Underlying illnesses, standard of living, age, sex, social or ethnic origin… A team of researchers from the University of Oxford tried this week to profile the people most likely to succumb to infection with the new coronavirus .
They may not be the first scientists to do so, but this is the largest study to date in a country. The “anonymized” health data of more than 17.4 million British adults were thus analyzed between February 1 and April 25, 2020. Among them, 5,707 died in hospitals after contracting the disease.
Advanced age, obesity, diabetes and asthma…
The first factor associated with Covid-19 mortality is unsurprisingly age. And very, very far. In this study, published Thursday, the over 80s would have 12 times “risk” of dying than the 50-60 age group. This gap is three times larger when compared with the 40-50 age group. The prevalence of males is also confirmed: the probability of dying from the disease is doubled for a man.
Next comes the role played by co-morbidities. Pathologies such as diabetes, asthma and especially overweight prove to be aggravating factors, as already suggested by some previous studies. People suffering from morbid obesity are here 2.4 times more represented among deaths. Same thing or almost for severe cases of diabetes, those called “uncontrolled”. For severe asthmatics, the additional risk is around 1.25. On the other hand, contrary to the work of colleagues, members of the University of Oxford put the weight of hypertension into perspective (1.07).
Finally, these studies confirm that the ethnic minorities (blacks, Asians) and the poorest are overrepresented in the lists of the deceased. The risk of death is about 1.6 to 1.7 times higher. Professor Liam Smeeth, professor of clinical epidemiology at LSHTM, doctor of the NHS and co-director of the study states that these differences “are not due to underlying clinical health problems” (cardiovascular disease, diabetes, etc.). Additional work is planned on this issue. The hypothesis of the over-representation of these populations among the trades most exposed to Sras-Cov-2 (health, trade, construction sites, etc.) must in particular be explored.
A study not yet validated but already criticized
Note that this study raises some reservations, however. Especially since it has not yet been validated by the scientific community – a practice that is more and more widespread in this period of health emergency, in the United Kingdom as in France.
What is questioning here are the comparisons that are made between the risk levels of one element (age, comorbidity, etc.) and another. As one of the Seb Bacon researchers confirmed to us, this study “suggests that a man aged 40 to 49 with controlled diabetes is less likely to die than a man aged 50-59 without diabetes” . “It is a reasonable interpretation”, abounds with his colleague Ben Goldacre, also joined by Le Parisien.
But these “risk scores” lack methodological relevance, according to Carole Dufouil, epidemiologist. “You can digitally conclude a lot from the model, but basically it is wrong,” says the research director at Inserm. She points in particular to confusion between the notions of causality (one thing causes another) and correlation (a simple link between two things). According to her, it is impossible to draw such conclusions as it stands.
