INTRODUCTION

Myron Genel

In 1992, Edward “Pete” Ahrens, who directed the clinical research center at Rockefeller University for more than two decades, published an insightful monograph, “The Crisis in Clinical Research. Overcoming Institutional Obstacles”(1). Calling for reforms in the National Institutes of Health (NIH) and within academic medical centers which collectively constitute academic medicine, Ahrens' book was read primarily by the clinical research establishment; it received a number of positive reviews and had little impact. In the four or so years since the monograph's publication we have witnessed the collapse of an ambitious effort by the Clinton administration to reshape our health care system and the subsequent advent of market-driven “reform” coupled with a new political imperative to reduce federal expenditures. As a consequence, the intricate system of financing academic health centers, the principal site for education and biomedical research, may be unraveling as academic medicine is caught in a squeeze between limitations of government support and the demands of a market-driven health system to eliminate all but the most essential services required for the delivery of patient care.

Is there a crisis in clinical research? Or is this merely self-serving hyperbole from the clinical research community? Certainly the issues are far more complex than simply a limitation of financial resources. Former NIH director James Wyngaarden(2) identified clinical investigators as an “endangered species” in the late 1970s, and similar concerns have been expressed repeatedly over the years in virtually every clinical discipline, including pediatrics(37). A declining cadre of physician-investigators who are truly equipped to pursue research opportunities, a dearth of adequate training opportunities, even for those who are so inclined, and a paucity of interdisciplinary collaborative models have been identified repeatedly as significant barriers to the conduct of clinical research. The increasing competition for scarce resources, however, appears to have intensified the debate. After a 3-y study, the Institute of Medicine has published its major report on clinical research career opportunities and has provided a number of recommendations for the federal government, in particular NIH, academia, and the private sector(8). One recent editorial speaks of clinical investigation as “an endangered science”(9). In recent months a special ad hoc committee appointed by the NIH Division of Research Grants has identified a variety of concerns with the peer review system for clinical research(10). NIH director Harold Varmus has appointed a special emissary for clinical research-the former director of the National Institute of Arthritis and Musculoskeletal and Skin Diseases-as well as an advisory panel on clinical research, headed by David Nathan, a member of these societies. The American Medical Association, through its Council on Scientific Affairs, has identified clinical research as one of its major issues for 1996. If this is not evidence of a crisis, clinical research is certainly receiving a substantial amount of attention.

The plenary symposium, summarized here, is from the 1995 annual meeting of the pediatric academic societies. Although there was a particular emphasis on the recruitment and training of clinical investigators in pediatrics, the symposium also dealt with bioethical issues in conducting clinical research in children and the effect of new health care organizational models on academic pediatric departments and on support for population-based research. The participants were selected for their particular expertise and asked to comment upon various aspects of clinical research form their own unique perspective and in the context of a rapidly changing health care environment.

William Kelley is executive vice president and chief executive officer of the University of Pennsylvania Medical Center and Health System and dean of the school of medicine. He has been in the leadership of virtually every professional organization devoted to internal medicine and clinical research, having served as president of the American Federation for Clinical Research and of the American Society for Clinical Investigation. Dr. Kelley was chairman of the Institute of Medicine's Committee on Addressing Career Paths for Clinical Research which issued its report in 1994. Russell Chesney is chairman of the Department of Pediatrics at the University of Tennessee, Memphis, and serves as secretary-treasurer of the Association of Medical School Pediatric Department Chairmen (AMSPDC). He has been president of the Society for Pediatric Research and the American Society of Pediatric Nephrology and is former chairman of the American Academy of Pediatrics' Council on Pediatric Education. Barbara Starfield is an internationally recognized health services researcher well known for her studies of primary care practice and the relationship to health care outcomes, especially as they relate to child health. She is a past-president of the Ambulatory Pediatric Association, a recipient of its George Armstrong Award as well as a number of other awards for her research in primary care and child health. Harvey Cohen became chairman of pediatrics at Stanford in 1993, just as managed care reached its ascendency in California. A product of Duke's M.D., Ph.D. program, he was previously associate chairman for research and development at Rochester and is a past-president of the Society for Pediatric Research. One of the leading figures within the American bioethics community, Thomas Murray has had a longstanding interest in clinical research, especially with respect to the care of newborns and children. Recently he has been engaged in probing bioethical implications of the revolution in molecular genetics, having been a member of the Working Group on Ethical, Legal and Social Issues (ELSI) of the human genome project from its beginnings. The responses of our distinguished panelists follow.:

CAREERS IN CLINICAL RESEARCH: OBSTACLES AND OPPORTUNITIES. A POSTSCRIPT

William N. Kelley

In 1991, the Institute of Medicine (IOM) established a committee to review careers in clinical research: obstacles and opportunities. The committee report was published in August of 1994, and its content was reviewed in a brief article appearing in the Journal of the American Medical Association(1). The formal report is available from the National Academy Press (2101 Constitution Ave., NW, Lockbox 285, Washington, DC 20055).

The committee took a broad view of the definition of clinical research, primarily to assure that all of those individuals throughout the United States participating in clinical research would not be excluded by a more narrow definition. The broad definition included all studies intended to produce knowledge valuable to understanding the prevention, diagnosis, prognosis, treatment, or cure of human disease. It also included biomedical and health services research carried out in humans, usually by health care professionals, as well as research in organs, tissues, cells, subcellular elements, proteins, and genes derived from humans. Finally, it was defined to also include the study of microorganisms as well as studies of other members of the animal kingdom when this research is directed toward the study of human disease. However, the committee focused most of its attention on, and believed that the major concerns were related to, those individuals who conduct research which requires “hands-on” participation with a human subject or, as defined by the committee, human research.

The committee members concluded that monumental opportunities exist for the translation of advances in molecular biology to improved human health. These advances have occurred as a result of a major investment of public funds in the support of basic research, and, indeed, the continued development of the field is highly dependent on the continuation of those funds.

Second, the committee believed that the pursuit of careers in clinical research is unattractive. The reasons for this are multiple but include prolonged training, the accumulated debt for the trainees, the lack of role models and mentors, the perceived instability of federal funding, the inadequate emphasis on research in the curriculum, and the competing demands of the practice of medicine that almost all of these individuals encounter. The combination of these factors has led to a chronic undersupply of individuals doing clinical research, particularly those interested in human research.

The committee also concluded that the current level of training and support of health professionals doing human research is now or soon will be inadequate. It was the general consensus that research support was fragmented, undervalued, and underfunded. Because hard data were not available, a small study was commissioned by the committee to examine the frequency of human research funded through the investigator-initiated mechanism at NIH. The most recent grant data available was for 1991. In that year, only 10.4% of funded grants involved human research, whereas 4.5% involved both human and laboratory-based research. A more important variable leading the committee to this conclusion, however, was the burgeoning of opportunities presented for clinical research as a result of the explosion of progress in basic biomedical research. This rapid progress has opened many opportunities for the translation of this fundamental research into human subjects, and hence the demand for health professionals who are qualified to do human research will substantially increase.

It was also concluded that health care reform would likely worsen the current situation. Although the committee's activities occurred during the time of the Clinton administration's direct involvement in attempts to reform health care, the same concerns clearly exist today. That is, the source of funding for support of clinical research and human research, particularly in its earliest stages, is often the academic health center (AHC) itself. As funds become tighter and tighter and these small, but positive, operating margins of AHC disappear, there will be no funds remaining for the start-up of research and for the development of multidisciplinary programs in areas, such as gene therapy, which are absolutely critical in underpinning the entire field of clinical research.

Finally, the committee concluded that there was no segment of society organized to ensure the survival of the clinical investigator.

The committee's recommendations were directed toward the key entities responsible for guiding and supporting clinical investigators in clinical research in the modern era. A number of the recommendations were directed toward the NIH where virtually all of the investigator-initiated funding for clinical research occurs. The committee strongly recommended that NIH develop a method for collecting and analyzing data related to human research. One of the frustrations which the committee experienced related to the serious lack of information over a period of years. In addition, although there was no evidence of bias apparent in the functions of study sections against, or for that matter, in favor of, clinical research, the committee recommended that NIH establish an oversight committee to provide continuous oversight on the function of study sections should such bias ever appear. This is a recommendation which has been made by others in the past. Other recommendations included 1) an effort to expand support for newly independent investigators; 2) expand support for centers and program projects where funding is more likely to relate to human research;3) expand general clinical research centers, where a very large proportion of human research is actually carried out; and 4) expanded support of NIH training programs, particularly the medical scientist training program. Finally, the committee recognized the serious impact of annual discussions in Congress about reducing NIH spending on the academic community. This negative impact is perhaps most serious on trainees and individuals just beginning their research careers. Any success in achieving multiyear stabilization of NIH funding would go a long way toward relieving or at least minimizing this stress.

The committee also directed its recommendations to universities and academic medical centers. There were strong opinions expressed that faculty needed more protection from clinical responsibilities, that they needed more academic recognition and rewards, that the infrastructure to support human research needed to be strengthened, and that efforts to encourage excellent medical students and postdoctoral trainees to pursue careers in clinical research would be substantially enhanced if the medical school curriculum and postgraduate training programs were more inclusive of clinical research training.

Many of the accreditation and certification organizations such as the Liaison Committee on Medical Education (LCME), Accreditation Council for Graduate Medical Education (ACGME) and the certifying boards can do a great deal to encourage the inclusion of clinical research experience in the training programs through medical school and postdoctoral clinical training. In addition, the certifying boards were encouraged to reexamine the potential burden of time limited certification on the productivity of clinical research careers in academic institutions. They were also encouraged to expand their clinical investigator tracks, particularly as they might relate to recertification.

Although not a formal recommendation of the committee, there was strong support expressed for the development of some kind of alliance to support clinical research which might involve the executive branch, Congress, industry, foundations, and professional societies. It was the consensus of the committee that any success in developing a private sector coalition which could work with the executive and legislative branches of the federal government would be extremely useful in the long term for the support of clinical research. In addition, such an alliance might also be effective in enhancing academic-industry relationships and in preserving academic health centers.

The final recommendation-which has been made by many others over the years and, indeed, has been implemented in various specific areas-is the development of a payback mechanism for educational debt. Clearly the mounting educational debt of the average medical student is a serious problem for them as they consider the possibility of an academic career over the long term. The average debt in many schools now exceeds $60,000. The effort of these students to choose a career in which that debt can be repaid can play a very important limiting factor in their choice of an academic career.

Since the publication of the IOM report, a number of things have happened which may well have a positive influence on individuals pursuing careers in clinical research. Perhaps the most important of those has been the decision of NIH director, Dr. Harold Varmus, to take a proactive role in stimulating clinical research, and, to the extent possible, clinical research training. There has been a focus on establishing a training program within the NIH which would be transportable to the General Clinical Research Centers. In addition, a clinical research liaison position has been established at the NIH which is filled by Dr. Lawrence Shulman, previously the Director of the National Institute of Arthritis and Musculoskeletal and Skin Diseases. Finally Dr. Varmus has formed a committee to examine the future of clinical research and how it might be protected for the future.

THE TRAINING PIPELINE FOR ACADEMIC PEDIATRICS

Russell W. Chesney

As a nation, the United States is in the midst of massive restructuring in medicine. Despite admiration for the scientific and technical advances of American medicine, our “national” health indices do not show sufficient value to citizens, business leaders, and state and local governments for investment in health care(1). Current investment expenditures equals 14% of our gross domestic product, or approximately one trillion dollars per year. Market forces are driving health care reform to improve efficiencies in what has been a largely independent and unregulated economic entity. The overriding factor driving this reform is cost containment or, to put it more succinctly, a reduction of costs. These large-scale and yet, at times, local attempts at cost containment place great constraints on the well-being of academic medical centers(2). Among the challenges faced by academic medicine are reformers with little concern for the traditional three-legged stool paradigm of teaching, patient care, and research. Cost containment also places constraints on the traditional sources of support for teaching, research, and patient care. There exists a universally perceived notion that we have overproduced specialist physicians(3). The educational mission that entices many physicians to enter academic medical careers does not appear to be as highly valued by our society(4). Thus the need for “reengineering” or “downsizing” of academic medical centers has led to great faculty distress and trainee confusion.

The challenges of academic departments of pediatrics are clearly just as perplexing as they face flat or falling clinical revenues. Reimbursements for each work unit have fallen. The stable, albeit low, funding for public patients (i.e. Medicaid) appears attractive to private-sector physicians. The competition for extramural research grants has intensified, particularly because large numbers of Ph.D.s in biomedical science have graduated over the past two decades(5). Education subsidies from states also are flat or falling as many states face enormous fiscal income constraints. Moreover, the facilities operated by pediatric departments are expensive and have high, fixed costs.

Additional challenges come from the fact that innovative technologies are perceived by state and local governments and insurers to be driving up the total cost of health care. The pharmaceutical industry is downsizing along with the remainder of corporate America, which clearly influences the availability of funds for drug-trial research. There exists a widespread recognition that the reward system for faculty will become based on productivity and the financial health of the endeavor in which the faculty member engages. Finally, traditionally low income or deficit departments, such as pediatrics, will not be encouraged to continue deficit spending by hospital and medical school administrators(5).

In the midst of these cataclysmic changes which alter our paradigm of academic medicine, recent advances in biomedical research are extraordinary. The unraveling of the human genome, advances in imaging technologies and structural biology, as well as immunology, are only a few of the remarkable areas in which advances have been made. Students of human biology find each copy of their favorite journal replete with exciting new finds. Molecular biology provides the tools that permit investigators to define diseases at a cellular level. The role of cytokines, receptors, and transduction mechanisms in cell processes allow for a clearer understanding of pathophysiology. We are in the “decade of the brain” with developmental neurobiology as its capstone(6). Pediatric research is at an exciting juncture. New and specific therapeutic agents are aimed at receptors, cell signaling mechanisms, and cytokine concentrations. Recombinant hormone therapy is being investigated extensively. There exists an appreciation that drug trials must include children in some fashion(7). Cogent investigations are occurring in the areas of the “new morbidities” in pediatrics: gun control, injury prevention, physical and sexual abuse, and the impact of violence.

Active clinical research has complex workforce issues. The pipeline for future pediatric clinical researchers begins in our residency and fellowship programs, which comprise the pool from which academic generalists and subspecialty pediatricians are recruited. Traditionally, a significant proportion of clinical investigators in pediatrics have emerged from subspecialty training programs. Pediatricians certified by the American Board of Pediatrics (ABP) through December 31, 1994, number 55,130; pediatric subspecialists certified as of this date equal 9,015 or 16.3% of the total(Table 1). This number does not include child neurologists who have completed pediatric training. It also does not differentiate pediatricians in private practice from those who are full-time members of a medical school faculty and who are more likely to engage in clinical research. Most importantly, large numbers of academic generalist pediatricians are usually certified by the ABP but often have extensive further training in such programs as the Robert Wood Johnson General Academic Pediatric Program. It is this group of pediatricians who are likely to conduct clinical investigations of common disorders affecting large numbers of children, or to be involved in “outcomes research” activities. This group, whose numbers are unclear, represents an important subset of pediatric clinical researchers.

Table 1 Certified pediatric specialists as of 1995

The reasons to attract academic pediatricians into clinical research are numerous. Clearly, clinical investigators can use their intellectual talents and focus on questions important to society. The clinical researcher plays a pivotal role in the bedside translation of scientific innovation. For example, the entire development of modern respiratory therapy, including the use of jet ventilators, of extracorporeal membrane oxygenation, and of liquid ventilation as well as surfactant and nitric oxide drug therapies results from translational research. Prospective and retrospective studies of standard therapies can save valuable health care dollars. Moreover, hands-on clinical investigation is a prerequisite to answering questions concerning pathogenesis, the role of social factors, and the role of new therapeutic interventions. The excitement of clinical research involves the acquisition of new knowledge that will have clinical relevance and potential use.

A factor in attracting trainees to a clinical research career is faculty role modeling. Ideally faculty members should encourage residents and fellows to attend local, divisional, departmental, and school research seminars. Several other available activities to stimulate research are the American Academy of Pediatrics' Resident Research Grants Program, of which 43 have been given out over the past 3 y: the Frontiers in Science program at the annual meeting of the Association of Medical School Pediatric Department Chairmen(AMSPDC), to which residents from one third of all medical school-based residency programs are invited each year, and travel funds provided by the regional societies for pediatric research. All residents should be made aware of the Pediatrician Scientists Development Program, a highly effective program which funds six to nine fellows each year for 2-3 y of intense laboratory-based behavioral or epidemiologic research. [The Pediatrician Scientists Development Program is sponsored by AMSPDC and funded by numerous organizations including the National Institute of Child Health and Human Development (NICHD), St. Jude Children's Research Hospital, American Academy of Pediatrics, American Pediatric Society, the March of Dimes Birth Defects Foundation, Toronto Sick Children's Research Foundation and others.] Residents should be encouraged to attend campus-wide or departmental research days. Some national organizations, such as the American Society of Pediatric Nephrology, provide travel grants to PL-2 residents to attend their annual meeting. Finally, in institutions where an active General Clinical Research Center(GCRC) exists, residents should have the opportunity to interact with clinical researchers and understand the rationale behind chemical investigation protocols. Although some of these activities may seem passive, the opportunity to meet and interact with active clinical researchers cannot be overemphasized.

Once a fellow is committed to a clinical research career, the curriculum for his or her training is critically important. In addition to clinical training, the development of a clinical investigator should include formal course work in biostatistics, epidemiology, biomedical ethics, clinical research design, and skills in writing grant applications(8). Other objectives include the acquisitions of skills as an educator, participation in the education of pediatric residents, and the continuing education of practicing pediatricians(8). Board requirements must be met and dealt with in a fashion to permit adequate time for research training. Finally, fellows should be encouraged to learn to focus on an important question in pediatric health care.

After an individual is recruited to a medical school pediatric faculty, several important factors must be overcome to achieve an adequate supply of clinical researchers(2, 4). Postresidency and fellowship training has often been fragmented. A faculty member may need to enroll in the formal courses described above if experience was not obtained earlier. Other issues include a perception that NIH grants are awarded to basic researchers rather than clinical investigators and, in many universities, the lack of interdisciplinary programs that can deal with the multiple overlapping disciplines required for clinical research. The paperwork required to develop a drug or technology trial can be onerous. Young faculty members are also burdened with student debt which may require extracurricular moonlighting activities. Recent AAMC data indicate that 1994 medical school graduates have a median debt of $64,436 which is greater for graduates of private schools (Table 2). Another important factor relates to the reward system within academic medical centers. Promotion and tenure decisions are frequently based on a value system related to publication numbers and the requirement for extramural funding, especially the individual R0-1 grant. Promotion committees are often composed of senior and tenured faculty who may not fully appreciate the effort involved in conducting first-class clinical research.

Table 2 Educational indebtedness of medical school graduates (1994)

Young pediatric faculty members focusing on clinical research need to join regional or national multicenter groups which are formed to conduct clinical research. Such groups include the 12-center National Institute of Child Health and Human Development (NICHD)-funded Neonatal Collaborative Group and NICHD-funded, seven-center Pediatric Pharmacology Research Units. Hematology/Oncology investigators should join either the Pediatric Oncology Group or Children's Cancer Study Group. Other NIH-funded projects include Specialized Centers of Research (SCOR) grants in certain focused areas, pediatric pulmonary centers, and multicenter groups focused around a clinical problem. Certain subspecialties have developed multicenter study groups, including critical care, nephrology (Southwest Pediatric Nephrology group, North American Pediatric Renal Transplant Cooperative Study), and immunology(Pediatric Immunodeficiency Group). Young faculty members should be encouraged to become clinical associate physicians in their GCRC (the CAP Award). With a group effort, more useful information can be obtained and the publication record of many of these groups is quite strong. It is also critical that departmental and school promotion committees understand the importance of such multicenter groups and their value to health-related research when evaluating the credentials of a faculty member engaged in clinical research.

Little accurate data are available to assess workforce issues relevant to clinical research in pediatrics. Although one can use data from large managed-care organizations, these data are incomplete and often exclude the“quaternary” level patient who is referred to an academic medical center. Sophisticated manpower questionnaires are being developed in some fields such as nephrology which will permit a probing analysis of what the current workforce does on a daily basis, but these studies have the danger of being self-serving and extolling the virtues of a given discipline. A focus on population-based needs, as for example, the number of patients per 100,000 population, fails to account for the degree of poverty in a region, its rural population, and racial makeup, all of which influence disease incidence. Workforce studies must accurately account for time spent in patient care, research, teaching, and other service activities. The clinical investigator should also be seen as an individual who will be working in an integrated system with epidemiologists, health policy analysts, statisticians, and pharmacists.

One indication of the future workforce comes from the selection by residents of entry into subspecialty training. The career plans of 1994 general pediatrics examination first-time candidates are given in Table 3. The choices of those trainees in terms of subspecialty are provided in Table 4. An important aspect of this data are that 61% of first-time candidates choosing a general pediatric career are women, whereas men appear to be choosing subspecialty training in greater numbers, especially in the more technical subspecialties of cardiology, critical care, gastroenterology, nephrology, and pulmonology.

Table 3 Career plans of first-time candidates for 1994 general pediatric examination
Table 4 1994 general pediatrics examination: first-time candidates (2583) entry into: pediatric subspecialties

In summary, the pipeline for clinical researchers is potentially adequate, at least in terms of current pediatric residents choosing subspecialties. Programs are in place that can attract the resident who is stimulated by the intellectual excitement of clinical research. These residents, once identified, should be encouraged to attend the AMSPDC Frontiers in Science program and other specialty meetings. Training must be systematic and include study design, epidemiology, biostatistics, and exposure to hands-on patient studies. Many pitfalls must be overcome, including fragmentary training, lack of interdisciplinary training, and effective ways of dealing with education-related debt.

One must recall that the challenges of clinical research are potentially rewarding because of the translation of basic science advances into patient-related studies. In some instances, multicenter trials may offer the best hope for achieving success; however, the individual faculty member should gain recognition. In the future, multicenter studies must adequately fund investigators for the extensive time commitment to clinical research. Well conducted clinical therapeutic trials will underpin appropriate treatment methods, and outcomes research can save scarce clinical dollars. Tomorrow's clinical researcher will require more integrated training and will need an appropriate reward system, but can answer a difficult group of questions that can help many children.

PATIENTS AND POPULATIONS: MEETING OF THE TWAIN

Barbara Starfield

In 1980, the plenary session of the joint pediatric societies heard a presentation entitled “Patients and Populations: Necessary Links between the Two Approaches to Pediatric Education”(1). It might have also been subtitled “Never the Twain Shall Meet” because, although the theoretical justification was present even at that time, the means for achieving linkages were lacking. Today, the rapidly emerging health care revolution is providing the imperative for the linkage. What seemed like a good idea 15 y ago, on heuristic grounds, is now the order of the day. To keep the science of pediatrics moving ahead with the tide of health services delivery requirements, we must merge the two approaches: patients and populations.

Back in 1980, the theoretical argument for linkage was based on observations that certain aspects of knowledge could not be gleaned from research using only medical center patients. This included knowledge of the distribution of health problems, knowledge about how health problems initially present for care, and how they respond to management. Patients seen in medical centers are very unrepresentative of the general population. For example, data from the 1987 National Medical Expenditure Survey indicate that within a 1-y period of time, only 1 in 12 (8.6%) of children and adolescents in the United States are seen in hospital clinics. As a result, research done in such settings leads to a systematic neglect of certain types of problems, potentially misleading information about the significance of symptom complexes, and suspect information about the effectiveness of various therapeutic modalities. Research to inform decisions about adequate medical care, whether it is clinical research or health services research, requires access to non-hospital settings and especially to community settings with defined populations rather than patients.

Fortunately for research, the way in which health care reform is progressing makes linkages between the population and patient approaches much easier. That is, health maintenance organizations (HMOs), managed care, and integrated health systems all share the characteristic of defined populations. Prebudgeting for services at specified locales, whether or not the practitioners are themselves paid by capitation or fee-for-service, means that it is much easier to know where people receive their care and what care they receive. As a result it is now possible to track the natural history of problems from their inception through their progression to their resolution.

Information systems are much better now than they were 15 y ago. Computerization makes it possible to automate medical records and thus to facilitate the transfer of information; this works to both improve care as well as to provide the substrate for research to study illness phenomena and their responsiveness to interventions. The challenges, however, have not changed. The science of medical practice is still built on quicksand.

We are in the midst of efforts to base medical practice on guidelines; most professional groups are developing guidelines with no sound basis for making the judgments necessary for guidelines development. The literature is extremely sparse on just those characteristics that are needed for informed guidelines: the relationship between symptoms and subsequent disease; the adequacy of screening and diagnostic tests for the vast majority of clinical problems as they present, especially to primary care practitioners; the efficacy of the vast majority of treatments that are in widespread use; and, even, ways to measure whether or not these treatments are effective (let alone efficient). None of the challenges of guideline development can be accomplished without applying epidemiologic approaches to clinical data. The literature that does exist, and from which decisions are currently being made, is largely inapplicable to primary care. Take, for example, the matter of decision making about the need for diagnostic workups and therapeutic intervention. Most of the research on cut-points for abnormality in diagnostic testing derives from studies with patients who are not representative of the community because they are seen in medical center clinics. Patients presenting in primary care facilities have a much lower probability of illness than those seen in medical center clinics, even when their clinical presentation is identical(2).

This means that we overestimate the likelihood of disease in primary care when we apply clinical norms that derive from research in medical centers. This results in the paradoxical situation in which the country with the most highly developed research base, the United States, overuses technology in its delivery of health services. Unfortunately this does not guarantee our population anywhere near the best health levels(3). Consider, for example, how poorly the United States ranks among 11 Western industrialized nations on health indicators in various age groups, using the averaged rankings across several different measures: infancy, 10.3; childhood, 8.5; and adulthood, 4.8(4). Clearly, whatever is wrong with our health system has a greater impact on children than it does on adults. We are 7th of 11 in age-adjusted death rates, and 10th of 11 (at the bottom of the ranking) on years of potential life lost (and this remains the case even if deaths due to violence are removed from consideration). This relatively poor performance is not a result of “bad” behaviors of our population: we rank better than 4th of 11 for percent of population over 15 smoking(5), first (lowest) for alcohol consumed, and 2.5 (very low) for motor vehicle deaths per million miles driven(6) for the same 11 countries.

In contrast, our costs are much higher than elsewhere, a direct result of our heavy specialization and excessive and often inappropriate use of technology. It is curious that some specialists defend our need for yet more specialists on the grounds that new technologies would create increasing work for them. A more population-oriented view would consider the needs of the population, based upon research findings, rather than the availability of technology as the basis for these types of policy decisions.

Similarly problematic is the basis on which the medical effectiveness-or“outcomes” movement-is based. We are moving, correctly in my view, to a health care system in which the practitioners, or the organizations in which they work, will be judged on the basis of the results they achieve. This will be true both for primary care as well as for consultative and specialty care. Here we are faced with two dilemmas: we know little about what makes for effectiveness (for the same reasons that we lack information on which to base guidelines) and we do not know how to measure results. Up to now, it has been customary to base judgments of impact by changes in biomedical phenomena: test results. Yet we are increasingly recognizing that test results are often very imperfectly related to what is of most interest to people. People do not care about laboratory values in the abstraction; laboratory values mean something only in relationship to what they signify: threats to life expectancy; threats to the ability to function in daily life; threats to comfort; and threats to the seeking of higher purposes and greater goals. Our ability to measure these manifestations of health is primitive, yet is compelling, because our performance will increasingly be judged on our ability to contribute to them. Unfortunately, we are further behind in pediatrics than in adult medicine in measuring these aspects of health, and we have fallen even further behind in the last 15 y.

It is only very recently that investigators have begun to address the need for health status assessment in childhood. Complicating the task are the differences in developmental stage which make it impossible to use the same tool over the entire range of childhood, and the challenge of directly assessing health status because of our reliance on proxy respondents, generally mothers. Nevertheless, it seems possible to develop an instrument that, at least, uses the same domains (if not the same items) throughout the child range, and which, at least down to age 5 y, incorporates some aspects of assessment derived by interacting with children themselves. We have been using an instrument based on this conceptualization in adolescents aged 11-17; it has good properties of reliability and validity when used in community settings, such as schools, and in clinical settings(7). It considers health status to be composed of six domains: discomfort, satisfaction with one's health, disorders, achievement of developmental and social expectations, resilience, and risks. The same concepts should be relevant at younger ages as well, although the items that comprise the scales would be different. Whether or not the instrument proves sufficiently sensitive to reflect changes that occur in response to interventions remains to be tested; one major methodologic challenge is to devise a gold standard against which to test the validity of change reported by respondents. Past research provides ample evidence of lack of agreement between patients and physicians, between children and their parents, and between parents and children's teachers on various aspects of perceived health status. Who is to be believed? Against what criteria are we to interpret what we observe? Further advances in the technology of health status measurement-and therefore outcomes assessment- will clearly require large and diverse populations, not medical center or other clinical groups of patients, for testing of the whole concept of health status, methodologic aspects of its assessment, and the matter of cultural relevance of what is measured as appropriate outcomes in different populations.

As the country moves toward managed care and enrolled populations, there must be some way to compensate health services facilities for enrolling sicker or more vulnerable populations. Without such a means of “case-mix adjustment,” health service systems will find ways to avoid serving these more vulnerable and therefore more costly groups. How does one determine which populations are sicker than others and therefore in need of more extensive and expensive services? The most obvious solution is to pay more for populations that have higher proportions of individuals with chronic illnesses or disabilities. The problem is that there is a high degree of variability within populations with chronic illness or disability, so that some individuals require much more care than others. Neither the presence of chronic illness nor disability, per se, identifies individuals who require more resources than others. Past research has demonstrated that individuals who require higher resource inputs during one period of time are likely to continue to require continued high resource inputs. That is, higher users of services tend to remain high users over relatively long periods of time; the converse is also true. The problem, however, in paying health facilities according to past experience with resource inputs is that it is eminently “gameable.” That is, the surest way to guarantee a high capitation in prebudgeting for services is to encourage use of services in the prior budget period. Given that physicians themselves initiate the demand for at least 40% of all patients visits and, of course, almost all costs for services such as diagnostic tests and therapeutic interventions, paying on the basis of prior use will greatly inflate the costs of care to an unacceptable degree. The only reasonable solution is to devise a way to characterize the burden of morbidity by using some measure of health status. Our own population-based research has demonstrated that populations enrolled in health plans, both those that are prebudgeted as well as those that are under fee-for-service arrangements, can be so characterized by the various combinations of their types of illnesses in as a prior period. This method, known as the ACGs (for Ambulatory Care Groups) can predict subsequent use of resources at least well enough to be useful to the many HMOs that are now using them(8, 9).

In the 1980 talk at these very meetings, a case was made for increasing the cadre of physician-researchers who could successfully understand and implement the linkages between patient-based and population-based research. Subsequent developments have hardly been a success story. We still lack a critical mass of physicians who understand and are challenged by the vast array of topics that lie in the boundary between the two lines of endeavor. Without such researchers, decisions concerning the appropriateness of and payment for needed care will be made by those who have little appreciation of clinical realities. I end this paper, therefore, with the same plea I had in 1980-for more attention in pediatric training to the need to combine the clinical and the population approaches. The last two sentences of the earlier presentation still pertain today: “with the decline in availability of funds, child health research is easily sacrificed. It is now essential that the inherent links between the two types of research be strengthened.” Today, we are helped by the imperatives of our implicit health services reform. Let us finally arise to the challenge.

THE EFFECT OF MANAGED CARE ON THE FUNDING OF CLINICAL RESEARCH

Harvey J. Cohen

Over the past 20 y, tremendous strides have been made in the basic biologic sciences and in the areas of high technology that have poised pediatrics, and other areas of medicine, to be able to translate these exciting developments into new diagnostic and therapeutic modalities that have the potential to better enhance the health of children. Indeed, there have been advances in every field of pediatrics in the past 20 y that have utilized the knowledge gained from the basic sciences. We can and should be proud of these accomplishments. With the advent of managed care, including both discounted fee-for-service and capitation, further advances using this new biology and new technology are at risk. For the purposes of this discussion, clinical research is defined as including basic in vitro studies in materials of human origin, basic explicative pathophysiology, small therapeutic trials, large scale and often multicenter clinical trials, health services and health policy research, and outcome studies.

The clinical investigator who is uniquely trained in both medicine and the sciences has been the central coordinator of these activities. The needs of the clinical investigator include: an adequate number of patients to study a new diagnostic or therapeutic modality; time to plan, develop, carry out, analyze, and communicate the results of these investigations; support in terms of an appropriate milieu; and adequate numbers of individuals, space, and funding to perform these clinical investigations. The traditional funding for clinical research has come from either government contracts (federal, state, city, and county), hospital operations, physician practice plans, Medicare, and private donors. Private donors have included individuals, foundations, and pharmaceutical, biotech, and high tech companies, and other industries. Most of the clinical investigations have occurred within the context of the academic medical center. This is true whether the clinical investigations are instituted by an individual or are developed in conjunction with a pharmaceutical or biotech company.

Academic medical centers have grown tremendously from 1960 to 1992. Although growth has occurred in both basic sciences and the clinical sciences, the greatest amount of growth has occurred in the clinical faculty of the medical school, and there has been a more dramatic increase in the amount of funding for the academic medical center from the growth of the faculty practice plans than from the growth of NIH grants. In 1993, more than 50% of the revenue for academic medical centers came from clinical income. With the pressure from managed care to decrease revenue for physicians, this source of income from the academic medical center is gravely threatened, and thus will not be available to support investigative time and effort. The problems with the academic medical centers are that they are large and that most of their services are available elsewhere at lower costs. Major teaching hospitals costs are almost twice as much per day as those in non-teaching hospitals.

When one thinks about the needs for clinical research versus the needs for managed care, there appears to be a misalignment of goals(Table 5). Clinical research requires additional financial support, whereas managed care tries to decrease any extra cost that is present in the system, other than that necessary for the clinical care of the patient. The results of clinical research, although potentially of benefit to the patient, usually increase the costs of the care of these patients. Managed care looks at ways to decrease this cost. Clinical research is time consuming and often inefficient because of the need to plan, test, analyze, and report information that is gathered. Under managed care, physicians are required to be more efficient in the use of their time. Much clinical research requires a large number of patients, and this often necessitates referrals from many physicians. Under managed care, the incentive to the general practitioner is to decrease the number of referrals to specialists, because that results in a potential decrease in his or her own income. Finally, the goals of clinical research, for the most part, have concentrated on designing diagnostics and therapeutics to improve care, whereas the goals for managed care are to decrease the cost of care.

Table 5 Misalignment of goals

There are both advantages and disadvantages for academic medical centers in a managed care environment. Academic medical centers have an aura of quality that has allowed them to charge higher prices to support their teaching and research activities. However, payers are increasingly reluctant to pay these higher charges. In academic medical centers, faculty provide a large set of physician employees who can be coordinated with the hospital in a much more united way. However, as is true in many institutions, faculty guard their departmental independence and sometimes find it difficulty to unite effectively with their hospital. I feel that this alignment is necessary for our survival under managed care. Academic medical centers provide rare tertiary services that somewhat insulate teaching hospitals from competition and often attract patients from outside their local market. However, high tech services provided by teaching hospitals are only a small percentage of the actual clinical activities within the teaching hospital, most of which can be provided elsewhere. Finally, the large size of most academic medical centers and teaching hospitals may shelter them from a volatile market. However, size can breed complacency and also increase the difficulty of performing the changes that are necessary to survive in a managed care world.

Academic medical centers could easily be destabilized financially by managed health care systems. In addition, medical center budgets face reduced federal funding for graduate medical education and stricter cost accounting for research grants. Faculty practice plans face reduced revenues as managed care organizations cut back referrals, negotiate deep discounts, or refer to specialists at less costly facilities. Increased reliance on clinical practice revenue to support salary could limit faculty members' time for research.

Another effect of managed care systems on clinical research is the impact of adult specialists on pediatric subspecialty care. As adult specialists experience a decreasing patient volume, they are more likely to lower the age at which they care for a patient who previously would have been sent to a pediatric specialist. This decrease in the number of patients not only affects the financial viability of pediatric subspecialists and their research activities, but also decreases the number of patients available for clinical investigation.

These are the issues. There are no simple solutions. It is clear that academic medical centers have to change their organizational structure to be more valued as patient care partners with both payers and medical groups. They have to develop cost-cutting strategies, emphasize the quality of their work, and become leaders in outcomes research and measurement. Some of these cost-cutting strategies, however, are opportunities to fund clinical research. Many institutions are developing critical pathways which need to be studied and evaluated. Changes in the practice of the primary care physician are occurring with a decrease in the referral rate. The effects of these changes on the quality of care also need to be investigated.

We need to realize that as academic medical centers we are in three distinct businesses-patient care, research, and education. We will find it increasingly difficult to subsidize one business with the other. Thus, we will have to look at our clinical research as a business with its own sources of revenue and expenses, and we have to be able to develop mechanisms to support research programs that are, for the most part, independent of the clinical sources of income. We need to look at nontraditional funding for clinical research. This should include joint ventures with industry, payers, and employers to perform clinical investigations. Payers should, and the more enlightened ones will, support the clinical aspects of protocol-driven experimental investigations. It is also apparent that a significant amount of funds are present in the arbitrage of insurance companies. Whenever possible, the academic medical centers should try to assume the financial risk for the full care of patients, and, by instituting appropriate utilization management and quality assurance endeavors, try to manage that risk as effectively as possible with the goal of using the risk pool of dollars to support research. Health systems or hospital foundations need to be established to solicit funds specifically for clinical research. In addition, the development of physician hospital organizations (PHOs) can permit academic physicians to share in cost reductions with the hospital and use dollars to help support the research enterprise.

To fund clinical research in a managed care environment, we need to lower costs for delivering care while maintaining quality, develop networks to enhance the flow of patients, develop alternate sources of funding, focus clinical research in just a few areas, and have clinical investigators better publicize the results of their clinical research, both in terms of outcomes and potential effects on costs.

Finally, and I think very importantly, we need to mobilize our constituencies to act as advocates for support for clinical research. These constituencies include the primary care physicians (with whom we also need to develop business partnerships), patients and families, disease-oriented organizations, and the media. Using all of these approaches, we can and must develop strategies to finance and enhance clinical research at this critical juncture in the history of medicine. I feel that with time the pendulum will swing back to a recognition of the importance of funding clinical research with clinical dollars. How long will it take? I don't know, but this too shall pass.

If anyone tells you they know the answers as to how to have clinical research survive in a managed care environment, they are-at best-fooling themselves. However, clinical research will survive and potentially thrive, but it will need to be different from what any of us had imagined it would be. The stakes are high. If biomedical research disappears, then health care as we know it also disappears, and that will make us, as a nation, second rate. The challenge for leaders in pediatrics at academic medical centers is to face the issues of managed care, to adapt as we need to, but to hold dear and support the important aspects of clinical research so that the children of tomorrow are healthier than the children of today.

THE ROLE OF INFORMED CONSENT

Thomas H. Murray

The year 1995 marked the 50th anniversary of the end of World War II. Discussions of the ethics of research with human subjects were profoundly influenced by what the world learned about how the Nazis treated human beings in the name of science. Our first lessons on that experience came in the trials of war criminals at Nuremberg, including physicians and scientists who participated in experiments which still today are chilling in their cruelty and stark inhumanity.

The Nuremberg judges articulated principles with which to judge the offenses of the defendants who came before them. These principles have come to be known as the Nuremberg Code. The first article of the Code says this:

The voluntary consent of the human subject is absolutely essential. This means that the person involved should have legal capacity to give consent; should be so situated as to be able to exercise free power of choice, without the intervention of any element of force, fraud, deceit, duress, overreaching, or other ulterior form of constraint or coercion; and should have sufficient knowledge and comprehension of the elements of the subject matter as to enable him to make an understanding and enlightened decision(1).

The first sentence of the Nuremberg Code, taken literally, would prohibitall research on children, at least young children, as well as on all other persons incapable of giving morally meaningful free, informed, and understanding consent. The prohibition implied by this first article of the Nuremberg Code, if observed strictly, would ban therapeutic and nontherapeutic research alike on children and others who cannot consent for themselves. Because it deals only with research, the Code would not ban therapeutic innovation on children. But it leads to the bizarre implication that although we may try new treatments on children, we are not permitted to study the effects of such treatments systematically.

In practice, we have recognized some of the limitations of the Nuremberg Code. We do permit therapeutic research on children, relying on parents to protect their children's welfare. We even permit nontherapeutic research on children when the risks are minimal. Does this mean that the judges in Nuremberg were wrong? Not at all. Their task was to condemn and punish almost unimaginable evils committed in the name of scientific research. They were not asked to provide a detailed analysis of the moral subtleties residing in a wide variety of research efforts with different aims, in different settings, on different populations.

The legacy of Nuremberg for the ethics of research with human subjects has both positive and negative aspects. On the positive side, Nuremberg reminds us that even competent scientists can overreach. Contrary to a myth widely held after the war, some of the worst crimes were committed, not by the dregs and misfits of Germany's scientific community, but by its prewar leaders, internationally esteemed researchers(2). Nuremberg also reminds us that the human subjects of research must be treated with respect, and that informed consent is a vital manifestation of that respect.

On the negative side, Nuremberg has contributed to a tendency to confuse two related but distinct claims: 1) that it is wrong to do research on a human subject without that person's consent when he or she is capable of giving or refusing consent and 2) that all research on human subjects must be justified by some variety of consent-proxy consent, deferred consent, vicarious consent, and so forth. The first statement is about as widely shared a truth as any available to us in ethics. The second statement, however, does not follow from the first. There are large categories of persons-children, the severely demented, unconscious patients in the emergency room to name a few-for whom genuine consent by that individual is either unavailable or else fails to carry the moral significance that consent by a competent, informed adult does.

The variants of consent we have invented are, of course, all fictions. They may be well intentioned fictions, but they are fictions nonetheless. They are a reflection of the negative side of Nuremberg's legacy-the tendency to see informed consent as the primary moral warrant, or justification, for making a person the subject of research. It is not surprising that consent may have appeared as such to the Nuremberg Tribunal. Had Nazi experimenters followed strictly a rule that forbade using human beings in research without their informed consent, none of their morally hideous research would have taken place. Would-be subjects would have refused. Making informed consent the centerpiece of the ethics of human subjects research accomplishes two good things: it prevents research harmful to its prospective subjects, and it demonstrates respect for the moral standing, the dignity of those people. What it does not do is provide useful guidance for circumstances in which people are unable to provide morally meaningful consent ordis sent.

Consent long has been recognized as a powerful moral warrant. Ancient Rome had the maxim, Volenti non fit injuria: with consent, there is no injury. In everyday life consent has a clear meaning and wide application. Suppose someone chooses a color to paint their house but, after the job is finished, they decide they don't really like the color and refuse to pay the painter. The fact of their consent deprives them of any moral excuse for failing to fulfill their part of the contract. They are morally obligated to pay the painter.

What should we do, though, with those many potential subjects of clinical research who are not capable of providing morally meaningful consent? We can try to create “Just So Stories”-how the camel got its hump; how babies can be described as “consenting” to participate in research. This is essentially what happened when the National Commission for the Protection of Human Subjects of Biomedical and Behavioral Research considered the issue of research on children in the 1970s.

Two prominent ethicist-theologians set the terms of the debate. Paul Ramsey argued that any research on children that did not aim at benefiting the individual child was wrong. Parents may be permitted to consent to enroll their child in research intended to benefit that child, but not to research that might involve even minimal or negligible risks, if that research was not intended to benefit its child-subjects. Ramsey asserted that any parent who consented to such research committed “a violation of the nature and meaning of the responsibilities of parenthood”(3).

Richard McCormick argued, in contrast, for the idea of “vicarious consent,” which he said was based on “a construction of what the child would wish could he consent for himself.” The nature of our moral commitments to each other, McCormick claims, means that we all ought to be willing to consent the participate in research. But he goes further with his claim that, because we ought to consent, it is “a reasonable construction or presumption of our wishes to say that we would do so”(4). He extends that logic to children: they ought to want to consent to becoming research subjects; hence we can presume that they would consent.

Neither Ramsey nor McCormick provided helpful ways for thinking about the ethics of research on children, because both framed the question as did the judges at Nuremberg. “Consent is the heart of the matter,” McCormick said(4). But, as Ramsey insisted forcefully, children cannot consent, or rather their consent does not carry the moral force that an adult's consent provides. So, Ramsey concluded, no nontherapeutic research, however innocuous, may be performed with children. McCormick, recognizing that a complete ban on nontherapeutic research would be harmful to children's interests, tried to create a variant of informed consent, “vicarious consent,” based on the dubious premise that equates what children would choose to do with what they ought to do. Those of us who have raised children know only too well that a major part of our job is trying to bring what our children want to do into closer alignment with what they ought to do.

The problem, for both scholars, was that they accepted the framing of the problem bequeathed by the Nuremberg Tribunal. Ramsey was led to a denunciation of all nontherapeutic research with children. McCormick reached more sensible policy conclusions, but based on an implausible “just-so-story” about the relationship between children's wishes and duties.

A better solution for understanding the ethics of research on children is to recast it as an issue in the ethics of parental discretion. What may parents agree to on behalf of their children? May they permit their children to be exposed to minor risks or inconveniences to benefit others? This, in essence, is what happens in nontherapeutic, minimally risky research. I offer a more detailed account of this manner of framing the problem elsewhere(5). They key point is that rather than try to hammer consent into a shape suitable for dealing with children who cannot consent, or creating a fictitious “just-so story” about how they would consent, we do better to think in terms of the scope of parental responsibility and parental discretion. Widely accepted ideas about the moral scope and limits of parental discretion provide a basis for justifying minimally risky nontherapeutic research on children, while at the same time giving strong reasons for opposing research that poses unreasonable risks without compensating benefits to the child. The key advance here is finding an ethical analysis that fits well with its subject, one that does not have to rely on dubious “just-so stories.”

Another area of clinical research is trying to shake off a similar affliction. Research on emergency treatments with noncompetent patients has been curtailed because these patients are not able to give consent for themselves, suitable proxies are rarely available, and treatment must be initiated without delay. This yields the same sort of absurdity noted above: physicians are permitted morally to use whatever therapeutic interventions they believe are most likely to benefit their emergency room patients, but they are not permitted morally to do the research that would show which of the possible interventions would, in fact, most benefit those same patients.

An absurdity like this is a signal that our framing of the problem is defective. As in pediatric research, we have framed the ethics of emergency research in terms of informed consent. Because the genuine article-informed consent before enrollment in a study-was unavailable, researchers created a“just-so story” and called it “deferred consent.” In practice it amounted to enrolling a patient in a study, seeking an appropriate proxy to provide consent, and, if consent was not forthcoming, removing the patient from the study. The Office of Protection from Research Risks (OPRR) pointed out in 1993 that deferred consent was not really informed consent, not even real proxy consent. And so emergency research was curtailed further.

If what we want is to protect the interests of prospective subjects, assure that they receive treatments at least as good as the available alternatives, and respect their dignity, we can do that. Requiring researchers to ask a close relative or other appropriate proxy for permission to enroll or continue with a patient in a protocol is a very useful protection. But it does not carry the same moral significance as the patient's own informed consent, and calling it “proxy consent” does not change that fact. The problem is the “one-size-fits-all” framing of the problem-the presumption that whatever protections we devise must be couched as some variant of informed consent.

We should provide stout protections for emergency patients who are incapable of giving consent. Acknowledging that for such patients, in such circumstances, genuine informed consent is a fiction, is simply admitting the truth. It is not a wholesale invitation to use these patients in any and all experiments. They are extremely vulnerable, in no position to accept voluntarily additional risks, and deserving of our deep concern. What we need are a set of protections well suited to their particular circumstances. Such an array of protections was suggested recently in a consensus statement(6). The statement describes the impossibility of getting consent in certain emergency circumstances and the desirability of research on alternative interventions. It then provides a set of protections which, if observed scrupulously, would protect prospective subjects in emergency research, even when consent is not immediately available. It is unfortunate that such a sensible and ethically defensible position took so long to be developed. But that too is one of the legacies of Nuremberg.