The relationship between inter- and transdisciplinary research and potentially transformative science is poorly understood. We use a case study of a long-term transdisciplinary research effort on hantaviruses combined with findings from studies of team science to generate a hypothesized model that links cross-disciplinary collaboration with transformative scientific outcomes. We show that potentially transformative research depends on the existence of an interesting and worthwhile problem to which participants can contribute in salient ways, human and material foundations within disciplines, collaborative mutualism across disciplines, and a transformative learning process that enables knowledge integration across diverse perspectives. Transformative learning theory suggests that new, integrated conceptual understanding is initiated by disorienting dilemmas. We argue that engagement in cross-disciplinary collaboration produces disorienting dilemmas that initiate transformative learning. Our hypothesized model provides a generalized framework for understanding how transformative learning occurs in cross-disciplinary collaboration and how that can lead to transformative science.
The One Health Initiative (www.onehealthinitiative.com) views the emergence of zoonotic infectious disease as resulting from the dynamic interactions of the ecosystems of wildlife, domestic animals, and humans (Monath et al. 2010). The initiative explicitly states that interdisciplinary and cross-sectional approaches are required for the prevention, surveillance, monitoring, control, and mitigation of emerging infectious diseases. The terms interdisciplinary and transdisciplinary describe research efforts that integrate a variety of academic perspectives with nonacademic perspectives of other stakeholders (Klein 2010). These types of diverse, collaborative efforts allow the detection, investigation, anticipation, and potential prediction of outbreak events—as well as responses to those events. In this article, we review a historical transdisciplinary response to a zoonotic infectious disease outbreak that is an example of the One Health concept and use it to generate an initial model for understanding the potential for novel, high-impact outcomes generated by such complex collaborations.
In 1993, a series of unexplained deaths in the American Southwest prompted an immediate response from various groups of health practitioners and biologists representing a variety of institutions. The etiology of the outbreak was determined to be an exotic, viral, hemorrhagic fever virus—the Sin Nombre virus, a member of the Bunyaviridae family of RNA viruses. The common native deer mouse (Peromyscus maniculatus) is the primary reservoir species that carries this zoonotic pathogen, which causes hantavirus pulmonary syndrome (HPS). To understand the reservoir population, biologists, mammalogists, climatologists, and evolutionary biologists became involved with the medical community. One year prior to the outbreak, the Institute of Medicine (Lederberg et al. 1992) released a report in which concerns were voiced regarding the likelihood of disease outbreaks going undetected until they were well under way, because of a lack of coordination between the relevant organizations. In contrast, the overarching outcome of the hantavirus investigation was the rapid determination of the causative agent and its reservoir through a broad collaboration of scientists and health professionals and the use of highly discriminating molecular epidemiologic technology. The outbreak and subsequent investigation led to the description of a novel clinical syndrome determined by unusual pathological findings; an exploration of a new form of medical shock; and the first identified North American viral infectious disease for which the reservoir species, rather than an arthropod vector, was involved. Ultimately, crucial links were found among the El Niño Southern Oscillation (ENSO), increased precipitation, vegetation primary productivity, deer mouse populations, and the occurrence of HPS.
Understanding the links between ENSO and the trophic cascade that leads to an increased risk of exposure to the hantavirus enabled public health officials to predict the risk and to release health warnings of potential outbreaks in subsequent years (Glass et al. 2000, 2006). Specific scientific outcomes and their associated contributors demonstrated an exceptional level of transdisciplinary collaboration from the local to the national level. The remarkable collaboration that emerged in response to the hantavirus outbreak was documented by Yates and colleagues (2002), whose work was awarded the first sustainability award by the Ecological Society of America. The collaborative effort transformed the research of the scientists who were involved, created new paradigms in the zoonotic infectious disease community, and left a lasting positive impact on medical community practices. This transdisciplinary group produced a wealth of outstanding science with an immediate impact.
The hantavirus investigation is a classic example of transformative research: “driven by ideas that have the potential to radically change our understanding of an important existing scientific or engineering concept or leading to the creation of a new paradigm or field of science or engineering. Such research also is characterized by its challenge to current understanding or its pathway to new frontiers” (USNSB 2007, p. 10). More recently, the US National Science Foundation postulated that transformative research is often associated with interdisciplinary discourse (www.nsf.gov/about/transformative_research/characteristics.jsp). Although a crisis was the catalyst for the group of scientists working on the hantavirus, there is an urgent need to catalyze such transformative research in the absence of immediate crises. However, little is known about how to effectively launch inter- and transdisciplinary collaborations. An emerging science of team science is beginning to yield an understanding of key factors present in effective team science (Stokols et al. 2008), but this knowledge does not always translate into understanding how to generate those factors (Pennington 2011). For example, knowing that it is crucial to develop a shared vision does not provide guidance about how to more effectively develop a shared vision across groups with differing perspectives.
The hantavirus experience highlights a number of key elements that support transformative transdisciplinary science, comparable to findings from several studies of research teams. In this article, these findings are then used to hypothesize models of such efforts that can begin to inform the process of generating transformative science. First, pertinent details about the teamwork and transformative breakthroughs that occurred are articulated. Then, those details are abstracted into categories of interacting issues, and the nature of each category is expounded through a synthesis of the hantavirus case with empirical data from studies of other research teams. Last, process models are hypothesized. In this article, we generate new ways of understanding the relationship between inter- and transdisciplinary research and potentially transformative science and identify transformational learning as a key mediating process.
The hantavirus story: Emergence of the hantavirus in North America
In the spring of 1993, the unexplained deaths of a young New Mexican couple from the Four Corners area was reported: They died rapidly of overwhelming respiratory failure with an unclear etiology. Additional cases were subsequently reported, primarily in healthy young adults, with an alarming case fatality rate initially exceeding 80%. The earliest patient symptoms mimicked influenza—that is, extremely nonspecific, undifferentiated symptoms, such as fever and general myalgia. Between 1 and 4 days after the initial symptoms, the patients typically experienced shortness of breath and coughing, which generally heralded the onset of rapid pulmonary distress, pulmonary failure, and death within 12 hours. The laboratory parameters for the patients indicated elevated white blood cell counts, low platelet counts, and atypical lymphocytes. Scientists recall a number of key interactions across disciplines that enabled major diagnostic breakthroughs. For example, in one setting, investigators listed all diseases that could account for any of the clinical, pathologic, or epidemiologic aspects of the outbreak (approximately 50). A facilitated differential diagnostic discussion of the identified possibilities allowed a consensus conclusion to exclude all but three potential etiologies: influenza A, viral hemorrhagic fever, and a novel agent.
During the 19 days after the first initial Four Corners area deaths, clinicians, public health professionals, epidemiologists, pathologists, laboratorians, and molecular biologists collaborated to determine the etiology of HPS. Named Sin Nombre, the hantavirus was a previously undescribed species in the bunyavirus family of viruses. This determination was enabled by the use of polymerase chain reaction (PCR) applied as a new field diagnostic technology. Mammalogists determined that homologous sequences of virus samples from different patients were comparable with homologous sequences of virus samples taken from deer mice (P. maniculatus), which confirmed the host–reservoir relationship (Childs et al. 1994). HPS, a zoonotic infectious disease, had never before been recognized in the Western Hemisphere. Other species of New World rodents carry varieties of hantavirus, but the zoonotic dynamics for each virus is different; most are not nearly as dangerous as the one carried by P. maniculatus.
The positive identification resulted from tests of frozen tissue samples that were obtained from deer mouse voucher specimens collected prior to 1993 and archived at the Museum of Southwestern Biology at the University of New Mexico and at the Natural History Collection at Texas Tech University. Once the HPS diagnosis was confirmed, questions arose about its source, because the medical community was skeptical that similar deaths could have been overlooked in the past. The virus was either the result of a long-term, coevolutionary, and host-specific process or a rapidly evolving virus that had adapted quickly to climate variability. Cross-examination of the deer mouse specimens quickly confirmed that the outbreak had not resulted from recent introduction of the virus; rather, the virus had been widely distributed in its rodent reservoir for decades, although there were times when deer mouse populations showed little or no infection. Although the hantavirus was first observed, reported, and identified in the spring of 1993, that was not its first occurrence. Retrospective studies have documented hantavirus-related deaths in New Mexico as early as 1978 (Zaki et al. 1996). Indeed, oral tradition among the Navajo revealed an understanding of the risks of mice in homes (Chapman and Khabbaz 1994). This finding prompted ecological research by biologists, mammalogists, evolutionary biologists, and climatologists that ultimately revealed the episodic nature of mouse population changes, the connection to ENSO, and the historical perspective on outbreaks (Parmenter et al. 1993, Brunt et al. 1995, Mills et al. 1997, 1999, Hjelle and Yates 2001, Yates et al. 2002).
The periods of the highest human risk of exposure are most likely to occur when high deer mouse population densities are correlated with a high percentage of mice that are infected with the Sin Nombre virus. Such high mouse and virus densities are associated with abundant seeds after substantial rains. Variable population cycles observed in the deer mouse at Sevilleta research station were linked to ENSO. ENSO events, commonly referred to as El Niño, typically result in higher moisture in the American Southwest, with increased snowfall during a late winter, compounded with increased early spring rain. Increased precipitation and recharged soil moisture results in enhanced spring primary production by plants, which produces large amounts of seeds and provides a nutritional, caloric food source that, in turn, rapidly increases the population of rodents (Parmenter et al. 1993). Because natural predators that control rodent densities, such as hawks, owls, and snakes, do not reproduce as quickly as rodents, there is a period of time during which rodent densities are higher than usual. ENSO conditions prevailed from the fall of 1991 through the spring of 1992, which led to increased deer mouse densities in the spring of 1993, when the outbreak occurred. Subsequent ENSO events in 1997–1998, 2002–2003, and 2006–2007 have been correlated with increases in the number of cases in following years (www.cdc.gov/hantavirus/surveillance/annual-cases.html).
The primary means of virus transmission from the deer mouse to humans is aerosolized particles of contaminated dust breathed into the lungs. No arthropod vector is involved. For example, in a rural setting, enclosed structures, such as barns and utility sheds, house infected deer mice that shed the virus through their feces and urine, which accumulate in the dust. The contaminated dust is disturbed by wind or other mechanisms, such as sweeping, which exposes unsuspecting persons to particle inhalation (Lee and van der Groen 1989). This method of exposure was an unexpected finding, because direct contact with a biotic vector (animal or human) is typically assumed in studies of infectious diseases. Previous assumptions that a biotic vector was required for disease exposure limited the ability of researchers to conceive of possible mechanistic solutions to the problem (Zeitz et al. 1995).
Our understanding of the virus, its occurrence, the reservoir, environmental drivers, and transmission mechanisms has enabled researchers and public health officials to create more-effective prevention programs. Satellite tracking of rainfall across the mountains of New Mexico, Arizona, and California can predict high and low deer mouse population densities (Eisen et al. 2007). Warnings for periods of high risk can be made well in advance of the actual periods of high risk (Glass et al. 2006). During these high-risk periods, deer mice can be effectively trapped or poisoned in human dwellings, which lowers the potential of human exposure. Disinfection of indoor areas with bleach—before sweeping—is the best-known preventative measure but one that is counterintuitive and not always followed (McConnell 2009).
Now that the causative agent has been identified, a small number of cases of HPS are observed in the American Southwest each spring and early fall. New biotechnology procedures have been developed that identify antibodies made by humans to denature the hantavirus. This reveals the identification of hantavirus antibody compounds that can be used to test for human exposure to the virus. Several seroprevalence studies have suggested that the virus–host exposure has been taking place for many years in both human and rodent populations, and changes in human behavior and wild rodent ecology at the human–wildlife interface have facilitated the clinical recognition of the disease (Calisher et al. 2011).
Ultimately, collaborations around the hantavirus resulted in the development of an informal network called the Research Association of Medical and Biological Organizations (RAMBO). In the year following its formation, RAMBO generated a number of new integrated projects associated with hantavirus research. The legacy of RAMBO persists today, and researchers from the medical, biological, and other disciplines continue to engage with one another and generate new transdisciplinary research. The present article is, itself, the outcome of collaborative interaction by researchers interested in the geoepidemiology of zoonotic infectious disease, some of whom are original participants of RAMBO.
Was the hantavirus investigation an example of transformative science? The 1993 investigation did not launch new fields of science or new scientific paradigms, nor did the investigation radically change our understanding of existing scientific concepts, except perhaps for the new recognition that mammalian viral disease can be transmitted without an arthropod vector. Nevertheless, the investigative processes that occurred were a compelling example of the significant knowledge gains that can result from bridging disciplinary silos. It was the first instance in which molecular biology (nucleic acid amplification technology) was brought to the front line of an outbreak investigation, to identify the causative agent of the outbreak and the phylogenetic relationships in the bunyavirus family. It was the first instance in which forensic pathologists played a principal role in the detection, investigation, and response to a major infectious disease outbreak caused by a novel pathogen. It was the first time that evolutionary biology was used to anticipate geographically distant infectious disease outbreaks, by predicting hantavirus outbreaks in South America, observed in 1995. The collaborative assembly of links across quite distinct disciplines enabled the subsequent climate and ecological modeling of disease risk of HPS in humans, which has led to strategic public health interventions with a demonstrated positive societal impact.
By all accounts, this diverse group of scientists generated a creative transdisciplinary environment that has profoundly altered the research and perspectives of all involved. In addition, these efforts generated entirely new transdisciplinary, multiorganizational, and cross-cultural connections that persist today. From this perspective, the hantavirus investigation was certainly transformative. The notion of transformative science is multiscale, from individuals to society, with cross-scale connections. Transformative science at the scale of new paradigms and new fields of science necessarily depends on preceding transformations at the individual and local scales. In the hantavirus example, these kinds of transformations were stimulated by fertilization across research disciplines.
Although the details of each are unique, many stories of successful transdisciplinary teams begin with unlikely, unplanned, serendipitous encounters across disciplinary perspectives (Roberts 1989). However, the vision of transformative research assumes that such research can be sought after and enabled. Therefore, in the next section, we analyze the factors that contributed to the success of this group in generating innovative outcomes simultaneously across multiple scientific disciplines, organizational contexts, and societal settings.
Observations on the process of transdisciplinary collaboration
In the hantavirus example, an immediate crisis with great societal importance was clearly the motivating factor that overrode any other issues. Young, athletic people were quickly dying from an unknown agent, and the group was charged to respond. The dire nature of the outbreak facilitated the remarkable degree of collaboration across disciplines and organizations that might not have been achieved in other ways. Their respective organizations positioned group members to dedicate crucial time to the investigation. The benefit to society was clear. The costs of collaboration were reduced. The decision to participate was relatively straightforward. The group members had no initial perception that their findings would be transformative and would gain them additional funding and recognition, although these outcomes occurred.
In the absence of such a crisis, other motivations must prevail in order for researchers to collaborate across disciplines. Individuals may choose to participate in a research team because they believe that cooperation will lead to an enhanced ability to compete for funding, to generate findings that will have high impact within their own field, or to make a contribution to society (USNAS 2004). Those who have participated in successful interdisciplinary teams consistently report that the most compelling outcome is the intellectual stimulation and creativity generated by the collective group (Pennington 2011); indeed, this is the motivation reported by those who remained in the RAMBO group. In general, individuals make a conscious or unconscious comparison between the perceived potential benefits (e.g., funding, recognition, intellectual creativity) and the perceived costs (e.g., time, effort, resources) of participation. The outcome of the comparison is influenced by the individual's tolerance for risk and by past successes of the individuals involved. Engaging across disciplines is one of the most difficult activities in which a researcher can engage. If researchers do not believe the benefits to be high, they will choose to invest their efforts in other directions. Many factors affect this decision (Collins 2002, Kostoff 2002, Rhoten 2003).
Although the perceived benefits may be high and the costs low, researchers may still choose not to collaborate. Individual researchers must be able to envision how they can contribute to the collaboration in meaningful ways. Transdisciplinary research teams operate in what have been referred to as VUCA contexts—settings that are volatile, uncertain, complex, and ambiguous (McCausland and Martin 2001). Inter- and transdisciplinary research is volatile in that voluminous information may be available for any particular problem, but it is rarely precisely the right information, and new information is constantly appearing from many directions. Such research is uncertain because many scenarios of collaborative research could potentially be relevant, and it is unclear which scenarios would be the most beneficial to investigate. It is complex because an effective research team is an evolutionary product that leads to the emergence of shared objectives. And it is ambiguous because, until a team evolves into a cohesive group with shared objectives, it is not clear how the team members can collaborate in meaningful ways (Pennington 2008, 2011). The VUCA nature of the context of inter- and transdisciplinary research implies that the problem and solution are often ill defined, and it may be difficult for individual researchers to envision how they might contribute. Understanding that this is the case and that it takes time for shared conceptualizations of the problem to emerge is crucial to the success of new inter- and transdisciplinary research teams.
Foundational elements for transdisciplinary problem solving: Disciplines, data, and specimens.
A number of key elements were in place that enabled the serendipitous encounters that led to the novel findings of the hantavirus group. First, the investigators involved in the original 1993 outbreak had invested decades of effort becoming experts in their own fields. Without that deep knowledge, they could never have accomplished the listmaking activity that narrowed their focus to three possibilities. They could never have envisioned that short-term mammal research could answer key questions about the epidemic. Deep disciplinary knowledge is a fundamental building block of inter- and transdisciplinary research (Mansilla 2005). Although there has been much discussion in the literature about the issue of disciplinary silos, silos of knowledge are not necessarily bad; by nature, silos restrict communication among one another, and there is little guidance about how to engage more effectively (Alrøe and Noe 2010, Pennington 2011). It can be difficult to precisely identify which combination of silos is most appropriate for a given problem; indeed, effective combinations emerge and evolve as a result of interactions across disciplines around a given problem (Pennington 2011). Leadership in these situations requires managing the complex team and enabling emergence and evolution rather than controlling the team (Plowman et al. 2007).
Not only was individual disciplinary expertise and engagement across perspectives indispensable, but disciplinary and scientific networks were also crucial for rapidly transferring ideas, knowledge, data, and samples relevant to the investigation. The key components of disciplinary science are deep, cohesive, collective understanding and shared concepts that enable community construction and the vetting of knowledge. In contrast to the stereotype of the lone scientist, scientists are embedded in complex knowledge networks that give rise to disciplinary communities (Csikszentmihalyi 2007
One aim of transdisciplinary research is to get natural and social scientists to collaborate, so as to achieve an integrated view of a subject that goes beyond the viewpoints offered by any particular discipline. The question of how transdisciplinary approaches can be practised remains a challenge, however, if the quantitative and the qualitative sciences are both to be included. To explore this question, a series of qualitative interviews was conducted with researchers involved in two recent Swiss and Swedish research programmes. In both these programmes natural and social scientists had to collaborate in problem-driven environmental research. Three findings from these interviews are discussed in this paper: (a) that the researchers have more reasons to offer for non-collaboration than for collaboration, and that most of the thinking about transdisciplinary collaboration takes place at the level of programme management, (b) that the researchers should be classified as Detached Specialists or Engaged Problem Solvers rather than as natural and social scientists, and (c) that if collaboration evolves in a problem-driven research environment it tends to take the form of division of labour. The conclusion this paper draws for problem-driven research is that, paradoxically, the pressure to produce usable results should be reduced if collaboration is to emerge.