How Inequality Threatens the Promise of Big Data
This year, the health team’s research has focused on big data and the world of innovation that it will open up. But while emerging technology will give us much more comprehensive data than we’ve had in the past, it won’t be perfect. Inequality will probably create significant data blind-spots/gaps in the future, because it does in the present and it has in the past.
For instance, racial minorities have long been under-represented in clinical drug trials domestically. Clinical Ambassador, an organization dedicated to reducing these disparities describes the problem:
“Lack of minority inclusion in clinical research limits the accuracy of scientific results and leads to drug approvals that should be denied, the licensing and use of dangerous medical products, and also means that prescriptions thought generally safe for all may be recalled due to bad reactions in minority patients simply because they weren't included in the study sample... The potential of harm to underrepresented populations may not be revealed until after a drug is on the market. Equally unfortunate, new formulas thought to be useless in the general population may be of real benefit to minorities- if we know. More clinical research is required to better understand these differences.”
But the reverse is often true as well: in many studies minorities and the poor are over-represented. I happened upon an article on clinical research organizations testing drugs “on poor people in faraway places. Russia, China, Brazil, Poland, Uganda and Romania are all hot spots for… clinical research or clinical trials,”
The article, rightly, focuses on the injustice of testing on vulnerable people:
“The growth of the clinical-trial industry in India needs to be seen within the social and economic context of the country. According to the United Nations, 40 percent of people in India are illiterate. The gap between rich and poor continues to grow. Many risk not knowing whether the treatment their doctor is prescribing is [a proven] treatment or a part of a clinical trial.”
But conducting studies disproportionately on impoverished people poses problems for the privileged as well.
In the recent past, we believed that age and gender are the variables that need to be accounted for to make research generalizable. We believed that being impoverished (or being a college student or unemployed) are all social, not physiological, conditions and therefore not a relevant variable. But separating the social from the physical is problematic, and we’ve had evidence of that for a good century now.
One of my favorite cautionary medical tales is recounted in Robert Sapolsky’s excellent essay, “Poverty’s Remains.”
The essay explains that since at least the 1500s, we’ve relied almost exclusively on the bodies of impoverished people for research cadavers, (for numerous fascinating/gruesome reasons, many of which are detailed in the essay). It turns out that living with the constant possibility of being attacked physically or losing your and your family’s access to food and shelter is sort of stressful, and that’s pretty much the condition of most people who live in poverty. Well it turns out, that this constant stress creates some physiological differences in human bodies, for instance, bloated adrenal glands and shriveled thymus glands. And since most cadavers were those of poor people, most cadavers had bloated adrenal glands and shrunken immune glands. However, when looking at these organs, scientists didn’t realize they were looking at evidence of a life of stress and poverty—they thought they were just looking at "normal" internal organs.
So around 1900, an Austrian pathologist by the name of Richard Paltauf autopsied a child who died of Sudden Infant Death Syndrome (SIDS) and got a glimpse of its normal, healthy thymus gland, and perceived it to be hideously enlarged–so enlarged, he thought, that it could have been responsible for the child’s death. After autopsying several others who died of SIDS, he became certain these “enlarged” glands were the culprit in SIDS, and so from the ‘20s to the ‘50s, irradiating the throats of infants to shrink these glands became the preferred SIDS preventive treatment. Obviously, this didn’t actually do a good job of preventing SIDS, but it did do a great job of giving people cancer.
It’s an important lesson. In the future, we’ll have massive amounts of new data from biosensors, self-trackers, data-donors, as well data from our interactions with retail and medical institutions. But this data will not come equally from all sectors of society. In fact, the factors that make someone more or less likely to participate in a study or generate or share data—for instance, access to the technology that will generate data, distrust of companies or government—may actually correlate to physiological difference. And unless we keep that in mind as the world of big data emerges, we run the risk of reproducing Paltauf’s mistake in the 21st century.
For more on the future of poverty, more specifically ambitious visions and actions for ending poverty, visit our latest foresight game with the Rockefeller Foundation, Catalysts for Change.
- Ben Hamamoto's blog
- Login to post comments
-