The Value of Genetic and Genomic Technologies: Workshop Summary (2010)

For the third clinical scenario, Bruce Blumberg of Kaiser Permanente presented a hypothetical case of direct-to-consumer genetic testing that was designed to raise a variety of issues that occur in real-life cases.¹ Genomic analysis for the prediction of common disease risk is controversial. Proponents support an individual’s right of unrestricted access to his or her personal genetic information, while opponents stress the lack of consensus on the genetic markers used in genomic profiling and the inconsistency of risk predictions. In addition to the ethical, legal, and social issues surrounding genomic profiling, questions persist regarding the clinical utility, safety, and cost-effectiveness of genomics-based risk assessment.

A number of clinical laboratories offer direct-to-consumer genomic profiling either for risk assessment for common diseases or for carrier status for less common Mendelian disorders. Currently, Blumberg said, the three most prominent genomic profiling companies are Navigenics, 23 and Me, and deCODE, and there are ten diseases that all three companies include in their risk assessment panels: age-related macular degeneration, atrial fibrillation, breast cancer, celiac disease, Crohn’s disease, prostate cancer, psoriasis, rheumatoid arthritis, type 2 diabetes, and deep vein thrombosis. There are also numerous other diseases that are included in the panels of only one or two of the companies. Examples include Parkinson’s disease,

Alzheimer’s disease, lupus, osteoarthritis, multiple sclerosis, lung cancer, kidney stones, gallstones, and gout.

To facilitate the panel discussion of genomic profiling, Blumberg presented the fictional case history of “Anne,” who is intended to be representative of the average consumer who might have his or her genome sequenced at a commercial facility. Box 4-1 provides definitions relevant to the discussion.

Anne is a recently divorced 36-year-old MBA financial analyst. She has always considered herself to be both healthy and health conscious. She is an only child, her mother is of English ancestry, and her father of mixed Eastern European descent. Her past medical history is notable only for a mildly abnormal glucose tolerance test during her second pregnancy at age 31, which was not medically followed after the pregnancy. Anne prides herself on her careful diet, and she runs on a treadmill at her workplace gym at least three times a week. Despite these efforts, she is 15 pounds overweight according to a table that she found in a popular magazine. She has never smoked. She has been tired lately, which may be caused by the demands of juggling single motherhood with a career. Anne’s mother is 67 years old and was treated at age 59 for melanoma, but Anne knows no further details. Her mother has recently had mildly elevated blood sugars,

and her doctor is considering oral hypoglycemic therapy. Anne’s father is 70 and is taking a statin for hypercholesterolemia and a beta blocker for hypertension. He had a very mild heart attack in his late 40s, and he takes prophylactic aspirin. Two of the father’s maternal first cousins are said to have died of colon cancer in their 40s, but, again, no details are available. His paternal aunt died of breast cancer in her late 30s. There may be other relevant conditions in the family history, but no physician has ever asked Anne about this.

Anne has read articles in the New York Times and elsewhere about the availability of genomic screening for health risk assessment. As is her usual practice in health-related matters, Anne has extensively reviewed the topic on the Internet and compared the tests offered by several different companies. Based on her personal and family histories, she is especially concerned about her future risk of diabetes and coronary artery disease, so she selects the laboratory that places the greatest emphasis on these conditions on its website, and she submits a sample prior to leaving for a vacation. Upon returning, a printed report awaits her. If she understands the report correctly, she is relieved to learn that her risk for type 2 diabetes is 10 percent below that of the general population. On the other hand, her risk of developing coronary artery disease sometime in the future is 20 percent above that of the general population risk. It is unclear from the report whether the risks have taken her family or personal histories into account or whether the calculated risks are based exclusively on the genomic results. As she continues to read the report, Anne learns that her breast cancer risk is 30 percent above the risk for the general population, and she is dismayed to read that her Alzheimer’s disease risk is double the general population risk. Finally, she is surprised and confused when she reads that she is a carrier for hemochromatosis and alpha-1 antitrypsin deficiency, two conditions with which she is entirely unfamiliar. She wonders if these findings might explain her recent fatigue. Anne immediately calls her doctor’s office, but the earliest available appointment is in 2 weeks. When she arrives for the appointment with her report in hand, she appears to be mildly agitated.

Janet Williams, a genetic counselor at Intermountain Healthcare, said that the first challenge in working with a patient like Anne is knowing what the SNP results mean in terms of actual clinical risks in the future. A genetic counselor would look at the family history and deal with known risks, such as cancer, diabetes, and coronary artery disease, and any SNP profile results would be considered in this context and also in the context of what can

be measured clinically. The real difficulty for a counselor is providing the patient with a useful risk estimate that takes all of the relevant risk information into account. Anne is very health conscious and very motivated by the information she has received. She thought she had her lifetime risks well identified, but this profile has opened up an entirely new set of concerns for her, and it has potentially distracted her from other issues that are still appropriate for her to be concerned about, such as the family history of diabetes. At this time, Williams said, there is not enough evidence concerning how to use genomic profiling results to provide useful clinical information for a patient. Thus, counselors are left in the position of not having an appropriate response to the patients’ questions and concerns.

David Witt, a clinical medical geneticist at Kaiser Permanente, said that while he is generally enthusiastic about bringing new discoveries to the clinical setting, he is not enthusiastic about the type of genomic profiling described in the scenario. The case presented illustrates some of the many pitfalls that can occur in screening, demonstrating why it currently has very limited value and in some cases can actually be harmful. Most of the risk estimates that can be obtained today through genomic profiling, Witt said, are so minimally useful that a person does not need to be concerned with them. Furthermore, the reported risks are often already being addressed based on family history information. Witt proceeded to review the profile results in detail and highlighted some of the issues each of the findings creates, especially how the reported risk levels can be confusing and easily misinterpreted and can possibly lead to unnecessary testing or procedures.

Interpretation of the results can pose a significant issue for a patient. Concerning the findings for hemochromatosis and alpha-1 antitrypsin deficiency, the patient has received information that was not sought and which is not understood, as indicated in this case by Anne’s lack of a clear understanding of the difference between being a carrier and having an expressed condition. This led Anne to seek counseling from her primary care physician, thus taking up the provider’s time to explain the results, assuming that the provider is capable of doing so, Witt said. Furthermore, the screening results have created unnecessary anxiety about something that may have limited personal or familial value.

Another important thing to consider, Witt said, is that the risks presented are based on the current markers that the individual laboratory has selected to test, but as more and more markers related to chronic diseases such as diabetes and heart disease are found, the results today could be modified in the future to indicate either a higher or lower risk or else could be invalidated altogether. Furthermore, how the testing company

presents the results can have a significant effect on the patient’s perception. The percent risk scores given for diabetes (10 percent lower), coronary artery disease (20 percent higher), breast cancer (30 percent higher), and Alzheimer’s disease (double) actually represent very small differences in actual numerical risk. For diabetes, this patient’s risk compared to the general population decreases from 5 to 4.5 percent, and for breast cancer has increased from 10 to 13 percent. None of these changes are overwhelming, Witt said, and they should not override any of the current recommendations for standard of care or for practices based on family history. An argument could be made that a patient perceiving a higher risk would be inspired to do more preventative practices, such as increased breast exams, but the misunderstanding about what the presented risk scores mean could cause significant emotional distress or lead to a demand for screening tests that are not warranted, Witt said. Additionally, there may be little positive value in identifying an increased risk for a disease such as Alzheimer’s, for which there are currently no options for screening or intervention.

Witt questioned what information of real consequential value Anne received for the $500 or $2,500 she spent for the genomic profiling. She received some risk figures that she does not understand and which are perceived as being more substantial than they really are. These results won’t translate into any significant change in her healthcare management, nor will they have any practical consequences that will significantly influence her overall health in terms of quality of life or duration. At the same time, however, these results have created significant anxiety for her, and they have an additional cost in terms of the healthcare provider’s time. Finally, there is the danger of an uninformed or inexperienced provider advocating that Anne or her relatives receive additional genomic testing.

Karen Kaplan, a science reporter for the Los Angeles Times, provided perspective on the public understanding of genomic profiling. The SNPs that are being used were identified in a GWAS study and turned into a commercial product, she said. Kaplan agreed with Witt’s analysis that these profiles don’t provide meaningful information about personal health. Additionally, a lot of consumers who are thinking about having a genomic profile done do not consider whether a test has been FDA approved. Often, just the fact that something is expensive causes some people to believe it is legitimate.

Kaplan mentioned that there have been commentaries published in the New England Journal of Medicine, JAMA, and other medical journals advising physicians what to do if they find themselves faced with this type of scenario. Most physicians do not receive extensive genetics training in

medical school, and many are just as confused as the patient about how to deal with these test results. The advice that is routinely given in these journals is to explain to the patient why he or she should ignore the results and to say that they do not supersede anything the patient already knows from family or personal medical histories. Kaplan speculated that as the cost comes down, more people will have genomic profiling done. There have been efforts to present it as a “fun” thing to do—spend a few hundred dollars and compare your SNPs to your friend’s SNPs. The attitude has been “What could be the harm?” The harm, Witt said, is that people might discover something completely unexpected that they are not prepared to deal with. Individuals who approach this seriously should talk to a genetic counselor before they begin the process, Kaplan advised.

Steven Woolf, a professor in the departments of family medicine, epidemiology, and community health at Virginia Commonwealth University and a family physician with a background in the evidence-based evaluation of screening tests, suggested that the roundtable should consider the issues of genomic profiling within the established frameworks for evaluating screening tests. Regardless of the type of test, he said, there is a standard set of analytic principles that are routinely applied when evaluating screening tests. Groups such as the U.S. Preventive Services Task Force and the World Health Organization generally consider five issues when assessing preventative interventions: (1) the burden of suffering from the target condition; (2) the accuracy and reliability of the test; (3) the effectiveness of early detection of the condition; (4) potential harms; and (5) the balance of benefits and harms (USPSTF, 1996). Most of what has been discussed thus far at the workshop fits into these categories, Woolf said.

The burden of suffering from the target condition is relevant because many of the diseases for which there are genomic tests are inherently serious. With regard to the accuracy and reliability of the tests, Woolf said that one should consider sensitivity, specificity, and positive predictive value. According to fundamental Bayesian statistical principles, if the condition has a low prevalence, even a test with very high sensitivity and specificity can produce a very high proportion of false positive results. Thus, there can be a very low positive predictive value even with a highly accurate test, if applied to a condition with low probability. This is important to consider for some of the conditions reviewed at the workshop, Woolf said.

One should also carefully scrutinize the reproducibility and predictive properties of genomic tests in terms of the precision with which they predict the future development of a disease. Woolf mentioned the Bonferroni correction, a statistical method employed when considering a test that looks

at multiple parameters. For example, the probability of producing erroneous or spurious information is statistically increased when multiple tests are done together (as occurs in a chemistry panel or whole body imaging). Woolf questioned whether this concept might be applicable to genomic profiling as well.

Is there a benefit to early detection of the particular condition? There has been a perception in society and in the medical community that knowing one has a disease or a risk for a disease has inherent benefits, regardless of whether the testing actually leads to improved health outcomes. However to be considered clinically effective, a screening test must improve the likelihood of positive health outcomes. Evidence of an association does not necessarily imply that there is a benefit that can be employed. Optimal study design, modeling, and the role of intermediate health outcomes as opposed to distal health outcomes should all be considered when assessing the effectiveness of a test. Relative versus absolute benefit is very important to consider, Woolf said, as is efficacy versus effectiveness. (Efficacy is the performance under ideal conditions and effectiveness is how well the test performs in real-world settings.)

When considering harms, one should include both the immediate harms of the test experience itself and the harms of the downstream cascade that may be set in motion by the test results (ranging from patient anxiety to the distal effects on employers and insurance eligibility). Potential harms also include false reassurance (e.g., a patient might decide she does not need to keep using the treadmill three times a week because of a particular test result) and false resignation (she might conclude, based on the test results, that there is nothing she can do to prevent the inevitable occurrence of disease). Costs are sometimes considered as part of the harms, but Woolf noted that there is some controversy as to whether they should be.

Finally, the balance of benefits and harms is complicated. The typical advice is that the clinician and patient should work together to review benefits and harms and consider personal preferences, and together they should make the choice that is best for the patient, as is done in genetic counseling. Unfortunately, with direct marketing of genomic profiling to the public, that collaborative decision-making process is bypassed and the consumer is exposed to these tests without that benefit.

Woolf pointed out that many of the fundamental issues raised during the workshop discussion apply broadly to all areas of medicine. For example, the need for infrastructure for improving the quality of care (e.g., anticoagulation clinics) is something being dealt with throughout medicine. The need for faster, real-world research on effectiveness, the idea of practice-based research networks, and the use of other venues for collecting real-world data are all being studied in health services research across many topics. Nor is the need for helping patients to make better choices

about complicated trade-offs unique to genomics. These are system redesign issues that are very important for health care, and they will certainly benefit genomics, but they need to be dealt with generically in medicine and are not necessarily specific to this particular topic.

Vance Vanier, CEO and president of Navigenics, said that more than 98 percent of the company’s business comes from national physician groups and medical directors of large self-insured employers. An in-house team of genetic counselors is available for both pretest counseling and, for 1 year afterward, for post-test counseling. To date, only a small subset of possible conditions that are deemed clinically actionable are available in the profile. The business is now regulated under the state versions of the Clinical Laboratory Improvement Amendments (CLIA), and Navigenics is currently the only company offering these tests to be approved in all 50 states. Navigenics does not test minors, Vanier said, and patients receive updates relevant to their genome so that they can take advantage of new discoveries.

Vanier said that when the company first launched its genomic profile product, it was popular for reporters to be tested, to speak with one of the genetic counselors, and then to write about their experience. He recalled one particular reporter who was homozygous for every marker included in the screen for macular degeneration, which increased her risk by a factor of between 5 and 10. However she was more interested in the fact that her risk of Crohn’s disease was 0.1 percent higher than normal. When questioned if she had any concern about the macular degeneration risk, she responded that she had talked with her doctor and he had told her to eat more spinach, so she was comfortable with that result. Later she mentioned that her sister has Crohn’s disease, so that was much more emotionally meaningful for her, and she interpreted the Crohn’s risk very differently than a physician or a genetic counselor would.

There are some who are of the opinion that it is “paternalistic” to assume that physician involvement is needed and who argue that if people can understand baseball statistics they should be able to understand genetics statistics. Navigenics is not of that opinion, Vanier said, stressing that physician involvement in genomic testing is extremely important, particularly physician education and medical alignment with regard to such testing. Overselling the usefulness of the genetic testing in this early period would be extremely detrimental, Vanier said. The question that needs to be addressed is how to navigate the course to the distant future when genomics will provide the preventative measures it has the potential for.

Behavioral change, Vanier said, is one outcome of genomic profiling that is frequently overlooked. Emerging data suggest that showing people

their genetic profiles is motivating and can compel them to do all the healthy things that physicians have been telling them to do for years but that they never do (e.g., exercise, diet, medical compliance). There are some who believe that if genomic profiling even slightly increases patient compliance and positive behavioral change, it will have been of enormous value to society, regardless of how the screening question plays out.

In terms of evidence of clinical utility, genomics screening is no different than PSA testing or mammograms in that it will take a long time to prove its value. Clearly, genetic profiling is not a substitute for family history, but family history has its own limitations as a screening tool. How often, for example, do people really know what their grandparents died of? How often do physicians in the average office encounter manage to get a good family history? As we move toward the future, Vanier said, we will need to look for the small proof points and small focused areas of utility along the way that can make a difference and that can accelerate the adoption of genomics in a clinically useful manner.

Marc Williams, moderating the discussion, noted that the cost of genotyping is rapidly dropping to the point where whole genomes can be sequenced as cheaply as, if not more cheaply than, single genes. How do we balance the desire to achieve economies of scale with the potential for downstream costs associated with obtaining volumes of information that really are not needed or even understood yet?

Woolf answered that the potential for economies of scale to make profiling more broadly affordable compounds his concerns about the potential misinformation that could be generated. The bottom line is whether there is evidence that performing a test is going to make people healthier. Woolf said that he supports research to collect that evidence, but as a policy matter, until the evidentiary threshold is reached, it is premature to advocate such consumer-oriented genomic tests, whether in isolation or as inexpensive sets that invoke economies of scale.

In some cases, Vanier said, the price that a consumer pays out of pocket for a bundled test is now less than what the laboratory would charge through a traditional third-party reimbursement system for a single indication. The cost of genotyping is coming down at such a brisk rate that within the next 12 to 24 months, for the same cost that one could obtain pharmacogenomic information for use in warfarin dosing, one could sequence an entire genome and at least acquire a dozen other pharmacogenomic indicators as well. Given the concerns raised with waiting five

days to get warfain pharmacogenomic results back from the laboratory, he asked, would it not be better to have all of the pharmacogenomic markers available in the electronic medical record, ready for when the physician has another drug to prescribe? Janet Williams responded that this assumes a level of data retrieval and cross talk between electronic medical record systems, laboratory systems, and pharmaceutical entry systems that does not exist right now.

Even if genome sequencing were free and entire genomes were decoded at birth and stored on a medical ID bracelet, Kaplan asked, what is the most that that could tell you? Genomic sequence data do not take into account epigenetics, environmental influences, or numerous other inputs. It is very easy for consumers to buy into the idea that if it is in their DNA then it must be real, but there is a big gap remaining between information and meaning. We can bridge the cost gap, Kaplan said, but what are the possibilities for bridging the information gap?

One unusual characteristic of genetics and genomics, Marc Williams said, is that there is no currently available standard for representation of genetic or genomic information in any of the available electronic health records or personal health records. Information is stored as text entries. To be actionable, these entries will need to be computable—for example, to be able to be entered into decision support algorithms. This is an infrastructure gap that is not currently being addressed, and it may be unique to genetics and genomics. Another challenge, Vanier said, is having a system where that baseline information can be acquired cost effectively, entered, and then continually made use of as understanding evolves.

Witt noted that many of the SNPs in use today for drawing conclusions about risk may not give accurate assessments. Perhaps if another thousand SNPs were added and assessed, there would be a different risk finding. Witt described a recent study in which specimens from five individuals were sent to two genomic profiling labs. While there was great overlap in terms of analytic validity (i.e., similar results in terms of SNPs), the risk assessments differed about two-thirds of the time, even as to whether the risk was decreased or increased. This is evidence, Witt said, that genomic profiling is not ready for public consumption.

Vanier agreed that there have been many examples showing that the same sample tested at the three major companies may receive different risk factors. While the analytic validity is good across all three companies

because everyone is using a CLIA lab, the different companies are using different marker selections, he said, which results in different risk scores. This reflects different philosophies among the companies in how they pick the specific markers. The Centers for Disease Control and Prevention estimates that only 7 percent of the markers identified are ready for use in screening. Whether through regulation or self-policing, the industry as a whole needs to do a better job defining a more conservative base of markers to be used. Unfortunately, whether in genomics or existing clinical care, whether in an academic or community setting, there will always be a breadth of standard of care. Thus it is very important, in the early years of a field like this, for patients to approach the task in partnership with the medical community.

As the author of the hypothetical case scenario, Blumberg speculated further that maybe seeing her increased risk for coronary artery disease in her genomic profile was the stimulus that finally prompted Anne to lose 15 pounds and that in the end, after her visit with her doctor, she has no regrets and would do it over again, finding value in the information obtained. So, considering Anne’s story as that of one individual, and not in terms of a broad screening program, Blumberg asked if denying her that opportunity or steering her away from it during a consultation prior to testing would have been paternalistic. Witt repeated his earlier comment that if Anne’s small increase in breast cancer risk leads her to do more self-exams on a regular basis or to go see a provider, then, to that extent, screening has had a positive effect. But looking beyond the individual patient, as a provider in a system that provides care to over 3 million people, he said that the public health perspective must be considered also. In any individual case, it could be argued that there could be some benefit. But looking at the big picture, he said, the benefits are not there. Witt speculated that most people will not take any action based on the results and said the costs are still prohibitive. He also questioned the quality of the counseling in some of the companies.

Further discussing behavioral change, Woolf said that obesity and smoking are common not because people lack motivation or because they do not know that being overweight or smoking is unhealthy. Any sophisticated understanding of why these unhealthy behaviors are so prevalent has to take into account various environmental and contextual factors. Thus, although motivation gained from a genetic test may be of some benefit, the real opportunities for changing health behaviors are associated with the natural, built, and social environments (e.g., living conditions, dietary habits, availability of safe areas to walk, and advertising). Vanier agreed, adding that genetics tests should be bundled in a suite of services that

include messages advocating a healthy lifestyle, health coaching, and other follow-on activities. The teachable moment is when the individual first sees his or her results.

Marc Williams said that geneticists have historically had very little to offer in the way of treatment, and yet they have still been advocates for testing in many situations. In a study of Huntington’s disease, for example, a selected cohort of individuals who had been screened had an overall self-reported anxiety score and self-perceived sense of health that were much closer to normal than those individuals who had chosen not to be tested, irrespective of whether the screening showed that the subject carried the mutation. In the insurance world, the single greatest predictor for healthcare expenditures is self-perceived health. One could argue that if, in fact, this testing led individuals to be less anxious and to have a better self-perceived health, that might then reduce expenditures. Is it possible to quantify personal utility? Witt noted that predictive testing for Huntington’s disease is different from many other genetic tests in that the results are not going to lead to any medical intervention or health test. The value lies simply with providing information for the person and, even in the absence of useful treatment, having this information can be of great value to people. There is a significant difference between predictive testing for Huntington’s disease and broad genomic profiling. Vanier noted that in the REVEAL study (Risk Evaluation and Education for Alzheimer’s Disease) the most interesting use of the individual risk information derived from the test was for use in financial planning.

One participant said that genomic profiling should not be prohibited, but rather should be used to increase medical knowledge and health awareness through patient education and continuing medical education for professionals. The participant also pointed to the sophistication of social network groups of patients who do understand genomics and who can manage their chronic diseases better because they seek out the best treatment and can understand scientific evidence. Witt agreed that there needs to be education but expressed concern that consumers of genomic testing are being thrust into this situation and their doctors may not be up to speed. The participant responded that the marketplace for genomic profiling is still small and that the medical profession has an opportunity to catch up and get ahead. Genomics should be in more medical school curriculums and addressed in residencies as well. The participant urged the medical profession not to underestimate patients’ ability to understand genomics. Vanier added that the Navigenics genetic counseling team spends about 45 percent of their time with the

physicians the company works with. Often a physician will call the genetic counselor to make sure he or she understands all the issues before seeing the patient face to face and providing the genetic results. Physicians usually do that for the first 10 to 15 tests they order, he said. There are clearly lessons to be learned, Marc Williams added, about how to communicate risk information more effectively to all stakeholders.

Marc Williams pointed out that while genomic profiling is currently paid out-of-pocket by consumers, there could be costs and consequences associated with the medical interventions that would take place as a result of the findings. He referred to the example of Jeffrey Gulcher of DeCODE who had his profile done and identified an increased risk for prostate cancer. A subsequent biopsy confirmed that Gulcher did indeed have prostate cancer, and he began treatments. From a personal utility perspective, this is an outstanding result, Williams said. But if every individual who received a similar risk result followed suit and went for biopsy, the consequences in terms of cost to health systems and payers—and potentially the patient—would be significant.

Roundtable chair Burke observed that there was general agreement among the panel that there is insufficient evidence at this point to claim clinical benefit from personal genomics and that there may be reason to be concerned about potential harms, such as cascade effects, false positive results, false reassurance, or about relatively trivial risks taking up practitioners’ time. As research is done to understand the effects of genomic profiling, she asked, shouldn’t there also be an effort to understand how to avoid harm? And how would such studies be designed? Absent the evidence of benefit, should practitioners be recommending against screening—not just a neutral position, but actually a negative recommendation?

Woolf responded that an empirical basis would be needed for any negative recommendation as well. There must be reasonably good evidence of harms to make the case that the harms outweigh the benefits. He agreed that studies need to be designed to be inclusive in terms of outcome measures that look at both benefits and harms (e.g., motivation to adopt healthier behaviors or an adverse chain of events).

A participant pointed out that FDA, CLIA, the courts, and the court of public approval will all factor into this new era of disease-gene association and that there will be liability issues to address. He offered the innovative biotechnology industry as an analogy. A biotechnology company is supported by investors and dependent on the public buying its product. If the company gets it wrong (e.g., produces a faulty product that causes harms or does not work), its board will vote to close the company down because it failed and is not making money. Checks and balances also occur out in the marketplace where consumers vote with their dollars. The participant said that it is fine to be ahead of the field, whether in biotechnology or in genomic testing, but one should keep in mind that he will be held accountable by investors and consumers, by regulators, and by the courts.

A participant suggested that the era of SNPs is dead since they provide such a small amount of information and that we are quickly moving to full genomic sequencing. The participant asked whether it is even worthwhile to continue discussing how to use SNPs. Marc Williams responded that the fundamental issues are more concerned with managing the information than with any specific technology. Making things clinically relevant is not dependent on whatever the technology currently being used is but rather on the level of confidence there is in the technology’s predictive value and if that information is actionable.

This is one of the challenges, Blumberg added, when traditional methods of evidence generation are so much slower than the advance of technology and technological methodology. Perhaps by the time that the current warfarin clinical studies produce results in 2012, another gene may have been identified that, when added to the protocol, makes the difference between clinical utility and nonclinical utility. The old method of one test at a time, one protocol at a time, one disease at a time, is at odds with the profusion of new data and technology that is now available.

Vanier said that this is an evolving process. The algorithms and infrastructure it takes to translate SNP information into an end use are not unlike those that would be used for targeted sequencing and which could then be adapted to whole genome sequencing. The obvious difference is the volume of data that will be put through that infrastructure. SNPs are a first step toward mass utility whole genome sequencing, he said.