DANVILLE, PENNSYLVANIA—Dana Atkinson begins to cry as she recalls the January phone call that cast a pall over her future and that of her children. "I'm just still processing. It's a lot. Especially when your kids are involved," says the 38-year-old nurse, who works at the Geisinger Medical Center here near the banks of the Susquehanna River. (She asked that Science not use her real name.)
Months earlier, during a routine medical checkup, a lab technician had drawn a vial of Atkinson's blood. With her consent, some of the sample went off to Regeneron Pharmaceuticals, a biotech company in Tarrytown, New York, where the stretches of her DNA that encode proteins were sequenced. Atkinson forgot about her donation until the phone rang at home this winter. The caller, a research coordinator at Geisinger, told the mother of three that she carried a rare mutation linked to Long QT syndrome—a heart rhythm disorder that can cause the organ's electrical activity to suddenly go haywire, triggering fainting, seizures, and even death.
The caller provided few details, but advised Atkinson to come in soon for a consultation. "Though I have a medical background, I barely understood what I was positive for," she recalls. But the nurse understood enough to worry about her own health and that of her children, who might have inherited the mutation. And she also wondered about her mother's death 38 years before. She had died in her sleep, just 6 weeks after giving birth to her only child. An autopsy revealed nothing—but Long QT is a syndrome that leaves no obvious clues.
During Atkinson's follow-up at a medical center in Forty Fort, Pennsylvania, a genetic counselor told her that although she carried the disease-linked gene variant, she didn't necessarily have the syndrome. And if she did, her doctor could help her reduce her risk of developing problems. The counselor then had Atkinson diagram her family tree, to pinpoint other relatives who might also carry the dangerous variant. Atkinson's children have had their DNA screened, but their mother has not yet shared her news or their results with her extended family. She's still not ready.
Ready or not, more than 400 other Pennsylvanians have received similar calls, letters, or electronic messages over the past 2 years as part of the world's largest clinical genome sequencing effort, the first available to average Americans as part of standard primary medical care. Known broadly as the MyCode Community Health Initiative and run by the Danville-based Geisinger Health System, the effort has so far sequenced the protein-coding DNA, or exomes, of more than 92,400 people. More than 166,000 have enrolled in the study, and the goal is to ultimately enlist half a million of the nonprofit's 3.3 million patients. Geisinger asks each participant whether they are willing to be contacted if the sequencing reveals a DNA variant that puts them at risk for disease—and roughly 85% have agreed to be part of the experimental program called GenomeFIRST Return of Results.
In addition to using the sequencing results to prevent and treat diseases, Geisinger hopes to answer myriad questions, from how primary care physicians with little genetics knowledge cope with advising patients informed of disease mutations, to the challenges of "cascade testing"—the follow-up with relatives who could also be at risk. The project will inform ongoing debates, including how much people should be told about what their genomes reveal. "This is the future of health care," predicts Michael Snyder, director of Stanford Medicine's Center for Genomics and Personalized Medicine in Palo Alto, California. "Incorporating a person's genome sequence information into disease risk assessment is a no-brainer."
Beyond the ethical, medical, and scientific issues being explored by Geisinger is the crucial question of whether widespread genomic screening as a preventive medical measure is cost-effective. Regeneron is paying for each patient's initial DNA sequencing, in return for access to those data and Geisinger's health records, but that isn't likely to be practical nationwide.
The cost of widespread genomic testing in the United States would fall mainly to insurance companies—and they will want to be sure that testing does not open the door to unsustainable health care costs, says David Veenstra of the University of Washington in Seattle, who studies the implications of using genomic information in health care. Veenstra, who is working with Geisinger researchers to assess GenomeFIRST's cost effectiveness, adds: "In about 5 years I think we are going to have a very serious discussion about whether this is something that should be implemented in almost a public health sense."
In the early 1800s, iron and coal mines dotted the rolling green hills of northeastern Pennsylvania, where this ambitious medical experiment is unfolding. By the mid–19th century, iron mills dominated Danville's economy. The widow of iron magnate George Geisinger, Abigail, used her fortune to open Geisinger Hospital in 1915, during a typhoid outbreak. Over the decades, Geisinger has become the region's main health care provider with 13 hospitals, including its flagship here and two research centers. Today, its physicians battle health issues very different from typhoid. Type II diabetes, obesity, and opioid addiction are at crisis levels.
In 2001, Geisinger lured Glenn Steele, then the dean of biological sciences at the University of Chicago in Illinois, and now chairman of xG Health Solutions in Columbia, Maryland, to its helm. A member of the National Academy of Medicine, Steele had a stellar record as a surgeon and researcher focusing on liver and colorectal cancer. His decision to go to the rural health system stunned colleagues, in part because he had spent years building one of the country's top human genetics programs at the university. He tried to coax some of the faculty to follow him to Pennsylvania, but they declined. What Steele saw—that his colleagues didn't—was that Geisinger was a potential gold mine for genetic data. It began using electronic health records very early, in 1990, and because this region maintains a relatively stable population, it has gathered data on the health of grandparents, parents, and children. This type of generational information is extremely valuable for tracking the roots of genetic disease—and Steele thought it could potentially allow Geisinger to reengineer medical care. "[In this area] you could look at multiple generations of outcomes and that's something that, outside of Iceland and a few other places in the world, just wasn't available," he says.
Once there, Steele quickly launched a "center for population health" to study disease incidence among the people in the region. In 2006, hoping to speed the hunt for disease genes, three Geisinger researchers—epidemiologist Walter "Buzz" Stewart, molecular biologist David Carey, and pathologist Glenn Gerhard—successfully lobbied for a biobank with blood and tissue samples volunteered by Geisinger's 3-millionpatient network. In 2007, the MyCode initiative enrolled its first volunteers. At the time Geisinger wasn't telling its patients whether their DNA revealed a disease mutation. Patients only gave Geisinger permission to analyze their samples and DNA for research. But that changed in 2013, with the launch of GenomeFIRST. Participants were offered a consent form explaining that Geisinger would tell them, and their physicians, of genetic findings relevant to their health.
As the number of enrollees grew rapidly, Geisinger partnered in 2014 with Regeneron to use its growing DNA-sequencing capability for free. The company mines both DNA data and a version of the electronic health records stripped of patient identity for links between genetic variations and diseases, clues that could identify targets for new pharmaceuticals. If the effort leads to successful drug development, Geisinger will receive royalties that will be reinvested in the nonprofit's health care system.
For now, GenomeFIRST tells patients about mutations in 76 genes that have been linked to 27 conditions, ranging from breast cancer to heart disease. All the conditions can be treated through surgery, pharmaceuticals, or lifestyle changes, or prevented altogether, says Michael Murray, Geisinger's head of clinical genomics. Disease gene variants with no clear-cut medical treatments, such as APOE4, which raises the risk of Alzheimer's disease, are not disclosed. Early indications are that about 3.5% of study participants will discover that their genomes harbor a disease-linked DNA variant, so Murray expects the impact of the population-wide scanning to be "profound."
Once a second DNA sequencing confirms a patient has such a variant, their primary care physician receives a message through a secure medical portal, and a copy of the genetic test report is embedded in the patient's record. The note also provides the physician with a link to an educational slideshow that describes the potential consequences of the risk variant and guides the doctor through talking to the patient. Stephanie Cabello, Atkinson's internist, initially thought counseling a patient about a mutation would prove daunting. "But [Geisinger] made it really quite easy," she says.
About a week after the physician is notified, the patient receives an electronic note through the Geisinger health portal—or a letter in the mail, or a call—that says, "We found important information in your DNA sample. This information may guide your healthcare and your family's healthcare now and in the future."
The note doesn't say what the mutation is. Rather, it asks the patient to call Geisinger. On that call, a research coordinator begins a scripted conversation explaining the genetic finding and the implications for disease risk. The patient is also encouraged to complete a family history and to discuss the news with relatives by sharing a letter Geisinger has prepared. It explains the result and advises how they can pursue testing.
The coordinator then tells the patient they can meet with their physician or a genetic counselor to further discuss the result. Not all the follow-ups go as expected. In 2016, one of Cabello's male patients learned he had a BRCA2 mutation. Although best known for causing breast and ovarian cancers in women, the gene can also raise the risk of many cancers in men. Cabello was excited by the chance to offer aggressive cancer screening that could save the man's life, but his reaction startled her. "He actually wasn't interested in knowing more about it, which was terrible," she says. He said he would let her know when he was ready to hear more. Fourteen months later, he still hasn't contacted Cabello.
When you start doing tests in people who you don't understand adequately, you risk putting those people at great harm [from invasive follow-up tests].
For some families, news of a potentially harmful genetic variation cascades beyond the Geisinger system, across state and national boundaries. Coal Township resident Patrice Molesevich, a nurse–turned–health plan case manager for Geisinger, joined GenomeFIRST in March 2015. A year later, a study coordinator called her about a mutation in the BRCA2 gene.
"I knew there was going to be cancer," Molesevich says, because her brother died at 37 from pancreatic cancer, her father died from prostate cancer, and her family tree was speckled with other cancers, including colon, lung, and throat. But the actual mutation surprised her because, as far as she knew, there was no history of breast or ovarian tumors, which are more often linked to BRCA2.
She met with a Geisinger genetic counselor, who estimated, based on the pattern of cancers in her family tree, that she had an 85% chance of developing breast cancer and a 7% chance of pancreatic cancer. An MRI then revealed early-stage cancer in Molesevich's left breast. Molesevich chose to have a double mastectomy and reconstruction. She was also prescribed a breast cancer treatment, anastrozole, for the next 10 years to further curb her risk.
The genetic counselor advised that her two children and her brother's should have their BRCA2 checked. Molesevich's daughter didn't have any obvious mutations; her son is resisting the test.
Molesevich then contacted her sister-in-law, Kerry Georgeadis, who had been left to raise three young girls, one a baby, when her husband, Molesevich's brother, died in 1992. Georgeadis was stalwart. "I never dreamt that we would still be dealing with cancer-related issues all these years later," she says, "but we are, and we are going to do everything that we can to proactively and positively fight this dreaded disease."
First, Georgeadis told 30-year-old Lynsey Towne, her middle daughter, who was living in New York City. Towne's OB-GYN helped her get tested, and good news followed: She had not inherited the mutation.
Ashley Meskill, 33, Georgeadis's eldest child, had a more harrowing experience. Her doctor on Long Island in New York wasn't familiar with genetic testing, and sent her to a small, local cancer center. When she learned she had the mutation, a counselor there simply rattled off the odds of developing various cancers and made her feel as if she had gotten "a death sentence," Meskill says. "That sent me into a ball of anxiety." Now she's seeing physicians at Memorial Sloan Kettering Cancer Center in New York City and she says she is slowly coming to terms with the news and what it means for her and her family.
Georgeadis's youngest daughter, 25-year-old Cydney Engle, chose to go to Sloan Kettering for her BRCA2 test. In March, she got the bad news. Now that the initial shock has worn off, she is considering a double mastectomy. For Engle and Meskill, the test results have sparked a healthier lifestyle—and activism. They are now training for marathons and started an Instagram account, @brca_sisterhood, to connect with others who carry this variation.
Some believe Geisinger actually returns too little genomic information to patients. It started with a list of 56 genes that the American College of Medical Genetics (ACMG) in 2013 concluded were sufficiently linked to diseases that had medical interventions for people to be told about a mutation. After consulting clinical geneticists at Geisinger and other institutions around the country, Murray and his colleagues added three genes for two conditions—ornithine transcarbamylase deficiency and hereditary hemorrhagic telangiectasia—and 17 additional genes that are known to cause cardiovascular diseases, bringing the total to 76. "I've seen the list and nothing jumped out as radical or aggressive," says David Miller, a medical geneticist at Boston Children's Hospital and co-chairperson of an ACMG working group that tracks this issue.
But to Stanford's Snyder, Geisinger is "not aggressive enough." In a recent exome sequencing study of 70 individuals, he found 12 with variations in genes that he considers "medically actionable." Half of these genes aren't included on ACMG's current list of 59 disease genes, nor are they part of Geisinger's return of results. Geisinger also does not inform patients about recessive mutations, which are relatively harmless to the carrier but could cause diseases such as cystic fibrosis in offspring if they inherit a second copy of the mutation from their other parent. In Snyder's view, people have the right to know about such genetic findings, as well as about mutations that confer elevated risk of Alzheimer's disease.
"Treatability may not be the only consideration people have regarding such information," says Lisa Parker, who directs the Center for Bioethics & Health Law at the University of Pittsburgh in Pennsylvania.
But James Evans, a physician and director of Adult and Cancer Clinical Genetics Services at the University of North Carolina in Chapel Hill, warns that physicians and the public "shouldn't be seduced by the occasional cherry-picked case reports" into thinking that broadly screening the DNA of healthy populations for clinical use is always a good idea. He calls that a "dicey proposition" and "dangerous."
Only a few population-based screening procedures, he notes, have proved beneficial over time: colorectal center screening, pap smears, and mammograms (although that has been contested for younger women). "When you start doing tests in people who you don't understand adequately, you risk putting those people at great harm" from invasive follow-up tests, Evans cautions.
Even if patients aren't harmed, the cost may be unsustainable, says Leslie Biesecker, who leads the Medical Genomics and Metabolic Genetics Branch at the National Human Genome Research Institute in Bethesda, Maryland. For example, people who learn they have mutations that cause Lynch syndrome are advised to get screened more frequently for colon cancer, which could catch a tumor in an earlier, and more treatable, form. But, Biesecker asks, how many patients had to be sequenced, at what cost, to discover the one with a possibly harmful gene?
Current estimates place the actual cost of genome sequencing in the low thousands of dollars per person, but that's just the beginning of the tab. How much does the additional colon cancer screening cost, and how accurate is it? False positives, Biesecker says, can lead to unnecessary surgery or follow-up tests, and additional costs.
Health economists commonly measure changes in population health using "quality-adjusted life years" (QALYs), shorthand for a year of perfect health. The current consensus in the United States is that interventions that add a QALY to a person's life span for less than $100,000 are cost-effective. So far, says health economist James Buchanan of the University of Oxford in the United Kingdom, no one knows whether using sequencing to screen for disease variants in a low-risk population beats that threshold.
Jason Vassy, a primary care physician at Harvard Medical School in Boston, and colleagues recently provided some rare data on the economic question, following 50 healthy adults receiving normal medical care and 50 who also had 4600 genes analyzed, at a cost of $5000. Physicians ordered twice as many follow-up tests for the people with DNA information, leading the group to conclude, in the 1 August issue of the Annals of Internal Medicine, that such an aggressive sequencing approach is not yet cost-effective.
However, Vassy notes that Geisinger's approach may make more financial sense as it's more targeted, limited to conditions for which treatments exist. "It's a top-down approach to identifying high-risk individuals in a large population for whom they can provide a medical intervention," he says.
Hard-to-predict human behavior also muddles cost-effectiveness calculations. Medical statistician Glenn Palomaki of Brown University's Warren Alpert Medical School, who studies the consequences of prenatal screening and genetic testing, notes that there is no health benefit to learning about a mutation if a carrier doesn't do anything—and the cost of DNA sequencing is wasted. In contrast, if a person overreacts out of fear, they might consume too many health care services. And will those who find out that they don't carry a mutation in the BRCA genes assume they are protected from breast cancer and neglect regular mammograms, endangering their lives and racking up more medical costs in the long run? Geisinger's researchers will try to analyze all those questions as Genome-FIRST amasses data.
Many have argued that adults who consider themselves healthy are afraid of genetic testing and will be depressed and anxious if they receive news they carry a disease-linked DNA variant. Biesecker suggests that Geisinger's recruitment success in a "middle-America swath of the population" contradicts the first concern. "People are eager and interested in this," he says, "If there is a way you can improve their health care with genomics and sequencing, they're up for it."
As for how they handle any bad news, the numbers present a more mixed message. About half the people told of a disease mutation meet with a genetic counselor; another quarter go to their primary care provider. One quarter don't do anything right away, says genetic counselor Adam Buchanan, who co-leads GenomeFIRST.
Patients support the program's research goals, says Mark Speake, a doctor who's been working with Geisinger since 1990. "[They] just seem pleased to participate. … That's their contribution to the next generation." He thinks this is particularly important for many who suffer from obesity, diabetes, hypertension, lipid disorders, and thyroid disease. "These things run in families and they are very common in our population." Three of Speake's patients have received news of mutations, results that he says didn't come as a total surprise, and didn't overly disturb them. "A light bulb goes off and the family history makes sense to them now."
Health care systems in Israel and Singapore, as well as others in the United States such as Kaiser Permanente, are watching Geisinger to inform their own genome-based medicine. But Speake isn't optimistic about implementing a similar program in other health care systems or as part of a government program. "I think it is a good start and good that it is voluntary. I think a lot of folks are worried about 1984, George Orwell and all that," he laughs. "[But] our patients trust us … they don't think we are prying into their background. We are just trying to take good care of them."
One GenomeFIRST element that is still a work in progress involves children. Between 2% and 3% of the participants enrolled are children; one couple even signed up their baby immediately after birth. Murray and his colleagues are working on a counseling system that provides extra support to families when parents are told a child has a mutation.
Atkinson's experience underscores the need to support such parents. In May, a blood test showed that her daughter has not inherited the Long QT mutation, but both her sports-playing sons, who are 9 and 12, have. For boys with this syndrome, sudden death is most likely to strike during adolescence—and during intense exercise. (Girls are at high risk in their early 20s.)
Atkinson's older son has now seen a pediatric cardiologist, had an electrocardiogram, and worn a heart monitor for 2 weeks. The results suggest that he has Long QT syndrome. A cardiologist has recommended he begin taking β-blockers to reduce the risk of his heart going haywire. Atkinson also purchased a $2500 automatic external defibrillator to jump-start his heart if it stops. The cost of the device was not covered by her insurance—even with a Long QT diagnosis.
According to the medical tests so far, Atkinson herself is a borderline case. But given her family history, she is taking β-blockers. Her younger son will be monitored regularly. Though it may take many years for others to resolve whether Geisinger's genomic approach is the future of medicine, she's a believer. "I thank God for this program, that this [mutation] was found and I'm not burying one of my kids."