Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
5 Brainstorming on a Service Delivery Model for the Future Sharon Kardia, Ph.D., Moderator Panelists: Debra Lochner Doyle, M.S., C.G.C.; Alexandra Shields, Ph.D.; Vivian Ota Wang, Ph.D., F.A.C.M.G., C.G.C.; Catherine Wicklund, M.S., C.G.C.; Frederick Chen, M.D., M.P.H.; Catherine DesRoches, Dr.P.H.; Bruce Korf, M.D., Ph.D.; and Sharon Terry The focus of the following discussion was to explore what new and future models of genetic service delivery might look like, including who will provide the services, what those providers need to know, how the providers will inform patients, where the services will be delivered, and what, if any, burden might be placed on providers using these new models. Health Care Delivery System Lochner Doyle began the discussion by saying that there is no one-size- fits-all service delivery system for genetics and genomics. Different models must be tailored and modified over time. Entrepreneurs outside the medical system may help to develop new ways of meeting the needs of the pro- vider and patient markets. Primary care physicians, according to Chen, are undertaking a renovation of the medical home. Elements such as informa- tion management, information technology, and the use of electronic health records are driving major change in primary care practice. Reimbursement for primary care, however, has not kept pace with other areas of medicine. Shields concurred that reimbursement must be addressed if the medical community is to realize the potential of genomics. The full continuum of services that are needed to deliver genomic medicine adequately and appropriately must be reimbursed and available for both commercially and publicly insured populations. If medicine has value, society must ensure that there is a floor beneath which citizens cannot fall; there must be funding for public hospitals, community health centers, and providers that serve unin- sured patients. Primary care and prevention must be adequately reimbursed; 41
42 INNOVATIONS IN SERVICE DELIVERY IN THE AGE OF GENOMICS the promise of genomics in everyday medicine could perhaps be used as a leverage point with which to incentivize the reformation of primary care. Chen raised the question of how genomics will fit into a system that is already overloaded. There are workforce studies that claim a massive short- age of physicians. In addition to looking at physician-to-population ratios, one must also examine the distribution of physicians across the population. Currently there is a maldistribution of providers, and this is likely to present difficulties for the practice of genomics as well. Korf used Figure 5-1 to illustrate the stratification of the world of genetics and its place in medicine. At the top of the triangle are rare mono- genic conditions, and it is in this space that medical geneticists have histori- cally practiced. These disorders, such as inborn errors of metabolism and congenital anomalies, are relatively rare, single-gene conditions. Despite their rare occurrence, an increasing number of diagnoses are possible, and new interventions are being developed for individuals with these kinds of conditions. Newborn screening, in particular, is having a significant effect on the incorporation of genetics into primary care. The middle tier, referred to here as common monogenic, represents FIGURE 5-1 Stratification of genetics in medicine. SOURCE: Korf, 2008. R0146 Figure 5-1 fixed image
A SERVICE DELIVERY MODEL FOR THE FUTURE 43 conditions such as breast and ovarian cancer and hereditary nonpolyposis colon cancer, which are not necessarily monogenic but are fairly common with well-established genetic associations. Individuals with these predis- positions do not look any different from other patients. The only way to identify that they are at risk is to conduct a family history, which is a lengthy and complex process, or to perform genetic testing, which has its own benefits and pitfalls. The base of the triangle, continued Korf, is the area in which data are least sound. Given the lack of data and understanding, medical prediction will likely be about as accurate as weather prediction, that is, it will be correct to some degree, but far from exact. Enormous amounts of data and statistical modeling go into a prediction, and stochastic factors can affect the outcome. As data continue to be developed, it will be critical that part- nerships within the professional genetics community guide the rest of the medical community through the disruptive technologies and new service models that arise. Terry pointed out that although there is much discussion about how providers and other sectors of the medical and research community must be integrated, it is crucial that the consumer community also be involved. There are many examples of consumer demand driving rapid change in other industries, such as computer technology innovation over the last 30 years. Similarly, companies such as Amazon and Flickr have used technol- ogy to make it possible to connect people, and their physical and digital media, all over the world. Consumers are forming communities online, demanding and receiving information, and determining for themselves what is important or useful to them. The genetics and medical communities should be thinking about how they can help to augment these consumer- driven innovations, rather than fearing or delaying inevitable changes. Wicklund added that providers of health care and consumers of health care may not have the same definition of an ideal system of service delivery. While both certainly want quality of care, consumers may wish for more flexibility in their care, including the ability to seek the information they want in the form they want. They may wish to take their health care into their own hands. Wang asked, what if, after exhaustive research on gene associations and geneâenvironment interactions, it turns out that the message is that patients need to eat better, sleep 8 hours, and exercise? Health commu- nication messages arenât always about complex diseases and treatments; the importance of helping people do very basic, yet difficult, things cannot be overlooked. Levin noted that research, data collection, and helping providers use new technologies are all important components of moving genomics into health care. However, it is also critical to have an educated patient base,
44 INNOVATIONS IN SERVICE DELIVERY IN THE AGE OF GENOMICS across socioeconomic groups, that is prepared to receive this information and understand the advantages, disadvantages, and limitations of genetic testing. Health Information Technology DesRoches reported on her recently published study (DesRoches et al., 2008) that attempted to establish a baseline estimate for the number of phy- sicians who are using a fully functional electronic medical record (EMR), one that can integrate family history into a medical record and provide physicians with clinical decision support, e.g., suggestions to order a specific test or offer a specific screening. The baseline number was 4 percent. Even in the largest practices in the nation, only a small minority of physicians use this technology. When asked specifically about clinical decision support functions, 21 percent of physicians reported that they had a computerized system for reminders and guidelines. Only 17 percent of them use it, which translates into about 2 percent of physicians nationwide using this technol- ogy. Among the 4 percent of physicians who have a system that prompts them to order a genetic test when appropriate, only 8 percent have ever ordered a genetic test because of the prompt. DesRoches noted that both Google and Microsoft have offered web- based tools that consumers can use to store their health information. This could potentially save time and energy by allowing providers to simply upload a patientâs information into their system to populate a medical record. Currently, two hospitals in the country have this capacity. It is often assumed that genomics can be incorporated into a physicianâs practice through the use of these types of health information technology. The current low utilization of available technology, however, suggests that it may be a very long time before genomics is fully integrated into everyday practice. Shields added that when health information technology is put in place, the most critical element needed is a method for tracking the use of genomic medicine and collecting adequate clinical detail to assess its effect on health, both in real time and over time. EMRs can and must be made more clini- cally rich to be useful in genomic medicine, structured so that they can be used for research, medical practice, and public health. Shields asked the audience to imagine an EMR that includes information such as race, e  thnicity, education level, housing situation, diet, levels of stress, and aller- gies that could both be informative for individual risk assessment and also provide data for future research to find patterns among the EMRs of mil- lions of people. One workshop participant expressed her opinion that EMRs and per- sonal medical records are a false messiah. She said that EMRs are pro-
A SERVICE DELIVERY MODEL FOR THE FUTURE 45 grammed for billing, not for patient care, which makes it very difficult to capture the subtleties, nuances, and innovations in medicine that exist in the standard medical record. Similarly, the personal EMR, if it mimics what patients currently bring to the office, is a hodgepodge of previous opinions and test results that are of very little use to the physician. The thing primary care providers and their patients need is more time for contemplation. Unfortunately, under the current billing system, providers are rewarded only for actions that are billable, not for the time it takes to interact with patients and consider the appropriate course of action. From the providerâs point of view, when a patient comes into the office, collecting data for an EMR research project is a very low priority. The top priority is helping the patient with his or her health concerns. Asking a physician to collect information about demographics and environment is a waste of the providerâs time. DesRoches responded that the vast majority of physicians agree that these health information technology tools are not very easy to use and that they take more time than they save. This is because the financial incentives for this technology are completely misaligned. The savings that one could garner from using the system accrue to payers, not providers; therefore, there is no incentive for providers to use the system. As providers are i Ânundated with more and more genomic information and innovation, there will have to be a technological solution that helps them filter the informa- tion and decide on a course of action. This system must be created to benefit the provider, however, or it will not be used in practice. One participant observed that residents-in-training are, like their pre- decessors, extremely intelligent. However, they have lost some skills (e.g., math, spelling, clear handwriting) because new technologies enable them to function without those skills. If a robust decision support system was fully integrated into clinical care, he asked, is it possible that physicians would lose their ability to think and become technicians who simply input information and await the response? DesRoches responded that the unintended consequences of the use of health information technology are unknown at this point, because use is minimal. As it becomes more widespread, researchers will have to keep a close eye on these types of consequences, which are a serious concern. Chen added that although it may be true that students and residents no longer know how to listen to heart sounds, physicians also no longer taste and smell urine to arrive at diagnoses. Disruptive technologies are called disruptive for a reason, and they encounter much well-meaning opposition before they are finally adopted. Companies such as Microsoft and Google are capable of grasping the complexity of health care, which some health information technology vendors tend to underestimate.
46 INNOVATIONS IN SERVICE DELIVERY IN THE AGE OF GENOMICS Data Collection Several participants noted that one emerging theme seemed to be how to collect data from clinical settings, not only from claims data, but from creative new systems that do not burden providers. Korf pointed out that one of the challenges to such a data collection system is developing a systematic method for capturing phenotypic data that can be compared and aggregated on a large scale. Despite the fact that human phenotypes have been studied for millennia, and genotypes only for decades, it is far easier to classify and categorize genotypes. In response, one participant reported on a study that examined phe- notype data (Roses et al., 2005). Experts on 17 diseases contributed their opinions, and the researchers were successful in classifying the phenotypes of diseases within disease categories. However, an enormous amount of money and time was required. The sheer amount of data necessary to clas- sify a patient as having a disease would prohibit large-scale data collection in a clinical setting. Another participant added that even if it is possible to collect excel- lent genotypic and phenotypic data, still missing from the equation are the molecular data gleaned from samples of blood, tissue, tumors, etc. There are systems in place, such as those at the Department of Veterans Affairs (VA), that collect data on millions of patients per day, yet neglect to collect samples for molecular profiles. There is a focus on building a system for archiving clinical information, but molecular data must be included if the system is to be robust. Another participant noted that there is a major disadvantage to collect- ing data through clinical practice, which is that because medical practice is fairly standardized, there would be no standard âcontrolâ group. Rather than answering an experimental question, the data would simply reflect current medical practice. Having all the data may be interesting, but it would not necessarily translate into better clinical care. She added that some of the discussion had focused on how to fit genomics into the current delivery system. It is quite likely, however, that the future system will be vastly different. The cost of sequencing a genome is decreasing rapidly, and at some time in the future everyone may have their genome on âGoogle Genomics.â An individual will be able to look at recent studies of gene associations and use that information in conjunction with his or her health history to make decisions about prevention and care. Perhaps, instead of newborn screening, every baby will leave the hospital with its genome on a CD. While randomized controlled clinical trials will still exist, the way in which data are translated into health will be very different than in our current system.
A SERVICE DELIVERY MODEL FOR THE FUTURE 47 Another participant argued that although evidence thresholds of when to move from research to practice will vary for different genes and condi- tions, the fundamental problem is balancing the need for evidence of clini- cal validity and utility with a rapidly changing world in which consumers are demanding information and industry is continually developing new technologies. Genomic technologies can serve as a wake-up call for the medical system as a whole. The system is already strained under misaligned reimbursements and maldistribution of resources. The rapid growth of consumer demand for genomic technologies may soon break the system completely. The challenge is to collect quality data and establish evidence thresholds that could be used to determine coverage and reimbursement in practice. However, that participant continued, relying on traditional research methods is not feasible for genomic innovations. A large cohort study on a million people for 20 years would result in data in 40 years and a trillion dollars spent. In the meantime, the medical community would have to live with uncertainty and assess each new technology as it is developed. Innova- tions that meet a certain standard of evidence can be moved into practice and studied in a postmarketing environment. Large health care providers such as the VA are capable of performing this sort of research and practice along with informed consent. There must be a technology assessment system in place that is robust enough to allow promising technologies into clinical care and then provide data on the clinical utility of these applications. Another participant asserted that technological innovation in genomics, like every other innovation in medicine, is likely to go charging forward with tremendous waste, major errors, disenchantment, reenchantment, and finally a stable foundation on which to build the future. It will take a long time to translate the knowledge that is gained through research into the wisdom to know what to do with it. Panelists were asked to list their top priority area for service delivery, and then additional audience participation was encouraged. Chen placed priority on developing the clinical utility of genetic tests. DesRoches favored realigning reimbursements for primary care physicians. Wicklund also placed a priority on reimbursement, emphasizing the need for reimburse- ment for services other than just procedures. Korf said that his priority was developing a well-educated group of health care providers. Lochner Doyle emphasized the need to address the lack of data. Wang stated that priority should be given to focusing on people. Shields stated that her priority was to obtain better clinical information to evaluate the impact of genetics. Terry favored creation of a coordinated system with participation from all stakeholders.
48 INNOVATIONS IN SERVICE DELIVERY IN THE AGE OF GENOMICS Discussion Sharon Kardia, Ph.D., Moderator One participant stated that perhaps the people sitting around the table are the wrong people to be brainstorming about innovation. The members of the Roundtable are present because they are experts, she said; they have learned and are invested in the current system. True innovation may have to come from a table of outsiders with disruptive minds who can grapple with these issues with unbounded thinking. There are many innovative models in which âoutsiderâ genetic service providers are taking risks and struggling with the issues that have been discussed here. Perhaps the âinsiderâ com- munity should collaborate with these other providers to move innovation along. Another participant responded that although the idea of collaborating with innovators sounds appealing, there is conflict when innovators are also trying to turn a profit. Innovation can be done creatively, but its evaluation should follow the rules of research and take place in a relatively controlled environment. Innovation without research and assessment could result in more harm than good on a population level. Collaboration between the pri- vate sector and the government in setting parameters for innovation would help to ensure that new technologies are beneficial rather than harmful. A participant reported that at a recent American Medical Association meeting, there was discussion about the idea of a government institute, sep- arate from the National Institutes of Health, that would study the delivery of health care. Examples of what might be achieved by such an institute can be found by examining health care improvements that were made because of battlefield observations. Statisticians who were working on military battlefield health identified and solved problems with innovative health care, rather than medical solutions. Eye injuries were a major problem. The statisticians interviewed soldiers and discovered that they were not wearing their goggles because they looked ugly. Eye injuries decreased significantly when new, more attractive goggles were designed. Catheter infections were reduced to nearly zero with the implementation of a checklist for health care providers. The checklist was incorporated into care in the State of Michigan, with similar results (although investigators were castigated for breaking all the institutional review board [IRB] rules). Rather than attempt to change reimbursement policies, the participant continued, perhaps a âNational Institute of Health Care Deliveryâ could tackle many of the problems in the system today, with physicians, ethicists, statisticians, and others taking a systems approach to health care. Another participant pointed out that three government agencies are charged with
A SERVICE DELIVERY MODEL FOR THE FUTURE 49 studying and improving health care: the Health Resources and Services Administration (HRSA), the Agency for Healthcare Research and Quality (AHRQ), and the Centers for Disease Control and Prevention (CDC). The agencies have not, however, had great success in meeting that charge. Another participant observed that much of the discussion has focused on money, that is, on reimbursement policy, laws about Medicaid and Medicare, and money for research. He suggested that universal health care is a way to save money, keep people moving out of the intensive care unit and toward preventive care, and increase the amount of time and energy that could be focused on making health care better. Additionally, primary care physicians will be entering the field of genomics whether they want to or not. Basic genomics education of these providers is crucial, and when providers begin to use the new technologies it is essential that they be given information that will help them understand the genetic test and the results. One participant said that newborn screening is by far the most widely used example of genetic testing. Nearly every baby is screened for 30 differ- ent diseases, regardless of the parentsâ ability to pay. Although this testing is different from the genomic tests that have been discussed at this workshop, there are lessons to be learned. For example, when a baby is diagnosed with a rare metabolic disease, the doctor is unlikely to know much about the condition. One-page sheets have been developed to accompany the lab report and give the doctor a quick overview of what needs to be known and done. Shields said it had been mentioned that genomic medicine could pos- sibly be used as a leverage point to address larger problems within the medical system as a whole. Similarly, perhaps genomics could be leveraged to reduce health disparities. If a health information technology system were implemented to collect clinical data on patients, data about environment and demographics could be collected as well. Very little is known about the experiences and exposures of poor people in this country, and data collection on this scale could help explain some of the health disparities. In addition, disparities in access to and utilization of genomic technologies suggest that there needs to be an effort to communicate and deliver infor- mation in ways that are meaningful to all people. There are opportunities all along the trajectory from research to care to address racial, ethnic, and socioeconomic disparities. One participant said that there remains a huge amount of work ahead in the integration of genomics into health care. The health care system in this country is complex and difficult to navigate, with or without money. Physi- cians must be given well-researched, straightforward tools that improve the time they spend interacting with patients, rather than tools that consume more time than they are worth.
