Opportunities and Challenges in the Clinical Research Enterprise
Allan Korn, M.D.
Senior Vice President and Chief Medical Officer
Blue Cross Blue Shield Association
All the cowboys and all the cows go in one direction. Why do they do that? It is because when they get to Abilene, they get paid. There is a lesson there. The Clinical Research Enterprise faces the challenge of driving the current system toward high quality care. The Enterprise is made up of diverse stakeholders with vested yet often conflicting interests. To make the necessary improvements, co-operation must be cast as tangible benefits for each of these stakeholders. This poses another obstacle however, as conflict of interest may cause doubt about the integrity of those who stand to gain too much from a particular outcome. Studies are done by groups who have an incentive to ask a certain question. It is recognized that a particular member may not ask the “right” question from the view of the overall enterprise. As a group, we need to figure out how to incent the individual members of the Clinical Research Enterprise to ask the questions that need to be answered and move beyond the second translational block from clinical knowledge to practice. Moreover, how do you create incentives in such a way that they do not exacerbate conflicts of interest, and/or mitigate those that
already exist. The cowboys and cows go to Abilene because they have an incentive to do so. We need to create a similar pull in clinical research today.
Next we are going to hear from Lou Sherwood from Merck and Co. about disease management and the importance of outcomes research. Myrl Weinberg will also discuss the idea of integrated patient-centered care. There is a changing relationship between patients and their healthcare providers that necessitates a more integral role for patients at every step of the decision making process. This is just one of many things that must be considered when rethinking the process by which health care decisions are made and implemented so as to improve the health of the country.
OUTCOMES RESEARCH AND DISEASE MANAGEMENT
Lou Sherwood, M.D.
Senior Vice President for Medical and Scientific Affairs
Merck and Company
The second frontier is getting physicians to practice evidence-based medicine (setting care objectives, collecting data, and being accountable for the results). We need a great deal of additional research to figure out how to do all of this optimally. The thousands of Continuing Medical Education lectures delivered every year in academic institutions and by the pharmaceutical industry do not accomplish the goal of reaching the second frontier. That goal involves starting with outcomes research and moving to disease management. Outcomes research examines the consequences of medical treatment that are noticeable to patients and their families. It includes typical dichotomous variables, such as whether people live or die, and the so-called humanistic outcomes such as quality of life, functional status, and patient satisfaction—things that are vitally important to patients and their families. The third variable is the associated costs.
The thousands of Continuing Medical Education lectures delivered every year in academic institutions and by the pharmaceutical industry do not accomplish the goal of reaching the second frontier.
Why do we want to measure outcomes? It is important to do so because people are beginning to look critically and measure what various products and their organizations do in relation to health care. In light of the consolidation of health care and the changes being made, it is critically important that these measures be examined. We must have a structure, look at process, measure outcomes, and try to continually improve what we do.
What is meant by structure, process, and outcome? If patients’ risk factors are not being addressed in a secondary prevention mode (the things that we
know work), we are not delivering quality care, and patients are not receiving the benefits of the research conducted over several decades. Movement from outcomes research to outcomes management or disease management occurs when one seeks to produce desirable outcomes in usual clinical settings. It involves marrying the practice of medicine with the principles of public health.
We need to move the paradigm ahead. In an informatics era it is unacceptable for physicians not to be tuned in to the advances in medicine and not to be implementing them in their practices.
Physicians will continue to deliver care to one patient at a time, but they must start collecting data across their populations of patients. It would be useful if the average pediatrician had a data base on all immunizations of the children in his or her practice. It would also be useful if the primary care physician who, on average, has 1,200 or 1,300 post-menopausal women in his or her practice had a database on their pap smear results and mammograms, as well as assessment of cardiovascular, osteoporosis, and mental health risks. We need to move the paradigm ahead because in an information era it is unacceptable and borders on the unethical for physicians not to be tuned in to the advances in medicine and not to be implementing them in their practices. Major changes in information systems and infrastructure are needed to help physicians achieve those goals.
Disease management is a process that assists payers and providers in improving clinical outcomes and quality of life, and in managing health care costs using the principles of quality management. What is needed is an infrastructure and a common set of outcome measures that are endorsed by providers. If guidelines are “shrink-wrapped on someone’s shelf,” they do no good. What is essential is that at the local level, in the medical group, practitioners must adopt a guideline—either their own or someone else’s—with an eye toward process. This process entails setting objectives, collecting data, and looking at what happens to people.
The problem is that we do not know how to perform this form of applied clinical research well. A major barrier is not just the lack of a universal electronic medical record, or the absence of an infrastructure, but the mindset of physicians. Physicians have had limited training and orientation to this way of thinking. Not only must we teach the students, but we also have to train the faculty. We train physicians to be independent thinkers, and that is commendable. But if they cannot block and tackle, who would want them on their football team? Imagine eleven independent thinkers running around the field! Teamwork is essential, and it is something we learn about in industry.
In medical practice, there is heterogeneity well beyond what the data support, and we know that the variation is related to different outcomes. The critical question that we have considered in today’s workshop is, how do we institute systems of care
that improve outcomes for a population? This is not to say that every patient should be treated in exactly the same way; yet certain common themes emerge that ensure optimal care if data are collected across all patients for common diseases. Again, modern information systems are critical for achieving this aim.
We must start thinking about managing the total cost of health care, not just individual cost centers. We tend to work in a “silo mentality,” where each segment of the organization focuses on its own budget, when these budgets are actually closely linked. The Department of Veterans Affairs, for example, in setting up an effective system to connect various segments of the department, has been a pioneer this area.
We have published studies containing findings taken from the records of about 50,000 patients with coronary heart diseases; these records were drawn from the practices of several hundred cardiologists.1 The findings were appalling, even in terms of accomplishing cholesterol screening, much less in achieving the outlined goals. In a follow-up study, 10 groups of cardiologists met to define the objectives for patients with coronary heart disease. They did have a common set of objectives, however, and they had to collect a common data set. The only common variable among the groups that did the best work was the presence of a nurse or nurse practitioner who made sure that the work was done, that the patients received the right care, and that patients’ risk factors were addressed.2 The study pointed to the need for stronger research efforts. Physicians have to drive the quality of care, but they also have to control the costs.
The new paradigm is a dynamic relationship in which information is shared and the physician, appropriately informed, can present choices to patients, who can then make informed decisions.
The old paradigm of a condescending flow of information from physician to patient no longer holds. The new paradigm is a dynamic relationship in which information is shared and the physician, appropriately informed, can present choices to patients and families, who can then make informed decisions. The key is to maximize provider resources and expertise, with the help of systems, but maintain a true partnership between physician and patient.
How do you change behavior? By setting objectives and goals and using
quality measures, by examining one’s own data, benchmarking, and systems support. Carrots are nice, but sometimes sticks are necessary. Our traditional educational programs, whether medical school curricula or other areas, are grossly inadequate. We must experiment in this arena, because we are remarkably short on knowledge of how to change behavior.
Merck recently sponsored a major observational study called NORA the one year follow-up study of a cohort of 200,000 women.3 The background for this study was the magnitude of the burden of osteoporotic fractures and the fact that bone density measurements have been shown to predict fracture risks, along with other risk factors. The study objectives were to report the occurrence of low bone mass in a large cohort of ethnically diverse post-menopausal women, examine the relationships between bone density and fractures, and relate other risk factors to fracture risk.
From September 1997 to March 1999, 200,000 women were randomly selected at 4,100 primary care offices. At baseline these women had a mean age of about 65 years. In this population, nearly half had low bone mass. Seven percent had frank osteoporosis according to the World Health Organization criteria. Forty percent had so-called osteopenia, i.e., their bone density reading was –1 to –2.5 (1 to 2.5 standard deviations below the mean). Sixty-three percent of the women had taken estrogen.
Those women who were currently taking estrogen had a higher bone density; those who had been taking it for 10 years or more had the highest bone density. Women who had taken estrogen for even 10 years or more had lost most or all of the bone density they had previously gained if they had discontinued taking it for 5 years or more. Within the one-year follow-up period, 1.5% of the women sustained fractures. When we looked at the bone density by cohorts, we found that the women with the lowest bone density had the highest likelihood of having fractures. Using a baseline questionnaire and an inexpensive bone density test, we were able to predict quite accurately the women in this cohort who would have fractures. Regardless of what instrument we used, the findings were essentially the same. The cohort continues to be followed and the two year data will be available soon. This is the largest cohort of post-menopausal women being followed in an observational study.
We have been experimenting with guidelines and pathways, but we really need to move to document outcomes and have physicians and others make measurements and be accountable. The real question is, how long will it take? We need providers, payers, and patients to increase their sophistica
tion so that they understand this new way to think about and practice medicine. We need to develop methods to resolve long-standing issues that have presented barriers to evidenced-based practice. These methods include changing medical school curricula, building systems and infrastructure, and orienting people in this direction.
INTEGRATED PATIENT-CENTERED CARE
Focusing the discussion on integrated approaches to patient care, Myrl Weinberg of the National Health Council described an initiative called integrated patient-centered care, which goes beyond disease management. This initiative takes into account the patient with multiple chronic diseases and looks at the whole person. It is particularly applicable to the older population. Information is shared across providers, including information that providers rarely receive about complementary and alternative treatments that people are self-managing. The question is, how does this effort fit into what has just been discussed?
The National Council on Quality Assurance is currently finalizing standards for the accreditation of disease management programs. Just as we begin to consider a more holistic, integrated, patient-centered approach to care, disease management companies are contracting with health plans to handle one disease at a time. The progression has been from viewing the patient as a body organ to viewing the patient as a disease. The Clinical Research Enterprise needs to ask the question, what kind of clinical research will guide the decisions as to what is put into practice and how practice is carried out? Otherwise, we will be performing more segmented, targeted clinical research, rather than anticipating a system that considers the whole patient.
Lou Sherwood of Merck suggested that having all the information about a patient in one place is the “holy grail” of evidence-based medicine. He noted that this approach is essential in an aging population and that it represents a new process in the delivery of care
Myron Genel of Yale University agreed that disease management is the product of clinical research and that the evaluative sciences will help guide proper decisions for improving disease management. He mentioned that efforts to change behavior are enhanced when the relationship between physician and patient is stable.
RESEARCH PRIORITIES IN PHARMACEUTICAL COMPANIES
Participants in the workshop turned to the issue of research funding, noting the many difficulties to be overcome. Sean Tunis of the Centers for Medicare and Medicaid Services began by suggesting that the osteoporosis study described earlier by Lou Sherwood of Merck and Company highlights a fundamental problem in research—conflict of interest. Researchers at Merck may not have de
signed the study and decided to fund it by considering what would be the most important clinical question to ask in osteoporosis prevention. Rather, Sean Tunis suggested that a study on the prevalence of osteoporosis in a huge cohort of women may have been appealing because it potentially serves the needs of a company that produces an important drug for the treatment of osteoporosis. He postulated that an important clinical question about osteoporosis prevention in women at low risk would be whether taking calcium supplements twice a day had as much effect on bone density loss as taking Fosamax.
Lou Sherwood replied that Merck did conduct a three-armed study that examined the effects of calcium supplementation. Results showed no additive effect of the calcium over and above the alendronate and little or no benefit from calcium alone. This has been well documented in the osteoporosis literature.
Sean Tunis responded that the essential issue is that the priority-setting mechanism for Merck may not be what payers want. He observed that the large osteoporosis study was supported by those who produce scans to detect low-density bone mass and by those who sell a drug to treat osteoporosis. Because we are faced with trying to figure out how to direct limited financial resources to the most important research questions, there are reasons to think that the constructive interaction of self-interest noted in the case of the osteoporosis study may not be the way that the Clinical Research Enterprise produces the most valuable knowledge for the money.
Lou Sherwood replied that Merck’s incentives for the study are aligned with those of the National Osteoporosis Foundation in terms of helping to identify women who were at risk for this disease. He noted the importance of making physicians and patients aware of the advances that have taken place in osteoporosis. A main incentives for conducting the study was to help educate physicians as well as post-menopausal women. He acknowledged that the pharmaceutical company has profit incentives, but not in a way that does not enhance health. The key question is, what can we do to improve health? If a woman is diagnosed with osteoporosis, and there is a well-studied medication that is appropriate for that person, we might expect it to be used. It is a question of where the incentives are. Those who are paying for osteoporosis care may not want to deal with this issue. Fortunately, the Bone Mass Measurement Act was passed, and women covered by Medicare can now be reimbursed for bone density measurement.
Patricia Salber of General Motors Corporation noted that the osteoporosis study revealed that the number of women with clinically significant conditions related to their osteoporosis was relatively small. She raised a question as to whether all those women need Fosamax simply because, according to some guideline, their bone mineral density is considered low. Do we want to ask the question, who amongst all those people actually needs to have the treatment? The key question then becomes, who would be interested in funding a study to examine which patients have the condition but do not need the drug?
Rick Martinez of Johnson & Johnson entered the discussion by speculating
that payers are experiencing growing pains, transitioning from consumers of research to potential sponsors of research. He also stressed the importance of differentiating among the kinds of research under consideration. The motivations for supporting basic research and clinical research are very different. Certainly, the motivations behind pharmaceutical-sponsored research are different from those driving government-sponsored research through NIH, which is more curiosity driven. Industry is product-driven. The question of how payers inform their unique decisions will remain unsettled until the commercial insurers and investor-owned HMO’s conduct studies for their products.
William Crowley of Massachusetts General Hospital noted that the first translational level of research is largely and quantitatively funded by two groups in this country, NIH and pharmaceutical companies. Agencies such as the Veterans Administration, the Department of Education, the Department of Defense, and the Environmental Protection Agency, as well as some professional societies, do some research but in a limited role. Clinical trials are quantitatively driven by the pharmaceutical industry first, the biotechnical industry second, and now the NIH. He asked, what happens after that, at the second translational block? Although some infrastructure is in place for the first and the second groups, essentially no infrastructure is in place for the third. The cost to put it in place would be enormous, considering the information technology needed, the methodology required, the variety of community-based organizations affected, the large number of patients involved, and the breadth of disease states. The relative cost of such an effort needs to be kept in perspective.
Consider for a moment that $1.5 trillion is spent on health care in this country, and let us say that 1% is the amount to be devoted to such an effort. That is $15 billion, compared, for example, with a total annual budget for the NIH of $24 billion. We have to tie the spectrum and the quantitation together because we cannot discuss one without the other.
A paradigm shift in all types of research is that the individual, investigator-driven research is giving way, at all levels of research, to research conducted by large coalitions of multidisciplinary groups. Even the genomic research is performed by a multidisciplinary group. Those in industry have learned how to handle, encourage, and reward team research. In contrast, payers, providers, academic centers, and medical schools have almost no cultural background in using this approach. The paradigm is individual, investigator-initiated research, which is not “selfish research,” but rather inquiry-driven research, without which much collateral research would not be performed. The demand from the public, the demand from the genome project, the demand from everyone at this workshop is large, multidisciplinary groups. That is what the members of the Clinical Research Roundtable are so excited
Happy are the circumstances when the public health and the private wealth converge.
about. It is the coalitions that receive funding, not individual groups, because there is such a major paradigm shift afoot.
Rick Martinez noted that more than 100,000 physicians are currently in various residency programs. Most of those physicians will finish without any exposure to research methodology and with no experience working with biostatisticians or with those who design clinical trials. The situation is unfortunate, especially given the promise of genomics. The problem, however, is not unique to medical schools; even undergraduate education silos its departments of psychology, statistics, and mathematics to operate in relative isolation from each other.
Hugh Tilson of the University of North Carolina emphasized that the ethics and politics of who pays for research should be revised often, remarking: “Happy are the circumstances when the public health and the private wealth converge.” So much work needs to be done that if industry receives some gain from the research and is willing to pay for it, the effort is commendable. Hugh Tilson commended Lou Sherwood on the published article on the osteoporosis study, not just because it is a good article, but because Merck scientists are listed as authors.
As a scientist who worked in industry for 15 years, Hugh Tilson commented that he was never amused when it was suggested that he should not be listed as author of an article when he had done a great deal of the scientific work. The suggestion was made, he said, because of the perception that if he worked in industry he could not be trusted and, consequently, the content of the article could not be trusted. Conflict of interest is a difficult issue, and one that we need to keep addressing squarely at this roundtable.
The training of epidemiologists and outcomes researchers is inadequate even for industry research, much less for the research urgently demanded by the roundtable. Nor is training adequate for accomplishing the large public health practice and systems research agenda.
Another challenging issue is how to use the $16 billion to do what is needed on the right side of the research agenda, i.e., on the public health side. If there is funding for this effort, researchers may be drawn to the field. On the other hand, if no researchers are trained to do the work, no amount of money will bring results, at least not in the short run. The training of epidemiologists and outcomes researchers is inadequate even for industry research, much less for the research urgently demanded by the roundtable. Nor is training adequate for accomplishing the large public health practice and systems research agenda.
It is shocking that in America today no centers are funded to do public health systems research. We must close that gap if we are to have any understanding of the circumstances that cause us to be healthy, or at least the circum
stances that help us forestall the otherwise inevitable, inexorable, and sometimes rapid progression from wellness to illness. We need to stop that progression at the population and environmental level, and this effort will require an enormous investment in an infrastructure that must be created from scratch.
Hugh Tilson noted that the Centers for Education and Research in Therapeutics (CERTS) are another greatly under-funded area, and he mentioned that five years ago, no centers existed for education and research into the translation of what is known about treatment into improved practices of therapeutics. Such centers might address questions that do not necessarily have the proprietary interest that would draw industry sponsorship.
Recognizing the gap in funding, Congress appropriated funds for the CERTS as part of the FDA Modernization Act. Those seven centers are now funded through AHRQ collaboratively with the FDA. Applications to AHRQ for therapeutics research must include a proposal for a program of research that improves health specifically by improving the translation of what is known into what is done therapeutically. These centers provide the opportunity to fund the capacity for the research and not just the research projects themselves. In another stroke of genius, AHRQ funded centers conditionally on something that academia has always had a difficult time doing. Does anyone want to guess what? They have to work together! Representatives from the seven centers are required to come to the table quarterly as collaborators to examine the development of cross-academic synergy. This mechanism pulls the group together in a coordinated activity chaired by a national coordinating center, directed by Robert Califf of Duke University, and a national steering committee, chaired by Hugh Tilson.
Clinical researchers must be trained not only in clinical research methods but also in the dissemination of the information.
The centers were instructed to develop public-private partnerships as part of their long-term sustained viability. They were charged with creating the context in which the long-standing distrust and adversity between the private sector—the pharmaceutical industry, particularly in therapeutics—and academia and practice could be overcome. The purpose was to create a context for industry and academia to come as partners and develop research projects jointly, and to propose these projects to industry and voluntary partners for co-funding and collaborative work (including industry researchers as authors on research articles). Federal funding was to be reserved for projects for which private funding was not forthcoming. We are working hard on continued evolution and application of a set of principles for public-private partnership to ensure that all of the interests at the table can be heard, respected, and advanced.
Bill Sigmund of Pfizer stated that sometimes the misperception of industry-based research is that it is tainted and of low quality, but much effort is made to
ensure quality. Clinical researchers must be trained not only in clinical research methods but also in the dissemination of the information. Patients vary in their ability to make sound health care decisions, and we need to work on developing the best ways to communicate and disseminate the information to all of them.
Concluding the session, Allan Korn posited that demand-driven medicine is effective, as demonstrated by the efforts of the American Cancer Society. On the other hand, we face an issue when, for example, patients demand bone marrow transplants that are probably more harmful than helpful; demand spiral CT scans if they have been tobacco smokers, long before the NIH reports whether or not results from these scans are meaningful; or demand spiral CT scans of the coronary arteries as screening tools rather than diagnostic aids. We want to demand the right things. One reason we exist as a roundtable is the continued pressure created by unrestrained expectations in an environment of limited resources. We have to be very careful when we open the door to the expectation that all demands can be satisfied and allow everyone through it.
The purpose of the session on opportunities and challenges in the Clinical Research Enterprise was to tie together points made in early sessions of the workshop, explore opportunities for new approaches to research and patient care, and examine the challenges inherent in research funding, such as provision of incentives and conflict of interest (see box).
A representative from Merck and Company described outcomes research that can assist with disease management, citing recent research on osteoporosis prevention. A representative from the National Health Council described an initiative—integrated patient-centered care—that extends beyond disease management by taking into account the whole person, who may have multiple chronic diseases. Finally, participants discussed incentives in research funding and the conflict of interest that can result. They acknowledged a trend toward, and a need for, a paradigm shift from investigator-driven research to research conducted by large coalitions of multidisciplinary groups.
Highlights of Session on Opportunities and Challenges in the Clinical Research Enterprise
What are the opportunities for new approaches to research and patient care?
What are the challenges facing the Clinical Research Enterprise?