Imagine this scenario: At a routine visit, your doctor administers a new genetic test that shows you have a hugely elevated risk of a heart attack in the future. You’re in shape; you feel fine. But the prediction is in your DNA.
The next day, you tell your employer that your doctor wants you to make some adjustments — switch to a less physically taxing role, or maybe lower your stress levels in an effort to save your life. Can your boss legally deny you these accommodations? Under current law, yes.
That’s because the federal rules designed to protect against employment and insurance discrimination based on genetics were not written with this technology in mind. While the Genetic Information Nondiscrimination Act of 2008, or GINA, protects workers from being fired over their genetic test results, and the Americans With Disabilities Act protects those with active disabilities, neither law compels an employer to provide accommodations to help mitigate a person’s future health risk.
This could become a problem, as a new type of personalized medicine called polygenic risk scoring becomes increasingly popular. The technology uses DNA to estimate a person’s likelihood of developing a specific disease. Legal and public health scholars worry that as genomic tools advance quickly, our legal frameworks are falling behind — in some cases leaving those who use genetic tests vulnerable to outright firings.
“These polygenic risk scores are going to become an increasing part of your life,” said I. Glenn Cohen, a Harvard Law School professor who specializes in the intersection of bioethics and the law and has co-written a report on the issue. “To me, it’s a great opportunity to reopen questions about what the rules should be — questions that, from a policy perspective, are legitimately hard.”
To understand the legal lag, it helps to first consider how the science of risk prediction has evolved. Many people thinking of genetic testing envision monogenic tests, which search for specific gene mutations that cause specific disorders, like Huntington’s disease or cystic fibrosis. But polygenic risk scores scan for an array of variants throughout a person’s genome, synthesizing the information into a cumulative estimate of their risk of developing a disease like diabetes or cancer. If a monogenic test is like finding one severed wire that cuts off sound to a speaker, a polygenic risk score is like a soundboard, with dozens of switches that can combine to produce a dangerous frequency.
When GINA, the main statute that prevents genetic discrimination, became law almost two decades ago, companies were barely able to map individual genomes at an affordable price. GINA was written for the severed wire, not the soundboard.
Sam Trejo, a quantitative sociologist at Princeton University who studies polygenic risk scores, warned that GINA was “definitely not sufficient” to protect people from the types of situations they could encounter in the coming decades. “We need more robust protections there — treating DNA like something akin to a protected class,” he said.
Even among medical experts, polygenic risk scores still come with a lot of unknowns. A high risk score for heart disease, for example, is not detecting a clear structural defect, but instead, a complex mix of genetic influences that could involve your taste buds and cravings, your propensity to exercise or even a personality trait that drew you toward a high-stress job, Dr. Trejo said. Since the algorithms behind the scores are simply drawing correlations, the root causes and implications of a score remain opaque.
If the medical field doesn’t fully understand how to interpret polygenic scores, it makes sense that workplaces don’t either. Still, doctors agree that lifestyle changes can help offset certain genetic predispositions — which is why they might encourage patients with certain polygenic scores to minimize their risks on the job.
That’s where the legal concerns lie. Researchers at Harvard Law School and Yale Law School who recently examined the regulatory gap uncovered a key loophole pertaining to workplace accommodations. A construction company, for example, cannot fire a healthy worker because of a polygenic test result itself, fearing future health costs; but the company can turn down the worker’s request for a less strenuous role. That’s because GINA does not require employers to provide accommodations, and while the A.D.A. does, that law does not consider a genetic predisposition alone to be a disability that invokes those rules.
In a recent case, Darby v. Childvine Inc., a plaintiff with a BRCA1 mutation linked to breast cancer was fired after she requested time off to undergo a double mastectomy. The U.S. Court of Appeals for the Sixth Circuit held that “a genetic mutation that merely predisposes an individual to other conditions, such as cancer, is not itself a disability under the A.D.A.”
There also exists a legal dead zone for workers who begin experiencing mild symptoms of a disease, since GINA protects only those in whom a disease has not yet manifested, while the A.D.A. requires accommodations only for those whose disabilities impair major life activities. Thus, if supervisors at a chemical manufacturing facility know of a healthy worker with a high polygenic risk score for a rare lung disease, the employee could be fired immediately upon developing mild shortness of breath, the scholars said. GINA protections could vanish, while the standard A.D.A. protections would not yet apply, leaving the worker to rely on difficult-to-prove legal exceptions to save the job.
In addition to its workplace rules, GINA also prohibits health insurers from adjusting premiums or denying coverage based on genetic information. But the law doesn’t extend to life, disability or long-term-care insurance, which is why many people with genetic variants linked to a high risk of dementia cannot obtain long-term-care insurance.
Polygenic risk scores are still far from their full potential. The scores do not offer medical certainty; they only estimate a person’s risk of developing a given condition relative to the population used to train the algorithm. Since they rely on the relatively small effects of numerous genetic variants, the tests for some conditions will need data from far more people in order to decrease statistical noise and sharpen their accuracy. As of now, some results could be misleading, particularly for people of non-European descent, who are underrepresented in many models.
Still, the tests’ popularity is growing: The direct-to-consumer market has burgeoned, and at least one company has filed with the Food and Drug Administration for approval of its test to measure inherited genetic risk of coronary heart disease. Population geneticists and some physicians are arguing for more widespread use in clinics, saying the strongest algorithms could inform key patient care decisions, such as when to begin prescribing statins.
Some also say polygenic risk scores could factor into so-called composite scores, integrating genetic data with other traditional metrics like diet, smoking status and body mass index to increase their accuracy. But legal scholars see an ambiguity here, too, as factors like age and sex are explicitly excluded from GINA. If genetic data only makes up a portion of your score, does GINA still apply?
“I don’t want to leave anyone with the idea that this is a strong reason not to do it,” Cohen said. “My message is really for lawmakers: If you think these polygenic risk scores are good for public health, maybe you can plug these gaps.”
