Search for a report, a publication, an expert...
Institut Montaigne features a platform of Expressions dedicated to debate and current affairs. The platform provides a space for decryption and dialogue to encourage discussion and the emergence of new voices.
26/11/2020

Covid-19’s Racial Divide: Why Access to Ethnic Data Matters

Covid-19’s Racial Divide: Why Access to Ethnic Data Matters
 Anne Bouverot
Author
Chairperson of the Board at Technicolor and the Chairperson of Fondation Abeona
 Tanya Perelmuter
Author
Director of Strategy & Partnerships, Fondation Abeona

On April 10 2020, Ontario health officials said that collecting race-based coronavirus data was not needed since "regardless of race, ethnic or other backgrounds they’re all important to us". Although alarming statistics about the high level of contamination and deaths amongst African Americans were already coming from the US, they assumed that the situation would be different in Canada, which has a universal health-care system. Two weeks later, the city of Toronto, defying the province officials, launched racio-ethnic data collection, saying that, to save lives and respond effectively during the pandemic, it was important to understand which groups were more at risk. 

Preliminary findings released in July show that 83% of people who had reported infection identified with a racialized group (Black, South Asian, Southeast Asian, and Latin American), 51% were living in low-income households and 27% in households of 5 or more people. In the US, where the collection of ethnic statistics is common practice, the numbers are particularly cruel: African-Americans are twice as likely to be contaminated, and three times as likely to die from Covid-19. In Chicago, Blacks account for 67% of deaths while they make up 32% of the population. 

Could the situation be different on the other side of the Atlantic? It is difficult to know, since France has traditionally embraced an idea of universalism, masking diversity in order to treat everyone equally: a color-blind approach that finds its roots in the French Revolution, the Dreyfus Affair, and later, the dark days of the Vichy government. The French, as a lot of other European citizens, are very sensitive to privacy issues regulated by the EU General Data Protection Regulation, which governs the collection and use of private data. With only some rare exceptions, ethnicity statistics are prohibited and data on the official 25 criteria of discrimination are extremely difficult to collect and use. 

In May 2020 however, CNIL, the French national data protection agency, authorized the first cross-disciplinary study, EpiCOV, to be conducted by Inserm, France’s national research institute on human health. The goal was to constitute a statistically representative cohort of 135,000 French residents and map their SARS-COV-2 immunity status against common social indicators.

In the US, where the collection of ethnic statistics is common practice, the numbers are particularly cruel: African-Americans are twice as likely to be contaminated, and three times as likely to die from Covid-19.

The first part of the study, published on October 29, confirmed high exposure of the most precarious to the Covid-19 epidemic over 3 criteria: crowded housing, manual work, especially in the area of care done by women, and neighbourhood density. 8.2% of residents of so-called "priority" neighbourhoods tested positive to the CoV-2-SARS antibody test, compared to 4.2% elsewhere in France, almost double.

The most revealing results demonstrated a 9.4% exposure among immigrants of non-European origin, 6.2% among their children, compared to 4.8% among European immigrants and 4.1% among those whose parents were born in France.

Since racial and ethnic data collection is not allowed in France, researchers relied on its best proxy: place of birth. As Joe Cressy, the Toronto chair of the Board of Health who spearheaded the city’s data collection, pointed out, "we have long known that it’s your postal code, rather than your genetic code, that is the biggest driver of health. The social determinants of health - race, income, housing status - have long determined who gets sick, who lives, and who dies". Unsurprisingly perhaps, these results are in line with surveys carried out in the US, in the UK and in Canada, even if ethnicities may be different.

Individual privacy is of course of paramount importance, but so is a right to non-discrimination. These two basic human rights should not be opposed to each other, but rather work hand in hand – with the relevant data properly anonymized and protected.

Although we are far from a consensus on this in France, the subject is slowly starting to be voiced. Sibeth Ndaye, a former government spokesperson, wrote in Le Monde in June that, for ethnic data collection, "it is urgent to re-examine the question of the representation of people of color in the public, political, economic and cultural life of our country, and to reopen the debate around ethnic statistics in a peaceful and constructive way".

Although we are far from a consensus on [ethnic data] in France, the subject is slowly starting to be voiced.

In early March, just before the first lock-down, Institut Montaigne also published Algorithms, please mind the bias!. With health as one of the key areas of application, this report recommends an "active fairness" approach - testing algorithms for potential discriminatory behavior before they are deployed for general use.

We believe that diversity-blind approaches do not prevent discrimination, far from it, and that decisions informed by data, collected in a privacy-respectful way, are better decisions. We hope that the debate about data, fairness and equity will continue.

 

 

Copyright: INDRANIL MUKHERJEE / AFP

Receive Institut Montaigne’s monthly newsletter in English
Subscribe