notice that our open letter was signed by [Fellow of the Center for the Study of Tropical Medicine at the University of Oxford in Thailand] Dr. James Watson, who initiated the procedure for retracting the hydroxychloroquine article. Sometimes you need to act very quickly and as openly as possible. And involve as many experts as possible to see if the results are consistent.

already Has The Lancet contacted you?

- Not yet.

Now the third phase of vaccine research is starting in Russia, thousands of volunteers are being invited. Should this phase be postponed due to issues that have arisen, or can it be continued?

— To continue or not, it is necessary to decide not on the basis of our questions. It is necessary to pay attention to the doubts that Russian scientists have already expressed after the announcement of the third phase. They said that after the first phase, there is too little data on the safety of the vaccine. Even if you do not think about data problems, there are not enough volunteers. Six combinations were tested. Four groups of nine people and two groups of 20 people were selected. The largest sample on which security was checked was 20 people. This is very small for the first phase.

However, this particular vaccine does not really bother me, because the principle of its action is not that innovative.

You are following research on COVID-19. Were there many publications from Russia at all?

— I saw interesting works, but so far the contribution is not very significant. But this can be explained even by looking at the official incidence data. After all, the virus came to Russia later than to the US and China. Logically, they are a little behind. This does not bother me, because Russia has a very good and strong scientific community. I'm worried about speedy peer review and poor data analysis prior to publication on such an important topic. Of course, not only the authors of the work are responsible for this.