Genomic Messages: How the Evolving Science of Genetics Affects Our Health, Families, and Future - George Annas, Sherman Elias (2015)
Chapter 2. Personalized (Genomic) Medicine
Enthusiasm for gene-centered medicine is contagious,
and I am not immune to it. In my view, however, the
fundamental issues remain. . . . Enormous amounts of new
knowledge are barreling down the information highway,
but they are not arriving at the doorsteps of our patients.
—Claude Lenfant (2003)
We Americans think of our health care system as the “best in the world,” and that probably explains why changes in health care provoke widespread anxiety. We also see scientific progress and technological advance as good things and hope that they will ultimately make our health care, and our lives, better. Evolving genomic medicine, also known as “personalized medicine” and as “precision medicine,” holds great promise for improvements in medical care, but it also will bring sometimes unwelcome changes in the way we interact with our physicians and other health care personnel. We know change is coming, but we are anxious about what it will mean for us personally and for the American health care system.
Former Obamacare adviser Ezekiel Emanuel, an expert on the Affordable Care Act, begins his book on the future of American medicine, Reinventing American Healthcare, with the story of Erin and Justin and their two daughters. Justin dies tragically in a skiing accident. A few years later, Erin gets very sick after a trip to Mexico. Antibiotics fails to cure her, and eventually she is diagnosed with a strangulated colon that necessitates emergency surgery. The surgery discloses cancer, which the surgeon is pretty sure he has removed, so chemotherapy may not be needed. The surgeon, however, recommends a genetic test to see if she has Lynch syndrome, a genetic disease predisposing one to early colon cancer (Erin is forty-eight). The genetic test does not find Lynch syndrome, but specialists recommend additional genetic testing. Emanuel tells this story to illustrate how much medical care costs, the incredibly complicated insurance system we have, the problem of defining a pre-existing condition, and the reluctance of many health care providers to see uninsured patients, even critically ill ones. The Affordable Care Act will make payment easier for many Americans. For us, however, the point of this story is that genetics is still seen as an afterthought in medical care, and that as cutting edge as it is, Emanuel never mentions genetics again in his book about our current health care system.
Emanuel’s failure to describe the genomic medicine of the future is understandable: it’s still too early to tell how this evolving area will play out, or even how it should be described. President Obama, for example, called it “precision medicine” in his 2015 State of the Union address. And what is still widely referred to as “personalized” medicine may actually be more impersonal than today’s medicine. Personalized, genomic medicine requires electronic health records, and you may already find your physician interacting more with a computer than with you during office visits. You will need a basic understanding of the new language of genomic medicine, and the most likely ways it could impact your care, to be able to navigate the evolving genomics-enabled health care system. Personalized medicine has been one of the most hyped “revolutions” in modern medicine. In this chapter we put the current state of personalized medicine into perspective and suggest ways you can use genomic information in everyday health care for you and your family. Our goal in arming you with this information is to help make you a better advocate for your own health and that of your family. We begin with the rhetorical shift from personal to personalized medicine.
Forty years ago, a grandfatherly physician from Kentucky came to speak to Sherman’s medical school class about his memorable patients. He also offered the medical students two pieces of advice. First, after entering a patient’s hospital room, always sit down, even if the only place to sit is the patient’s bed. Second, always touch the patient. (Today the physician would have to first wash her hands.) It doesn’t matter what you do: take his blood pressure, feel his abdomen, look into his ears, or just put your hand on his forehead to feel for a fever. The critical element is physical contact.
This is, we think, the kind of “personalized” medicine most of us hope for, if not expect, from our doctor when we are sick or hurt. Of course, we presume that our doctor is competent and skilled, and has access to the latest medical technology, but that’s not enough. We want a personal relationship with our doctor; we want high tech, but we also want “high touch.” None of us want “cookbook medicine”; we are all unique, and we rightly expect our doctor to take into account our medical history, together with the results of physical exams and medical tests, to develop an individual treatment plan that is optimal for us.
Several terms have been used interchangeably with personalized medicine, including precision medicine, stratified medicine, targeted medicine, and genomic medicine. The best definition of personalized medicine so far comes from the National Academy of Sciences: “The use of genomic, epigenomic, exposure, and other data to define individual patterns of disease, potentially leading to better individual treatment.” The last clause is not really part of the definition, but rather expresses the hope that genomics will improve patient outcomes. Personalized medicine is a misleading label for the use of genomics in health care because it implies that you are your genes, and by treating your genes physicians are treating you. This is far too limiting a view of personalized medicine. As physician Marshall Chin put it in the context of improving patient care: “Truly patient-centered care is individualized care by clinicians who appreciate [that] patients’ beliefs, behaviors, social and economic challenges, and environments influence their health outcomes.” Chin continues that the “best care” spans years and includes inpatient, outpatient, and self-care that is “tailored” to the individual needs of each patient.
All this helps explain why we prefer the term genomic medicine to personalized medicine: it more accurately states what it is, using genomic information (information from our genome) to help guide health care. To us, and most likely to you as well, personalized medicine conjures a different image, a more holistic medicine like that practiced by the grandfatherly Kentucky physician. Nonetheless, it is difficult and confusing to try to change labels once they have become part of our medical vocabulary. We use both personalized medicine and genomic medicine in this book (we don’t think the term precision medicine will have a long life), depending on which aspect of personalized genomic medicine is being emphasized. We encourage you to think genomics whenever you see personalized.
Genomic medicine does not always require sophisticated and costly genomic technologies. Your family history is the simplest, most straightforward, and least expensive way for you to assess the risks of inherited medical conditions. For all these reasons, it is important for your health and health care that you know your family history. Popularly called a family “tree,” it has its origins in genealogy, identifying your relatives by their names and recording your family lineage. Observations that certain maladies run in families have been documented throughout history, and most people believe they will develop the diseases their parents had. This is not always true, but understanding family relationships is the beginning of genomic medicine.
The Genetic Family History
Referring to direct-to-consumer genetic testing, one medical expert at the U.S. General Accounting Office went so far as to tell Congress that “the most accurate way for these [direct-to-consumer DNA profiling] companies to predict disease risks would be for them to charge consumers $500 for DNA and family-medical-history information, throw out the DNA, and then make predictions based solely on the family history information.” This is an exaggeration, but it properly emphasizes both the power of a family history and the insufficiency of existing data to support genetic testing inferences.
Taking a genetic family history involves systematically inquiring about your health status and ethnic origin, first-degree relatives (siblings, parents, and offspring), second-degree relatives (uncles, aunts, nephews, and grandparents), and third-degree relatives (first cousins). The birth dates for living family members, and age and cause of death for deceased family members, should be recorded. It is especially important to check if there is any relationship by descent from a common ancestor (called consanguinity or “blood relative”). It may be necessary to obtain medical records (including laboratory test results, imaging studies, and pathology reports) to confirm directly the diagnosis of a relevant disorder in a family member. Untoward pregnancy outcomes (for example, miscarriages, stillbirths, infants born with congenital anomalies) should be ascertained. Exposure to drugs (including prescription, nonprescription, and “street drugs”) and toxic chemicals should be determined. The amount and duration of alcohol and cigarette smoking should be evaluated. Depending on the situation, particular emphasis should be placed on obtaining detailed information about the health concern in question (for example, cancer, prenatal diagnosis, or dementia).
Genetic professionals usually assess the risks of inherited medical conditions by performing a pedigree analysis. A pedigree is a diagram of family relationships that uses standardized symbols to represent people and lines to represent genetic relationships. Pedigrees ideally show at least three generations, mark individuals affected with a specific diagnosis, and indicate ethnic ancestry (figure 2.1). This display can be helpful in identifying how genetic disorders are inherited. A pedigree can be easily drawn by hand and scanned into the medical record. In practice, most health professionals use questionnaires, often with checklists. There is an interactive software program that allows you to record your family history on a computer. One version, called My Family Health Portrait: A Tool from the Surgeon General, is available free online. There are also inexpensive apps for making your family pedigree.
2.1 Example of a pedigree showing type 2 diabetes and cleft palate. Abbreviations: dx, diagnosed; y.o., years old; P, pregnant. Personal image.
Compared to depending only on notes in the medical record, using a family history and pedigree tool increases the likelihood of identifying a person at risk for an inherited medical condition by about 20 percent. In fact, family history and other conventional tests (for instance, cholesterol testing and blood pressure) outperform genomic-based risk predictions for cardiovascular disease. When you hear about how genomic testing discoveries are giving us valuable information about the underpinnings of disease, including the promise that these discoveries will soon lead to long-awaited new treatments, you should be skeptical, like a good scientist. For many common diseases, such as type 2 diabetes and certain cancers, the initial hype about breakthroughs has not translated into meaningful changes in everyday health care.
As mentioned in chapter 1, the overpromising of direct-to-consumer company 23andMe led the FDA to warn the company to stop advertising its personalized genetic tests in late 2013. The FDA was concerned that the company’s national advertising campaign was making promises it could not keep, including that customers could “learn hundreds of things about your health,” for instance that you “might have an increased risk of heart disease, arthritis, gallstones, [or] hemochromatosis.” The company’s commercials also suggested that the genetic information they could derive from analysis of your DNA was sufficient to permit you to change your life and your health: “Change what you can, manage what you can’t.” The FDA was correct, we think, in effectively shutting down 23andMe’s disease-risk profiling service for now. Family history remains more informative than limited genetic screening—and family history can help determine whether genetic screening makes sense (for instance, in the case of the breast cancer mutations).
Physicians seem to agree, at least based on an unscientific poll of physicians conducted by the New England Journal of Medicine in late 2014. Readers were asked whether they would recommend genomic screening for a hypothetical healthy, asymptomatic forty-five-year-old male, Jim Mathis. Mathis told his internist he was concerned about his risk of cancer after completing a family tree at a genealogy workshop, where he learned that three relatives had cancer (breast, ovarian, and prostate). Of the approximately one thousand physicians responding, 40 percent would not perform any genomic testing, and most of those who chose to test would limit testing to cancer genes. Only 12 percent would do the entire genome. The summary of the results concluded by focusing on the patient rather than the physician respondents: “Many patients, like Mathis, are curious and will actively seek knowledge about their genetic risk.”
In the aftermath of the FDA’s action, 23andMe limited its services to building family trees, advertising that you can use its services to “build your family tree and enhance your experience with relatives.” You can also “learn what percent of your DNA is from populations around the world” and “contact relatives across continents, or across the street.” Other companies also offer family-tree-building services, including Ancestry, WikiTree, and Geni. Esquireeditor-at-large A. J. Jacobs has noted that by using the services of Geni, whose database currently has 75 million “relatives” from 160 countries, he was informed that he is related to more than 80,000 other humans. In his words, “My newfound kin include the actress and lifestyle guru Gwyneth Paltrow, a mere 17 steps away, and the jazz great Quincy Jones, a mere 22. . . . These folks have no clue who I am.” Jacobs envisions the near future when our family trees will include “all seven billion humans on earth”—since we are all genetically cousins—and asks, “If everyone is related, what does the concept of family even mean?”
It’s a good question, since humans are all genetically related at some level. One outcome of all this consumer genealogy research based on genetics could be a restructuring of groups that we identify with, away from groupings based primarily on geography or religion and toward groups based on genetic similarities. People want to know more than just which continents and what percentages of each are represented in their DNA. Medical researchers want to know more than that too. They want to know which people share which genes, so they can study how these genes are expressed in various environments. Members of genetically similar “groups” in the not too distant future may themselves form Facebook communities based on genetic similarities, as much out of curiosity as an exercise in family building. In short, we have so far used surrogate indicators like kinship, religion, and geographical origin to suggest what genes a person may share with others. Soon we will be able to sequence the entire genome of an individual, who can then directly compare his or her genome with that of others. Then, to the extent we want to, we can form new group identities based directly on our unique genomes and how they compare and overlap with the genomes of others.
It will be a while before genomic analysis can provide more useful information than the family tree, and in the meantime American physicians seem to be concentrating more on their new technologies than on us—the opposite of the personal medicine most of us want. One of the most famous physicians in the world, Arnold Relman, the late editor of the New England Journal of Medicine, describes the technically brilliant lifesaving treatment he got at Massachusetts General Hospital after he broke his neck in a fall in 2013. There were no personal conversations with physicians, who spent most of their time “with their computers” and examining “copious reports of the data from tests and monitoring devices.” Relman notes, “What personal care hospitalized patients now get is mostly from nurses.” Nurses, of course, have always been the 24/7 caregivers in hospitals, but have they now almost entirely taken over the job of communicating with patients as people? Can genomics help reverse the trend of physicians in hospitals concentrating more on the computer and the patient’s electronic medical record than on the patient him- or herself, or will genomics make medical care even less patient centered? We think that genomics, at least in the immediate future, will encourage the trend toward more impersonal medicine (rather than personal medicine), and only strong patient and family voices are likely to reverse this trend.
When you are admitted to the hospital, you can expect “Big Medicine.” Many economists, and even some physicians, think that giant private corporations, which see maximization of profits as their primary goal, should be seen as role models for health care. We strongly disagree. We should be maximizing the health of Americans, not the profits of their caregivers. Surgeon-author Atul Gawande, for example, has suggested that our health care system has much to learn from the Cheesecake Factory, whose 160 restaurants serve more than 80 million people a year. Gawande believes that one key to the success of the Cheesecake Factory is its size, that economies of scale permit it to provide food and service of greater quality at lower cost. Size gives their restaurants buying power, allows them to centralize common functions, and enables them to adopt and diffuse innovations more quickly than do small independent operations. Our health care system has a similar goal: trying to deliver a range of services to millions of people at a reasonable cost and with a consistent level of quality. Yet American medicine has been largely unsuccessful in meeting these goals, producing instead skyrocketing costs, mediocre service, and unreliable quality.
How does genomic medicine fit into the Cheesecake Factory model? A key to the success of the Cheesecake Factory is standardization: making sure each dish looks and tastes the same time after time. This means that there can be no room for varying the recipe. The Cheesecake Factory has thirty types of cheesecakes on its menu. What if there were 13,360 kinds of cheesecake? In a discussion at Harvard Business School, Gawande noted that “healthcare is incredibly complex simply because of the myriad ways—13,360 and counting—that the human body can fail.” He observed that the delivery of health care is “arguably the delivery of 13,360 service lines, town by town, to anyone who needs care.”
Now think about what would happen at the Cheesecake Factory if each cheesecake had 3 billion ingredients, just as the human genome has 3 billion base pairs, and no two cheesecakes had the same ingredients, just as no two individuals have exactly the same genome. No two cheesecakes would look and taste exactly the same, just as no two people are exactly the same. The point is that there is no standard genome; personalized genomic medicine is directly dependent on DNA variation. As genomic medicine is increasingly used in health care, the focus will be on your genetic uniqueness as spelled out by your DNA. Genomic medicine will be adding a lot more information to a system already overloaded by more information than any other industry. This flood of genomic information will require sophisticated bioinformatics and computer-based translation in the doctor-patient relationship. We think genomic medicine will eventually make for better medical care, but it will not make medical care less expensive or more efficient, especially in the short run.
We can accept the Cheesecake Factory as a good model for the hospital cafeteria without applying it to the intensive care unit or the operating room. The real challenge for our health care system will be how to adopt many of the good features of megacompanies like Walmart and Amazon, such as emphasis on quality measures, cost efficiencies, and customer satisfaction, without losing sight of the fact that every patient really is unique (and not just genetically) and deserves to be treated and respected as an individual. We remember, for example, the hospital administrator who, in the midst of the 1990s fad of managed care, said he didn’t see any reason why a hospital could not be run like a ball bearing factory. And he would have been correct if ball bearings got sick, suffered, died, and had families that cared about their welfare. There are, nonetheless, routine steps that should be taken with every patient. Gawande is on firm ground here when he proposed that surgeons adopt checklists, like those used by pilots before flight, to use before they begin surgery. Having all surgeons check to make sure they are doing the right operation on the right patient at the right site, for example, makes perfect sense and should be mandatory.
Other experts, like author Robin Cook and physician Eric Topol, see a future in which your smart phone becomes an “avatar physician.” It will monitor your heart rate, respiration, and so on, and use this data to determine if you are having a heart attack and call 911. It may also be able to predict heart attacks or other conditions and warn you about them. Topol also envisions everyone having their DNA sequenced and made part of their electronic health record so that clinicians can “match individual with treatments.” The goal is not, Topol assures us, to replace your physician with your smart phone and computer, but, with your smart phone doing the routine monitoring and testing, “your flesh and blood primary care physicians will have more time to talk to you when you do need to see them.” We’re not convinced this is medicine’s future, but we concede that self-monitoring devices are proliferating. You may even want to try one yourself since, for at least some people, they can help provide the motivation to increase exercise or change to a healthier diet. We strongly agree with Topol that your medical and genomic information is yours and you should have complete and direct access to it.
Genetically Isolated Populations
Studying isolated populations has produced much of what we know about genetics and heredity. Understanding why will help you understand much of the genetics in this book. It has long been known that in animal communities inbreeding leads to problems and outbreeding leads to “vigor.” The same can be said for human populations. You and your first cousin can expect one-eighth of your genes to be identical by descent from the same ancestor. If individuals in a small community marry only among themselves, over generations everyone becomes genetically related. From a shared gene perspective, all marriages are among cousins. An examination of the family trees (pedigrees) of couples living in small, isolated communities reveals that their roots appear to twist back into themselves multiple times. This results in “genetic bottlenecks” within the population. For example, grandparents might be related to each other in multiple ways; for a couple marrying, each grandparent may be related to the grandparents on the other side. It’s easy to see how, generation after generation, the likelihood of a couple sharing the same ancestral genes steadily increases.
This inbreeding effect has led some autosomal recessive diseases (diseases expressed only if the person has gotten the same mutated gene from both parents and is thus homozygous for the mutation) to be more common in particular populations. If a couple shares a mutation in the same gene, there is a one-in-four chance of having a child affected with the disease coded for by that gene. Examples include Tay-Sachs disease among Ashkenazi Jews; beta-thalassemia among individuals of Mediterranean descent; cystic fibrosis among individuals of northern European descent; sickle cell anemia among individuals of African descent; and alpha-thalassemia among individuals of Asian descent. A number of unique isolated populations have been studied extensively in genetic research. We describe three: the Mormons, the Amish, and the Hutterites.
Geneticist Ray Gesteland of the University of Utah believes that “more human disease genes have been discovered in Utah than in any other place in the world.” This is because of three large data banks: the Utah Population Data Base, the Utah Cancer Registry, and more than 185,000 genealogical records from the Mormon Family History Library. For more than three decades, researchers at the University of Utah have collaborated with the Mormon Church in numerous genetic investigations, particularly in cancer research. Mark Skonlick, a geneticist at the University of Utah, for example, used Mormon medical records and pedigree information to help establish Myriad Genetics in 1990. The company later identified and isolated two breast-ovarian cancer genes, BRCA1 and BRCA2.
Another religious group, the Amish, have 275,000 members in more than thirty states. The Amish avoid most modern technology and prohibit or limit \use of telephones, television, and automobiles. Genetic studies on the Amish began in the early 1960s, conducted by Victor McKusick of Johns Hopkins University, after David Krusen observed that a form of dwarfism (originally described as achondroplasia but later diagnosed as Ellis–van Creveld syndrome) was frequent among the Amish. Cartilage hair hypoplasia (CHH), an autosomal recessive form of dwarfism associated with immunodeficiency and higher risk of developing lymphomas and leukemias, was first characterized in the Amish. McKusick found advantages in doing genetic studies with the Amish: their genealogic records are extensive; they are interested in the causes of illness; there is a high coefficient of inbreeding; they are relatively immobile; and they keep children with genetic abnormalities at home rather than institutionalizing them. Nor do the Amish object to modern medicine. The Clinic for Special Children in Strasburg, Pennsylvania, may be the only clinic in the world where you will find DNA sequencers inside and hitching posts outside.
Earlier in his career, Sherman was involved in a research project at Northwestern University studying the Hutterites (figure 2.2). The Hutterites first settled in what is now South Dakota, establishing three communal farms (colonies). When a colony reached a certain size, usually around 150 people, it would split. By drawing lots, half the families would remain in the old colony and half would move to the new one. Today there are more than 40,000 Hutterites living in almost four hundred colonies. Hutterite colonies often have a one-room schoolhouse where the children are taught secular subjects for half the day and religion the other half. Hutterites follow a communal lifestyle with shared goods, eating in a dining hall where the men and women sit apart. Children eat separately.
The Hutterite communal lifestyle is ideal for human genetic studies: they are exposed to similar environments, eating an old-style Germanic diet, with no smoking, no birth control usage (the average Hutterite couple has five children), and only occasional alcohol consumption. Most importantly, virtually all current Hutterites are descendants of the original founding population. The colony preachers keep meticulous genealogy records and have recorded the marriages, births, and deaths of colony members for hundreds of years.
2.2 Sherman with the Hutterites. Personal photo.
Over the years, Sherman drew blood samples from hundreds of Hutterite men, women, and children. Coincidentally, Sherman’s last name, Elias, is a common Hutterite name. Members of the visited colonies often forewarned new colonies that a doctor named Elias would be coming. The Hutterites would frequently ask if he had Hutterite ancestry, and when he told them he didn’t, they told him he just needed to look harder. The extensive pedigrees and medical histories of the Hutterites, linked with the studies performed on their blood samples, permitted invaluable insights into the genetics of such conditions as diabetes, asthma, breast cancer, and repeated miscarriages.
Two main points from the genetic research done on these populations merit emphasis. The first is that a key element of research in genomic medicine is obtaining an accurate, extensive family history. This is why unique populations such as the Mormons, Amish, and Hutterites are of such interest to genomic researchers. With or without genomic information, family history (including ethnicity) can provide highly valuable information for you and your physician that can directly affect your health care. The second is that genetic conditions can be inferred from family histories and later confirmed by genetic analysis. As whole-genome screening becomes more widely used, however, genomes can be (and will be) compared directly, without the need to look for kinship relationships. For example, in 2014 another “isolated” population, the indigenous population of Saudi Arabia, embarked on a project to do whole-genome sequencing on 100,000 members of this population to determine to what extent marriages among first cousins resulted in genetic abnormalities in their children.
Electronic Health Records
Genomics will be a key element in the development of evidence-based medicine because individuals with varying genomes will respond differently to the same drugs. In A Study in Scarlet, Sherlock Holmes observes to Dr. Watson, “It is a capital mistake to theorize before you have all the evidence. It biases the judgment.” Unfortunately, for a large part of medical practice, there is surprisingly little evidence for what works, or doesn’t work, and what is best for individual patient care. Recognizing this, about twenty years ago a small group of medical researchers and educators began advocating for what is now known as evidence-based medicine (EBM). Simply put, EBM “is the integration of best research evidence with clinical expertise and patient values.” The goal is to improve patient care.
In comparative-effectiveness research (CER), patient outcomes of one approach for managing a disease are compared to other approaches—for example, comparing the effectiveness (in terms of a defined outcome, such as survival rates) of two or more drugs for the same disease. CER is an effort to find out what works best, but it can also be an effort to save money by not paying for useless or marginal care. This approach can help us determine if we are getting the biggest bang for our health care dollars. Collecting data on treatment outcomes is essential to improving the quality of medical care, and we think you should enthusiastically support this effort. It is widely recognized that much, if not most, medical care today can accurately be described as “gray” care—care that is routine, does little if no harm, but does not do much, if any, good. The goal with genomics is to introduce it into clinical medicine in a way that both improves patient outcomes and is cost-effective.
Some treatments and screening tests can do more harm than good. Prostate-specific antigen (PSA) screening for prostate cancer is a good example. In 2012 the U.S. Preventive Service Task Force made a highly controversial recommendation that affects nearly 45 million men annually. After decades of routine use in clinical practice, the task force recommended that we abandon screening for prostate cancer with PSA blood testing because it causes more harm (for example, unnecessary biopsies, overdiagnosis, overtreatment, and complications, such as urinary incontinence, erectile dysfunction, and bowel dysfunction) than benefits (that is, potential survival advantages from treating prostate cancer). Even the suggestion that men give up annual PSA screening met with stiff resistance, particularly from the American Urological Association and prostate cancer advocacy groups, who have a financial stake in not changing current payment policies. Understandably, this conflict has led to considerable confusion, frustration, and even anger among men. Currently we are left with a middle-of-the road “shared decision process” that satisfies no one. To share decision making with your physician is fine, but the decision is not an informed one without the relevant evidence. For more on screening tests, see Appendix B.
How does taking a simple family history stack up in this era of EBM and CER? And how does a family history compare with genomic testing? In both EBM and CER, groups of patients, not individual patients, are analyzed to compare the effectiveness of alternative medical treatments. This may seem to be at odds with the fundamental tenets of genomic medicine, where the focus is on the individual patient’s unique and specific disease characteristics, coexisting conditions, genetic factors, risk factors, and personal values and preferences. Mostly, however, genomics will be used not to treat you uniquely but to treat you the same way as other people with your relevant genetic traits. In this respect, “personalized” medicine can be seen as genome-based group or sector medicine (“stratified” medicine), with the group or sector of the population based on genetic commonalities, similar to those identified in the four genetically isolated groups we discussed.
Proponents of genomic testing, particularly direct-to-consumer companies, have suggested that personalized medicine will encourage people to positively modify their lifestyles, especially making changes in diet and exercise. However, the same can be said for taking a family history. One randomized controlled trial showed that systematically screening for family history and tailoring prevention messages can be effective in improving health behaviors by increasing physical activity and fruit and vegetable intake. If all you get from having your DNA tested is a recommendation that you should exercise more and eat a healthier diet, taking a family history is much more cost effective. Of course, it’s another matter if genomic testing (which can be indicated based on a complete family history) can establish that you carry a particular deleterious gene, such as the BRCA1 or BRCA2 mutation. This mutation puts you at substantial increased risk of developing breast cancer, ovarian cancer, or both, and its identification could lead you to consider proven, effective preventive measures, including the drastic step of having a bilateral prophylactic mastectomy and removal of the ovaries, as illustrated by Angelina Jolie Pitt (discussed in chapter 1).
As we have suggested, genomic information will become much more useful to you and your physician when it is integrated into your medical record. This will not be realistic until your medical record is kept digitally. Most hospitals either have converted or are in the process of converting their medical records (now usually called a “health record”) from paper to computer. Electronic health records (EHR) are a digital version of the old paper records. In addition to being more legible than handwriting (which can prevent medical errors), EHRs have many advantages over paper records. EHR allows computer storage of medical records, laboratory tests, digitalized images (for example, X-rays, ultrasounds, and MRIs), electrocardiograms, fetal heart rate tracings, and so forth. This requires far less space and lower associated costs compared to storing hard copies. EHRs allow rapid retrieval, searches, and collating of medical information. They permit your doctor to easily chart information about you over time (for instance, blood pressure and weight) and track when you are due for screening tests (such as a mammogram or colonoscopy). In a hospital, EHRs can also be used to track infections, as well as trends in admissions and discharges. In fact, your records are designed to be accessed by everyone involved in your health care, including you. Because of incompatible systems, however, the information in EHRs doesn’t travel easily out of the practice or hospital.
Transition to the use of EHRs is well under way. The adoption of EHRs by the American health system will expedite the introduction of genomic medicine into everyday practice. The likely future interaction of the EHR and genomic information is illustrated by the medical treatment of a hypothetical patient we’ll call Olivia. Her treatment can be seen as a best-case scenario of the future of genomic medicine. Olivia’s case can be usefully contrasted to the case of Erin, the patient whose story we related at the beginning of this chapter, who had no unified health record, electronic or otherwise.
Olivia came to the United States with her family from China when she was four years old. Shortly after she turned seventeen, she suffered a grand mal (generalized tonic-clonic) seizure. Using an electronic referral system, her primary care physician sent Olivia (and her EHR) to a neurologist, whom we’ll call Dr. Good. At a university medical center, Good performed a number of tests, including an electroencephalogram to measure electrical activity in the brain. Good decided that the best drug for her condition would be carbamazepine (Tegretol®) and entered the prescription in Olivia’s EHR. The software performed a “carbamazepine pharmacogenetic adverse event algorithm,” which required Good to enter a number of mandatory “data elements” for the “decision model” before the pharmacy could accept the prescription.
A history of allergy to carbamazepine was queried, as well as any potential interactions with other drugs Olivia was taking. To determine the correct dose, the computer needed the patient’s height and weight. The possibility of pregnancy was important because fetal exposure to carbamazepine is associated with an increased risk of birth defects (including head or facial deformities, spina bifida, and heart defects). Olivia was a young female, so the computer prompted Good to order a pregnancy test. Personalized genomic medicine also came directly into play. Olivia’s Asian heritage produced a prompt to check her HLA-B*1502 status, which increases the risk of developing a potentially deadly blistering skin reaction to carbamazepine called the Steven-Johnson syndrome. Because she had this genotype, the computer suggested alternative drugs, and Good ordered one of them. Olivia’s treatment illustrates how the EHR will advance genomic medicine by embedding a patient’s genomic information into clinical decision making. Decisions affected include disease risk assessment, accurate disease diagnosis and subtyping, drug therapy and dose selection, assessment for adverse drug reaction, and family planning. The EHR has the potential to bring a tremendous amount of “personalized” information directly into the doctor-patient relationship.
Integration of genomics into medical care is evolving and will be gradual. It is part of overall efforts to improve patient outcomes and confront ever-increasing health care costs. Increasing the quality of medical care, however, does not automatically decrease its overall costs. Medical care is not like making cheesecake. Few of us want to go to Walmart for our physicians and hospitals, looking for the cheapest health care possible. It is likely that more information will lead to more expense. A study of 30,000 patient visits to more than 1,000 office-based physicians, for example, found that use of EHRs resulted in ordering additional imaging and laboratory testing.
The genomic information that will be included in your EHR will evolve. At first, as with Olivia, information about particular genes will be entered. Eventually, we think likely within a decade, we will transition to entering your whole-genome sequence, all 3 billion base pairs. There is still a question as to whether the “raw” genetic data will be entered into the EHR itself or housed in a cloud-based or other storage system separate from the physician and hospital records. Regardless, as genomic information is used more widely by clinicians, systems will have to be in place to regularly “interrogate” your genome. Some of this information will be of critical importance to your health care, but much, if not most of it, will be useless, and some will actually cause confusion and even potential harm. The challenge, of course, is to maximize the useful genomic information and minimize the marginal, useless, or even potentially harmful genomic information. As the director of Duke’s Center for Human Genome Variation accurately predicted in 2011, “Within the next few years our ability to identify pathogenic and potentially pathogenic mutations—as well as huge numbers of mutations that no one can vouch for as dangerous or safe—will almost certainly outstrip our ability to act on the information.” Nonetheless, we expect that within a decade or less everyone (at least all adults) will have direct access to their entire medical records by personal computer, including their genome. This will put a lot of discretion in your hands, and you will have to decide how much you want to know about your genome. Your answer will likely be highly influenced by what actions, if any, you can take to modify the effects of your genes on your health. Answering this last question is what most contemporary genomics research is about.
In the chapters that follow, we also address the hype of genomic medicine and how it has been oversold in large part to maintain public support (and more importantly, federal NIH funding), as well as by the promise of financial bonanzas for private industry. This doesn’t mean that genomic medicine is not going to be useful. Just the opposite: we believe it is evolving to be a major force in medical practice. It will likely affect virtually every aspect of how we diagnose, treat, and prevent illnesses and medical conditions—everything from diabetes, to infections, to infertility, to prenatal care, to dementia, to cancer. In other words, we think genomics will follow the path of most transformational technologies, whose impact is overestimated in the short run but underestimated in the long run.
WHEN THINKING PERSONALIZED
CONSIDER THESE THOUGHTS
Whenever you see “personalized” or
“precision” think genomics.
Your personal family history can often
tell you more than your genome.
Your family history can help determine
whether to have genomic testing.
Much of what we know about genomics comes
from the study of isolated populations.
Genomics will not be used routinely in medicine
until your genomic sequence can be stored on,
or linked to, your electronic health record.