3Barriers to Interdisciplinary Research and Training

Jargon

Scientists trained in a discipline learn to speak a specific language and adopt the analytical and methodological constructs that have accumulated in that discipline. This constitutes a form of professional socialization that serves as an important part of the training experience, but it can present obstacles to interdisciplinary research.

We speak the language of our discipline, which raises two problems: first, we may not understand the languages of the other disciplines; second, more dangerously, we may think that we understand these, but do not, because although the same terms are used in different disciplines, they mean something very different in each.

— Margaret A. Somerville⁵⁵

In addition, the problem exists that “different disciplines are continually rediscovering one another's discoveries, because they all have different names for them” (P. Smolensky, IOM Workshop, 1999).

Communicating with another discipline requires time and work. An extensive effort must be made to learn the language of another field and to teach others the language of one's own. Many have recognized that this barrier must be overcome before successful collaboration can occur.^8,29,40,52

Intellectual Turf

By definition, interdisciplinary efforts bring together researchers who have different expertise. Pride in one's discipline and its methods can be instilled during graduate school.⁵⁹ As a consequence, other disciplines might be viewed as less rigorous or important. To work together, people must understand and appreciate the value and limitations of both their own and others' methods. Groark and McCall²⁴ have written about the distrust between researchers and clinical providers: each group believes in its own superiority. The same can occur between disciplines.^16,38 Heated discussions can result when people with different backgrounds try to assert the correctness of their views (D. Tracer, IOM Workshop, 1999; P. Smolensky, IOM Workshop, 1999).

Team Building

Teamwork requires trust in another's skills and expertise. If these are outside one's field, as implied in interdisciplinary research, they might be difficult to evaluate.⁵² Interdisciplinary team members sometimes have difficulty in evaluating each other's performance on a given project because the criteria that are appropriate for such an evaluation are not familiar.⁶ Moreover, reward systems and practices regarding authorship differ among disciplines.^33,41 Good communication skills are helpful in alleviating such problems, but most members of interdisciplinary teams lack training and proficiency in such skills.¹⁵

Relationships among team members affect productivity.⁹ Team members need to be able to compromise and cooperate.^24,40,52 A feeling of community can facilitate interactions in an interdisciplinary team.¹⁷ Many stress that good communication among team members is essential for the process to succeed.^8,18,40

In building an interdisciplinary team, clarity regarding roles, expectations, and authority—particularly with regard to sharing of data and resources—is important for success.^12,22 Mutually acceptable policies for disseminating research results (including authorship) and facilitating achievement of team members' personal and professional goals need to be developed.⁴⁰ A possible consequence of not having a team process is that crucial voices will be missing in defining and solving the research problem.³⁹ All that suggests that it takes interpersonal skills to conduct interdisciplinary research and that expertise in disciplinary methods alone might not be sufficient.

Leadership

Interdisciplinary research teams need leaders who understand the challenges of group dynamics and who can establish and maintain an integrated program. Leaders need to have vision, creativity, and perseverance. To establish a successful interdisciplinary program requires education of scientific colleagues and administrators about the potential value of interdisciplinary research. To coordinate the efforts of a diverse team requires credibility as a research scientist, skill in modulating strong personalities, the ability to draw out individual strengths, and skill in the use of group dynamics to blend individual strengths into a team. Some have suggested that the best persons to direct interdisciplinary teams are mature scientists with well-established research careers who have conducted interdisciplinary research of their own.⁵²

Facilitating Interactions

In the experiences of the committee members, chance interactions can promote interdisciplinary collaborations. The casual discussion of research at the coffee machine, the fortuitous meeting in the corridor with a colleague from another department, an interesting seminar, or interactions among students and postdoctoral scientists—all can trigger collaboration. Buildings that isolate laboratories and research groups from one another can limit this type of interaction. But, bringing people together through the creation of central facilities or common areas can increase its probability.

Common location of faculties leads to interactions, scientific discussions, and possibly new insights on research data. For example, Washington University invested $28 million in a new laboratory building to house a shared imaging facility, animal facility, and psychology laboratories with the intent of encouraging interdisciplinary collaboration.⁴³ More recently, the University of California, Berkeley launched an interdisciplinary “Health Sciences Initiative” that includes the construction of two buildings that will house laboratories of researchers from several departments.⁵¹ The intent is to encourage daily interactions among the investigators in a variety of fields, such as physics and molecular biology. Similarly, in an effort to integrate the intramural neuroscience research, NIH is planning a center that will co-localize the basic and clinical neuroscientists from nine institutes.⁵³

Architectural design can promote interactions,¹¹ but virtual proximity, through the Internet or videoconferencing, provides another opportunity for achieving intellectual proximity as needed and deserves further research and consideration.

Academic Structures

A 1998 editorial in Science²¹ asserted that the modern university is “partitioned along academic lines that no longer truly reflect today's intellectual life. These academic groupings are now just categories that accountants and business managers use to build a budget. The issue is most pronounced in the scientific disciplines.” Others have written about the conflict between the scientist's professional need for autonomy and identity and the organization's need for an efficient bureaucratic structure.¹⁸ Despite some pioneering approaches (e.g., Rockefeller University's absence of departments and the University of California, Berkeley Health Science Initiative, which coordinates hundreds of scientists in multiple schools⁵¹), institutions change slowly. Thus, in considering interdisciplinary research and training, the continued presence of some system of academic departmentalization must be acknowledged.

Academic departments create an environment within which training and research occur. Discipline-oriented departments constitute a functional authority structure in charge of teaching, faculty recruitment, advancement, and promotion—as well as degree programs and courses. Funding processes reinforce the departmental structure. The departmental structures of universities have evolved primarily on the basis of scientific advances. However, departmental organization changes relatively slowly. The priority given to contributions in fields that correspond to departmental structures can inhibit interdisciplinary approaches.

Institutional policies regarding allocation of laboratory space, hiring, and promotion policies vary considerably, but the department chair generally has a major influence. Institutions vary in how they distribute credit for successful grants. Some allocate funds and resources among the investigators and their units, but others credit only the person listed as the principal investigator. That can penalize coinvestigators in other departments, in that their home departments might get no credit. The cross-departmental nature of interdisciplinary research is likely to compound the problems of allocation of credit and research resources when grants are funded across departments or schools.⁵²

Promotion and tenure policies and practices are major motivators and controlling devices for academic scientists. In a survey of 366 faculty spanning five disciplines in six universities, Moore⁴¹ found that quality and quantity of publications, followed by grantsmanship, were the most important criteria for tenure but that departmental politics guided all these considerations. It can be difficult, therefore, for junior faculty whose interests range beyond the formal subject matter of a given department to be viewed as either making substantial contributions or as being appropriate for advancement in a given department.

To counter some of those concerns, many universities have established interdisciplinary programs or centers that cross departments. For example, at the committee's workshop (IOM Workshop, 1999), Dr. David Tracer described a program in Health and Behavioral Science at the University of Colorado, and Dr. Donald Heistad described a center for aging research at the University of Iowa. Faculty members in such programs have a “home department” but participate in the research and training functions of the interdisciplinary effort. This approach does not eliminate the obstacles to interdisciplinary research within a department, but it does formally recognize the interdisciplinary endeavor.

Interdisciplinary research requires a commitment from university administration. A cross-departmental program can suffer if the administration does not consider the needs for faculty, space, and funds. Through control of faculty positions, the university leadership can promote collaboration, for example, by requiring a position to be jointly supported by two departments (T. Detre, IOM Workshop, 1999) or by ensuring that interdisciplinary programs do not drift back to a primarily single disciplinary perspective (A. Binder, IOM Workshop, 1999).

Funding agencies can facilitate commitment from high-level university administrators. For example, the National Science Foundation (NSF) Integrative Graduate Education and Research Training (IGERT) program requires support from the university administration to develop an application. Each year, NSF brings together the principal investigators for the grants and a high-level university administrator to discuss obstacles. This educates and involves the administrative officers and promotes the sharing of approaches among funded universities. Similarly, some National Institutes of Health (NIH) funding mechanisms call for institutional commitments. For example, the National Cancer Institute's Specialized Program of Research Excellence (SPORE) in Breast Cancer, which supports translational research, requires applicant organizations to supply a statement of institutional commitment that describes the mechanisms by which the program is given high institutional priority; the infrastructure for establishing, maintaining, and monitoring the program; and the plans for recruitment to strengthen the scientific capabilities of the program.⁴⁶ In addition, the institution is required to provide evidence of tangible commitment of space and money.

Requirements of institutional commitment as a condition of application and funding are likely to bolster institutional support. This support could influence the departments, eventually affecting tenure and promotion decisions. The IGERT administrator at NSF reported, for example, that some institutions are now making changes to their tenure procedures for young faculty to decrease the problem of tenure “jeopardy.” This problem results from the lack of recognition by the institution and faculty of interdisciplinary efforts in the tenure decision process (W. Jennings, Third IOM Committee meeting, 1999).

Publications and Professional Organizations

Professional organizations and journals are key ingredients of disciplinary identity.¹⁰ Their major impact is in providing an outlet for dissemination of information. Whitley⁵⁸ commented that “the existing set of journals in a science constrain and direct research topics and ways of working on them.” Although some societies (e.g., the Society for Neuroscience) are interdisciplinary by design, many are discipline-specific. Similarly, some journals (e.g., Science, Nature, and the New England Journal of Medicine) are broad-ranging, but the vast majority are discipline-specific. Multidisciplinary researchers, as opposed to more discipline-identified professionals, face problems of professional identity and outlets for publication in prestigious journals.

Research reputation is essential to obtaining support for research, gaining employment, getting promotions, and winning grants. Authorship of papers is perhaps the most important predictor of one's success in these activities.²⁶ The size of collaborative groups can limit the frequency of first-authorship and thereby pose problems for interdisciplinary researchers. New editorial policies on authorship are arising out of ethical concerns about multiauthor papers.³² To protect against fraud, there is a growing call to define the contribution of each author.^28,34,57 Acknowledging individual contributions to a research effort might provide a means to provide credit to people involved in interdisciplinary research, above and beyond senior authorship.

Federal Funding

Scientists perceive that they will encounter difficulties in obtaining support for and conducting successful interdisciplinary research because funds for such research come from many agencies with different programmatic emphases. In contrast with that perception is the apparent consensus among federal research sponsors that interdisciplinary research and training are essential (NIH directors, consultation, 1999 and W. Jennings, Third IOM Committee meeting, 1999). The importance of interdisciplinary research to NIH is reflected by a review of the Requests for Applications (RFAs) that are posted to encourage research in fields of special interest. An average of 23% of all RFAs issued in 1992–1999 by the National Institute on Aging (NIA), the National Institute of Neurologic Disorders and Stroke (NINDS), the National Institute of Mental Health (NIMH), Office of Behavioral and Social Sciences Research (OBSSR), the National Institute on Alcohol Abuse and Alcoholism (NIAAA), the National Institute of Nursing Research (NINR), and the National Institute on Drug Abuse (NIDA) addressed interdisciplinary research (Figure 3-1). In addition, interagency collaborations bring together multiple perspectives. As examples in chapter 4 illustrate, partnerships among NIH institutes, among government agencies, or between government and the private sector often provide a broad base of support for interdisciplinary research and training.

Peer Review

Although strong peer review is considered to be the basis of a successful funding system, it is a long-held perception that it puts interdisciplinary research proposals at a disadvantage. Many have suggested that the traditional proposal evaluation by discipline-specific study sections makes it difficult for interdisciplinary proposals to compete successfully.⁵² In an “open letter” to NIH, Howard Morgan⁴² wrote that, “peer review of cross-disciplinary research presents the problem of the definition of a peer.” Even if all disciplines covered by a proposal are represented in a review committee, unless the committee members themselves have tried to do interdisciplinary research in the field in question, they might not appreciate the issues. Morgan called on NIH to correct the problems associated with the review of interdisciplinary grants. In their responses, representatives of NIH acknowledged the legitimacy of the concern.^23,37

To address those concerns, NIH's Center for Scientific Review initiated a major overhaul of its system with the creation of a Panel on Scientific Boundaries for Review in 1998. In redesigning the integrated review groups (IRGs), the panel tried to ensure that all biomedical sciences were encompassed by the IRGs, each IRG (comprising several study sections) could cover the full range of expertise necessary for review, and all studies (basic and clinical) related to a particular disease would be covered within a single IRG (or a related set if necessary).² Within that structure are the four IRGs for the neuro- and behavioral sciences created in response to the incorporation of NIMH, NIDA, and NIAAA into NIH from the former Alcohol, Drug Abuse, and Mental Health Administration (ADAMHA).¹ The new peer review system is specifically intended to be more supportive of translational and interdisciplinary science. The first phase of the peer-review reform, completed early in 2000, created the IRGs. The next phase, expected to take another 2 years or so, will define the constituent study sections.² It will be important to monitor the new system and to track whether it increases the number of shared funding arrangements and collaborations and whether interdisciplinary proposals fare as well as single disciplinary proposals in the review process.

Duration of Training

The length of time required to complete training in more than one field—whether in the basic or clinical sciences or in a combination of the two—can be discouraging. The explosion of information in each scientific discipline raises concerns about how long it would take to attain expertise in one, let alone two or more, fields. The requirement for depth of knowledge in one field during graduate school might seem to preclude obtaining breadth in other disciplines within a reasonable period.⁵² The length of graduate training is increasing and the time spent in postdoctoral positions is also on the rise.^45,48 Interdisciplinary requirements might worsen the delay in starting a career.

That concern is especially relevant to clinician-scientists. To complete general and subspecialty training in most medical specialties requires 5 years or more after medical school. Little or none of this clinical training is directed toward research, so those interested in research must complete an additional 2 years or more either in the graduate student or at the postdoctoral level. The additional years of training burden come at a time when trainees are also trying to advance other parts of their lives, such as having families; this is a problem particularly for female physicians.

Debt

Educational costs are increasing. To pay for their training, many students are taking loans for their undergraduate or professional education. A 1997 survey of recent doctorates in psychology revealed that 63.8% of the graduates reported debt.^4,35 Of them, nearly one-third had debt of $41,000 or more. Debt was greater for those who specialized in clinical subfields (average, $35,000) compared with research (average, $15,000). The debt was greater for graduates with clinical PhDs from professional schools (average, $60,000). Medical students can carry an even higher debt. In 1997, the mean debt of graduates was over $80,000; some owed more than $100,000.^7,19,31 In fact, at the University of Medicine and Dentistry of New Jersey, the reported debt exceeded $100,000 for 40% of the 1998 graduating class.⁷ Such a burden of debt is a strong factor against a decision to continue in research, let alone interdisciplinary or translational research.

An additional disincentive is the relatively low salary in academic research careers compared with positions in clinical practice or private industry (for example,see the American Psychological Association, Full-Time Salaries of Psychologists⁵), especially in the face of large debt. Although some surveys have suggested that debt does not influence choice of specialty (for example, see Kassebaum and Szenas³⁰), it could influence the choice to pursue interdisciplinary or translational research. Graduates from an academic pediatric residency program were surveyed on factors that influenced their career paths; over half those responding indicated that salary expectation discouraged them from entering research.³⁶ Furthermore, the financial concerns might have a particular influence on the career paths of minorities and women. Analysis of data from an Association of American Medical Colleges survey revealed that “being a woman and being an underrepresented minority were associated with plans to enter salaried clinical practice.”²⁵

To encourage clinicians to engage in research, Congress passed a law in 1998 establishing Loan Repayment Programs (LRPs) related to AIDS research (AIDS-LRP), general research (General-LRP), and clinical research (CR-LRP).^27,47 These programs allow NIH to repay up to $35,000 per year in eligible education loans for participants who obtain research employment at NIH. Most graduate degrees (for example, PhD, DDS, DVM, DNSc) qualify a candidate for consideration, but the emphasis is on physicians. In addition, to qualify for the CR-LRP, a person must have a “disadvantaged background.” Others have recognized the value of these loan-repayment programs to encourage careers in clinical investigation and have urged Congress to pass legislation to address the national need.⁴⁴ The committee similarly believes that extending these loan repayment programs could provide an increased incentive to pursue interdisciplinary research training.

Job Opportunities

There is a concern among junior scientists that training in interdisciplinary fields will not prepare them for careers. “Jack of all trades, master of none” is a refrain that is often heard when interdisciplinary training is proposed.⁵⁰ It raises the question, Is there a market for people trained with an interdisciplinary perspective? Some have reported difficulties in finding positions for graduate students who received interdisciplinary training.⁵² But, a sampling of the employment opportunities posted in Science in 1975–1999 would suggest that the market is growing. In November 1975, 6.7% of the employment ads reflected multidisciplinary requirements. By November 1985, the percentage had grown to 11%. In November 1995, 19% directly or indirectly indicated the need for integration of approaches or suggested collaborative opportunities. In November 1999, 31% of the ads mentioned a collaborative environment or stressed team efforts. Furthermore, the development of large common laboratory space to facilitate joint research, as described above for the University of California, Berkeley and for Washington University, would be expected to accelerate employment of scientists who are competent in more than one field.

Private industry offers another avenue for those with interdisciplinary training. Representatives of industry indicated that they are looking for people who can work in an interdisciplinary environment (A. Cato, IOM Workshop, 1999).¹³ The process of moving a drug from the discovery process to clinical trials requires teamwork. The growth of biotechnology requires scientists who can work in more than one discipline. There has been a strong trend toward increasing employment of scientists in industry compared with academe. Data from the American Psychological Association indicated a dramatic shift of employment settings for PhD psychologists between 1973 and 1997 away from colleges, universities, and medical schools and toward for-profit settings.³ A 1995 membership survey of the Society for Neuroscience similarly showed an increase in the percentage of members working in industry from 3% to 10% since 1982.⁵⁴ A survey by the Federation of American Societies for Experimental Biology showed a similar trend for biomedical scientists in general.²⁰ In response to those trends, a study by the Committee on Science, Engineering, and Public Policy of the National Academies recommended that graduate programs provide greater scope and interdisciplinary training.⁴⁵

Staying Current

As the volume of information grows in a given subspecialty, the number of publications that a person needs to read to stay current in his or her own field forms another deterrent to becoming more interdisciplinary. Figure 3-2 illustrates the exponential growth in publications in selected fields of neuroscience. A search on key words in MEDLINE and various psychological and sociological databases yielded many times more references on each term in 1998 than in 1974. Keeping up to date in any of these fields is challenging. Keeping up to date in several would be daunting. A similar analysis in 1986 revealed the rapid increase in the occurrence of the word interdisciplinary in the title of papers between 1951 and 1982, perhaps suggesting a concurrent increase in interdisciplinary research.¹⁴

It is entirely possible that on-line journals will begin to break down the barriers of access to information. Accessibility at the speed of the Internet is making it far easier to learn what is being done in other disciplines, although the language and jargon barriers cited above will continue to haunt those developing the search engines for on-line publications. Citation services, such as Cite Trak, are now available to alert one to papers published in a particular field of interest. The scientific equivalent of bulletin boards or chat rooms, such as Brain Research Interactive (an on-line version of papers that are later published in BrainResearch), could provide an important forum for examination and interchange of ideas.

Midcareer Retraining

It is not just newly minted doctoral and postdoctoral scientists that face challenges in advancing interdisciplinary interests. A midcareer scientist who decides to retool or shift focus faces a difficult task. To continue a productive research career, funding mechanisms must be available to allow midcareer scientists to update themselves or retrain in another, quickly moving field. Such educational experiences require the opportunity to add to basic knowledge in one or more disciplinary skills or to broaden one's knowledge, not necessarily to create the capacity to conduct wholly different research alone, but, rather, to learn enough about other disciplines to work as a productive team member.

Special Challenges for the Clinician-Scientist

For decades, concerns have existed about the decreasing number of clinician-scientists.^49,56,60 Such people are needed to bring basic science into the clinic; they do clinical investigations. The percentage of graduating medical students with an interest in research has been declining.⁴⁹ The large debt, long training period and low academic salaries described above contribute to the problem. Other disincentives are lack of encouragement to pursue this career path and the clinical demands of academic health centers that draw staff away from research.

The participation of the clinician-scientist in interdisciplinary research poses additional challenges for professional advancement in most university systems. In some measure, this is due to the longer initial development times of most interdisciplinary projects before they yield results adequate for publication. If studies are concerned principally with the etiology of human disease, the transfer of preclinical principles to the bedside, or strategies for reducing the burden of illness, the problems of lead time to publication are amplified. Because many academic institutions have promotional policies that allow 4–10 years before faculty are “up or out,” there is a general perception that entry-level faculty should not participate in such studies. One strategy would be to delay participation in interdisciplinary research until midcareer, when tenure and other forms of job stability have been achieved. Unfortunately, that strategy could eliminate people from interdisciplinary research at the height of their creativity. Similarly, people who wait until midcareer for such a transition might lack necessary interdisciplinary strategies or approaches unless mechanisms of continued career development are in place.