Transdisciplinary approaches enhance the

production of translational knowledge

TIMOTHY H. CIESIELSKI, MELINDA C. ALDRICH, CARMEN J. MARSIT1, ROBERT A. HIATT, and

SCOTT M. WILLIAMS2
HANOVER, NH; WESTON, MASS; NASHVILLE, TENN; ATLANTA, GA; SAN FRANCISCO, CALIF; AND CLEVELAND, OHIO

The primary goal of translational research is to generate and apply knowledge

that can improve human health. Although research conducted within the con-
fines of a single discipline has helped us to achieve this goal in many settings,
this unidisciplinary approach may not be optimal when disease causation is
complex and health decisions are pressing. To address these issues, we suggest
that transdisciplinary approaches can facilitate the progress of translational
research, and we review publications that demonstrate what these approaches
can look like. These examples serve to (1) demonstrate why transdisciplinary
research is useful, and (2) stimulate a conversation about how it can be further
promoted. While we note that open-minded communication is a prerequisite
for germinating any transdisciplinary work and that epidemiologists can play a
key role in promoting it, we do not propose a rigid protocol for conducting trans-
disciplinary research, as one really does not exist. These achievements were
developed in settings where typical disciplinary and institutional barriers were
surmountable, but they were not accomplished with a single predetermined
plan. The benefits of cross-disciplinary communication are hard to predict a pri-
ori and a detailed research protocol or process may impede the realization
of novel and important insights. Overall, these examples demonstrate that
enhanced cross-disciplinary information exchange can serve as a starting point
that helps researchers frame better questions, integrate more relevant evidence,
and advance translational knowledge more effectively. Specifically, we discuss
examples where transdisciplinary approaches are helping us to better explore,
assess, and intervene to improve human health. (Translational Research
2017;182:123–134)

Abbreviations: SNP ¼ single nucleotide polymorphism; MDR ¼ multifactor dimensionality

reduction; DAG ¼ directed acyclic graph; GWAS ¼ Genome Wide Association Study; DiCE ¼
diverse convergent evidence; L.E.A.D. ¼ Locate Evidence, Evaluate It, Assemble It, and Inform

1
Present address: Department of Environmental Health, Rollins Biostatistics, University of California San Francisco, San Francisco,
School of Public Health, Emory University, Atlanta, GA Calif.
2
Present address: Department of Epidemiology and Biostatistics, Case Submitted for publication March 3, 2016; revision submitted October
Western Reserve University School of Medicine, Cleveland, OH 12, 2016; accepted for publication November 2, 2016.
From the Institute for Quantitative Biomedical Sciences, Dartmouth Reprint requests: Timothy H. CIesielski, Public Health Program,
College, Hanover, NH; Department of Genetics, Geisel School of Regis College, 235 Wellesley Street, Weston, MA 02493; e-mail:
Medicine at Dartmouth, Hanover, NH; Public Health Program, timothyhciesielski@gmail.com.
Regis College, Weston, Mass; Department of Thoracic Surgery,
1931-5244/$ - see front matter
Vanderbilt University Medical Center, Nashville, Tenn; Division of
Epidemiology, Department of Medicine, Vanderbilt University Ó 2016 The Authors. Published by Elsevier Inc. This is an open access
Medical Center, Nashville, Tenn; Department of Pharmacology and article under the CC BY-NC-ND license (http://creativecommons.org/
Toxicology, Geisel School of Medicine at Dartmouth, Hanover, NH; licenses/by-nc-nd/4.0/).
Department of Epidemiology, Geisel School of Medicine at http://dx.doi.org/10.1016/j.trsl.2016.11.002
Dartmouth, Hanover, NH; Department of Epidemiology and

123
 Translational Research
124 Ciesielski et al April 2017

Decisions; IARC ¼ International Agency for Research on Cancer; NTP ¼ National Toxicology Pro-
gram; CEBS ¼ Chemical Effects in Biological Systems; IOM ¼ Institute of Medicine; CISNET ¼ Can-
cer Intervention and Surveillance Modeling Network; CSTA ¼ Clinical and Translational Science
Awards program

Glossary Complex Systems

Systems with multiple interacting components and
Multidisciplinary emergent properties that often cannot be accurately
The aggregation of fully formed ideas that come from characterized with narrow or rigid research frameworks
distinct fields
Marginal Association
Interdisciplinary The association between one exposure (factor) and
The integration, adaptation, and harmonization of one outcome (disease) independent of other variables.
ideas that come from distinct fields If potential biases and other observational data imper-
fections are properly accounted for, this association is
Transdisciplinary thought to provide evidence for or against the involve-
The generation and utilization of research frame- ment of the exposure with the disease.
works and admixed ideas that could not come from, or
fit into, any one field
INTRODUCTION
Cross-Disciplinary The primary goal of translational biomedical research
A general term referring to the unspecified involve- is to elucidate the determinants of disease and apply
ment of more than one discipline this knowledge to improve clinical or population
health practices. Epidemiologists have been success-
Perspective ful in advancing this goal, particularly in the context
Intellectual orientation or viewpoint that can vary in of conditions with causal factors that have consis-
its capacity to assess and adapt to external input tently detectable marginal associations. However, in
the context of etiologically heterogeneous complex
Strategy, Approach, Process, or Method disease, causal factors may not have reproducibly
A general code of conduct or way of proceeding that detectable marginal associations because these dis-
does not have a rigid, prespecified, or detailed sequence eases have multiple interacting determinants. As a
or parameters result, progress in this area has been much slower.
Here, we take the perspective of epidemiologists and
Protocol or Procedure hope to generate further discussion by exploring a
A specific code of conduct or way of proceeding that general approach for increasing our ability to address
has a rigid, prespecified, and detailed sequence and pa- multifactorial health problems. Specifically, we advo-
rameters cate that epidemiologists take a transdisciplinary
approach, and propose that enhancing the opportunity
Communication for cross-disciplinary information exchange can help
A general term referring to the exchange of informa- by making relevant perspectives from multiple
tion, strategies, protocols, hypotheses, or ideas (through distinct fields available for scientific reasoning at
talking, reading, graphical image presentation, etc.) each stage of the research process, but perhaps most
importantly at the outset of defining a problem and
Information designing a research strategy. This increases our chan-
Data and facts ces of realizing information synergies, thereby allow-
ing us to frame better questions, gather more
Knowledge comprehensive data, and better exploit existing infor-
Understanding of the relevant causal mechanisms that mation to guide health decisions.
generated the data and facts (note that information and This general approach addresses the key issues identi-
knowledge have similar meanings and are often used fied in two sets of recent commentaries concerning the
to define each other; however, here we emphasize that future of epidemiologic research. The first group of com-
knowledge implies an understanding of why the data mentaries proposes that innovative thinking will be cen-
or facts are as they are) tral to progress in epidemiology and translational
Translational Research
Volume 182 Ciesielski et al 125

research (particularly in the current age of big data),1-3 cepts in his discipline (miasma), (2) looked for patterns in
and the second group posits that integrated approaches new ways (his maps), and (3) worked with relevant people
can streamline the development of effective from outside his immediate field (Rev. Henry White-
interventions.4-7 Here, we extend this discussion by head)?10 Essentially, the utilization of information from
exploring specific theoretical issues and examples that multiple disciplines throughout the research process cre-
illustrate how cross-disciplinary information exchange ates a transdisciplinary approach11,12 that can extend
has provided novel insights into disease processes, led knowledge beyond the limitations of the contributing
to more complete knowledge of causation, and thereby disciplines. Transdisciplinary research does not refer to
spurred the development of more effective interventions. the combination of fully formed ideas from distinct fields
Thus, the overall purpose of this paper is to reveal the un- (multidisciplinary research), or the integration of ideas
derappreciated utility of transdisciplinary research and from distinct fields (interdisciplinary research), rather it
fuel discussion about how it can be fostered. refers to the generation and utilization of research
The examples provided here clarify how enhanced frameworks and ideas that could not come from, or fit
communication between fields can cultivate creative ap- into, any one field.11,13 This emergent property of
proaches that make translational research both more transdisciplinary translational research can enable us to:
effective and efficient. They also emphasize the often un- (1) explore widely, (2) assess diversely, and (3) intervene
derappreciated advantages of teamwork.8 These exam- effectively in our efforts to promote human health. These
ples do not argue for the development of a single 3 areas provide the framework for our discussion below.
pipeline for the conduct of transdisciplinary research or
even that it is possible to know a priori how to design TRANSDISCIPLINARY PERSPECTIVES CAN HELP US
ensemble research strategies for all contexts. We also GENERATE BETTER HYPOTHESES
recognize that transdisciplinary research often involves A key to getting better answers is asking better ques-
difficult challenges,9 and it cannot be forced. However, tions, and transdisciplinary perspectives can generate
we propose that we can spur the development of produc- hypotheses that unidisciplinary perspectives might
tive transdisciplinary approaches if we create research otherwise miss. If we utilize a fuller set of scientific per-
and training environments that encourage cross- spectives, and tools from more than one discipline, we
disciplinary information exchange. By extension, failing can frame critical questions that are not apparent from
to take measures to increase our cross-disciplinary the data and tools of a single discipline. On the other
fluency will likely impede, or even prevent, the develop- hand, if we exclusively use canonical exploratory
ment of solutions to many human health problems. Stated methods, we will define relatively simple questions
differently, traditional silo-based research can only that will likely fail to identify many of the complex phe-
address a limited number of incomplete questions. nomena that lead to health problems.
As the examples here demonstrate, open-minded
cross-disciplinary communication can yield useful but Example: exploratory research in a single discipline
unpredictable results. Before researchers begin talking often fails in the context of complex disease.
to people from other fields (or at least start reading their
‘‘.the problem of identifying and quantifying
papers), it will often not be clear how different disci-
multiple component causes of disease is one of
plines may help each other address a given problem.
the most basic limitations in modern epidemi-
In addition, it will usually not be clear beforehand if
ology’’
the help will come in framing questions (the beginning),
Paolo Vineis and David Kriebel14
analytic methods (the middle), or information integra-
tion and intervention (the end). This communication In population health and medicine, we often employ
could result in the cross-disciplinary transfer of a single simple descriptive epidemiology techniques to generate
critical piece of information or it could generate a long- hypotheses about the causes of illness and disease15
term symbiotic relationship between researchers. The (e.g. univariate analyses from surveillance data, fre-
nature of this communication and character of its bene- quency tables, and histograms). Analytic epidemiology
fits are inherently unpredictable. Therefore, we will not techniques can then be used to test these hypotheses,
unnecessarily constrain this process by proposing the correcting for potential biases.15 This two-step process
use of rigid constructs or specific protocols. is foundational to epidemiology, but its effectiveness
Unfortunately, the utility of cross-disciplinary commu- can be limited when the first step is insufficiently infor-
nication is frequently overlooked although it has already mative. Stated differently, if the relevant causal factors
produced significant advances. For example, how long do not have detectable marginal associations, then stan-
would it have taken to reduce cholera transmission if dard descriptive epidemiology techniques may not
John Snow had not: (1) thought beyond some existing con- effectively direct our subsequent efforts.
 Translational Research
126 Ciesielski et al April 2017

Fig 1. Visualization techniques reduce jargon-based barriers to cross-disciplinary communication. Visual infor-
mation summaries can stimulate productive conversation in transdisciplinary teams by helping researchers to
reason with relevant factors that are beyond their individual disciplinary expertise. In this example, the putative
causal factors in the etiology of a disease are summarized in a jargon-free visual schematic. Discussing this sche-
matic allows the team to access additional relevant information from the collaborators (i.e. factor 7 may provide an
intervention opportunity). In this way, the team identified a previously unrecognized key modifiable factor even
though no individual had enough information to think of it on their own.

Most epidemiology training focuses on analytic the phenotype of interest.17,18 This method was
epidemiology and encourages the evaluation of effect developed by geneticists to detect gene-gene and gene-
modification (interaction) only if it is suspected a priori. environment interactions that are associated with a
This rule of thumb makes sense in the context of stan- phenotype, and it has demonstrated great utility in this
dard statistical models, because with these methods, role.17-26 Thus, this method, perhaps in combination
screening for all possible interactions is at best problem- with other machine learning techniques,27,28 can be
atic and at worst impossible. Therefore, if a critical used as a tool to identify multifactor patterns associated
cofactor is not suspected to be an effect modifier, it is with disease. For example, researchers have used MDR
not usually addressed in standard epidemiologic inves- to identify putative gene-environment interactions in the
tigations. The ability of a cofactor to have important ef- development of lung cancer (predictive single nucleotide
fects, however, is not contingent on our ability to polymorphisms [SNPs] differ by smoking status),29 and
suspect its role a priori. Therefore, we need to develop childhood asthma (several SNPs interact with indoor
exploratory methods to identify putative component dampness).30 Additional strategies for detecting multi-
causes16 whose etiologic role is not evident from mar- factor patterns associated with disease continue to be
ginal associations. A variety of potentially useful new developed in genetics,31-33 and techniques such as these
methods can be found in fields where techniques for can extend our capacity to identify combinations of
analogous problems have already been developed (e.g. factors that are linked to disease risk.
genetics, computer science, economics, and ecology). Another promising set of approaches from computa-
Such methods can advance complex disease research tional biology leverages visual data representations to
by expanding descriptive epidemiology beyond histo- help translate complex patterns into specific etiologic
grams and correlations, to include methods capable of questions.34,35 Visual methods may be of particular
generating novel multifactorial hypotheses. relevance in the development of transdisciplinary
An example of one such tool is multifactor dimension- discovery epidemiology because they reduce jargon-
ality reduction (MDR), a machine learning method that based barriers to cross-disciplinary communication
explores all possible combinations of categorical vari- (Fig 1), by replacing field-specific terminology with
ables to identify combinations that best associate with broadly-accessible visual aids. Additional nonstandard
Translational Research
Volume 182 Ciesielski et al 127

Table I. A transdisciplinary advance in analytic TRANSDISCIPLINARY PERSPECTIVES CAN HELP US

epidemiology: directed acyclic graphs BETTER ASSESS AND INTERPRET EVIDENCE

Directed acyclic graphs (DAGs) adapted from computer science Transdisciplinary perspectives can help us to gather
have: and utilize more relevant and comprehensive evidence
1. Helped us to better identify adjustments that introduce rather for vetting putative etiologic factors. This allows us to
than reduce bias44 better address our concerns about potentially
2. Provided a general analytic framework that can explain the
‘‘birthweight paradox,’’ and backs up our common sense
misleading findings.47 By including diverse types of
notion that trials of prenatal smoking to reduce infant mortality data and encouraging multiple modes of assessment,
are not a good idea45 transdisciplinary perspectives can rigorously evaluate
the strength of evidence supporting a given hypothesis.
computational tools, such as agent-based models and The process of including more and diverse approaches
other complex systems models, may be useful for that are encompassed by a transdisciplinary paradigm
learning about multifactor causes and system properties can result in a detailed understanding of the current un-
when disease risk is modulated by multiple nonlinear in- certainties and thus clarify for decision makers which
teractions that vary temporally.36-38 Furthermore, courses of action (or inaction) are most reasonable in
establishing transdisciplinary teams can promote light of the existing knowledge. It can also clarify for re-
communication between diverse subject matter experts searchers what additional evidence is most needed to
to advance the development of new complex systems advance our understanding of disease etiology and po-
models.39 These models can allow researchers to think tential interventions. In short, transdisciplinary ap-
about the relationships between putative causal factors proaches can improve our decision making by
from multiple fields in novel ways. Such methods create enhancing our ability to reason with imperfect and
a unique opportunity to develop the multilevel hypothe- incomplete evidence from many sources.
ses that are needed to address complex health problems. Example: transdisciplinary information allows for diverse
Of course, thoughtful trials and discussion are required convergent validation of findings. Genome Wide Associ-
to better clarify the strengths and weaknesses of these ation Studies (GWAS) are expensive endeavors and re-
nontraditional approaches. This process is already under- searchers would like to increase the usable knowledge
way for MDR23-28 and agent-based models.40-43 Because gained from these studies to learn more about disease eti-
nontraditional discovery epidemiology methods (e.g. ology and intervention options. It is becoming recognized
MDR) are prone to bias, as are all descriptive that, when utilized in isolation, GWAS analyses have a va-
epidemiology methods (e.g. unadjusted associations), riety of weaknesses that can hinder the discovery of ge-
validation with traditional analytic epidemiology netic risk factors.48-54 Essentially GWAS, like all
models is important. Furthermore, even analytic epidemiologic analyses, are prone to both type-1 and
epidemiology approaches that properly account for type-2 error, as well as the influence of unrecognized
known confounding and biases are limited in their biases. However, if GWAS data is systematically
ability to infer causality. Therefore, experimental, evaluated in the context of relevant evidence from
biological, and implementation research strategies will diverse areas, it can be part of a larger process that more
continue to be crucial for validating and characterizing effectively discovers and vets genetic risk factors for
the causal relationships suggested by any observed disease. For example, DiCE (Diverse Convergent
statistical associations. Evidence) is an evidence integration process that
We note that it is not surprising that transdisciplinary combines information from observational association
approaches can generate advances in descriptive epide- studies, bioinformatics, and laboratory experiments to
miology because in recent years techniques from other yield a metric that reflects the likelihood that a given
fields have enhanced the practice of analytic epidemi- genetic factor is involved in the disease
ology. In particular, directed acyclic graphs (DAGs) pathophysiology.54 As a proof of principle, this metric
from computer science have advanced our ability to identified the role of Hemoglobin S in severe malaria
communicate causal structures and identify bias, thus resistance54-56 and the role of PPAR-gamma in type 2
allowing us to build better analytic epidemiology diabetes54,57-59 when standard GWAS validation criteria
models (Table I).46 Here, we simply note that methods alone failed to detect these etiologically relevant
from other fields might help us advance descriptive factors. DiCE can also highlight potential false-positive
epidemiology as well. Importantly, if we fail to utilize findings in GWAS analyses, including those that reach
new pattern finding algorithms for discovery epidemi- canonical thresholds for statistical significance, and
ology, we will likely miss opportunities to identify suggest future research to address the ambiguous
modifiable component causes of disease. evidence. Overall, DiCE allows more diverse evidence
 Translational Research
128 Ciesielski et al April 2017

Table II. Cochrane Reviews63 can convert literature that is unwieldy and inaccessible into evidence that is
widely accessible and relevant to decision makers*

Many doctors and patients may ask:

Can vitamin C supplements prevent or treat the common cold (viral respiratory infections)?

Potential utility Unclear utility Feasibility Safety

Strong evidence The evidence is Inexpensive Thought to be without

suggests that regular inconclusive as to adverse effects.
vitamin C whether vitamin C can
supplementation can prevent the common
reduce the duration cold or reduce
and severity of symptoms if it is
common colds that started after cold
occur onset

*Information extracted from a scientific abstract and plain language summary that are freely available (in several languages) at the Cochrane
website.64

to enter the process of determining what leads to follow its utility in decision making.62 By promoting input
and how to follow them. This promotes well-informed from a broad array of sources, the available evidence
decisions and faster knowledge acquisition. and its quality have been considered and organized to
address the types of questions that physicians and
Example: transdisciplinary approaches allow us to work patients ask. For example, this approach has translated
with disparate inconclusive evidence. How can we even the complex literature on vitamin C and the common
begin to ameliorate a problem with as many potential cold into actionable information (Table II).
causes as the obesity epidemic? Again, a broad perspective
and systematic information integration can be useful. In Example: transdisciplinary approaches can clarify
research and intervention priorities. How can we properly
2010, the Institute of Medicine developed a framework
to promote this type of translational approach for allocate limited research, remediation, and policy efforts
combating obesity: the IOM L.E.A.D. framework.60 And to the environmental chemicals that pose the greatest risk
recently, Chatterji et al. discussed the application of this to human health? This situation represents another
type of approach in New York City’s policy decisions instance where there is incomplete and nondefinitive in-
regarding fat and calorie information for restaurant formation and a need to advance knowledge quickly to
food.61 In this framework, researchers: A) Locate minimize human health problems. One effective strategy
Evidence, B) Evaluate It, C) Assemble It, and D) Inform can be found in the IARC monographs.65 In this
Decisions. The structure for this translation process approach, multidisciplinary IARC working groups are
reflects our need to make policy decisions when there are assembled to discuss 4 aspects of a given exposure: (1)
many disparate pieces of relevant but inconclusive the potential for human exposure, (2) the evidence for
evidence. Imperfect evidence may only provide association with cancer in humans, (3) the evidence for
decipherable guidance in our research and intervention causation of cancer in animals, and (4) relevant
decisions, if it is considered in its totality. Narrow mechanistic/toxicokinetic evidence. This broad scope
assessments of the evidence from one field could prove of information is then converted by a cross-disciplinary
misleading or just plain false. Thus, the structure and consensus building process into a carcinogenicity
purpose of L.E.A.D. is analogous to DiCE, although it is assessment (e.g. probably not carcinogenic, not
more directly focused on implementation. classifiable, possibly carcinogenic, probably
carcinogenic, and carcinogenic). Recently, the National
Example: transdisciplinary teams can facilitate Toxicology Program (NTP) developed a similar general
information synthesis and decision making. Physicians strategy for integrating human, animal, and in vitro
need to quickly learn what the research evidence sug- evidence in chemical assessments.66 The NTP also
gests should be done for their patients. If that volumi- makes relevant evidence available for alternative
nous information is not comprehensively and cogently integration analyses by compiling it into publically
distilled in an ethical manner, then physicians cannot accessible databases (e.g. CEBS, DrugMatrix, and
effectively use it to guide patient care decisions. For ToxFX).67 Overall, these transdisciplinary approaches
over 20 years, the nonprofit Cochrane Collaboration can highlight the largest potential problems based on
has been using a network of diverse working groups to the available evidence and simply convey this
synthesize medical research information and increase information to both researchers and decision makers.
Translational Research
Volume 182 Ciesielski et al 129

These examples indicate that more comprehensive tors.71 The effect of each may depend on the context
and integrated information can allow for improved vali- defined by the other, and thus some causal factors
dation of potential risk factors and enhanced character- may not be detected by looking only at their marginal
ization of health problems. The long-standing associations (note that here we are referring to
transdisciplinary approaches (e.g. IARC monographs environmental factors in the broadest sense:
and Cochrane Collaboration) provide evidence that xenobiotic exposures, social/psychosocial factors,
these strategies are very useful and the newer techniques nutrition, etc.). Even when a marginal association is
(e.g. DiCE, LEAD, ToxFX) demonstrate that these ap- detectable, broader consideration of genetic and
proaches can be further optimized for efficiency. environmental variables taken together can illuminate
the mechanisms that create this marginal association,
TRANSDISCIPLINARY PERSPECTIVES CAN HELP US and provide information about etiologic subtypes that
DEVELOP AND COORDINATE INTERVENTIONS may benefit from distinct interventions.72,73
Furthermore, finding an isolated genetic cause of
Discipline-specific strategies can limit our ability to
develop policies and interventions that improve human disease may not suggest obvious interventions but if a
genetic factor is found to interact with a modifiable
health. For example, imagine our goal was to fill a barrel
environmental factor, then knowledge of the
with water and keep it filled. If the barrel has a hole in it,
environmental factor can create a prevention or
how would we best coordinate the efforts of a cooper
treatment opportunity. For example: children with
with that of a person getting water from a well? No mat-
genetic disruptions of phenylalanine hydroxylase
ter how hard each individual works, their efforts will be
function (phenylketonuria) can avoid many adverse
inefficient or even ineffective unless they are applied in
health consequences by eating phenylalanine-limited
the right order. The development of effective interven-
tions can suffer from similar issues. diets that would not be optimal for other children.74-76

Example: diverse perspectives allow us to handle

Example: interventions can be more effective when
etiologies that change in response to intervention. How
etiologic characterization is transdisciplinary. Chemical
can we design interventions that promote stable posi-
exposures, social exposures (neighborhood and family
interaction styles), genetics, educational strategies, and tive changes in complex dynamic systems when the
effect of the same action may vary temporally? RCTs
nutrients can individually be evaluated for their associa-
and experiments are ideal for learning about single
tion with neurodevelopmental outcomes. These variables
factors in systems that you can randomize and control,
may be studied separately by environmental epidemiolo-
but are less useful when multiple dynamic nonlinear in-
gists, social epidemiologists, geneticists, developmental
teractions are modulating disease risk. Some of the
pediatricians, neuropsychologists, and nutritional epide-
computational tools mentioned above (e.g. agent-
miologists, but if they work in isolation, information
based modeling) may be useful for learning about
about how to most effectively intervene is likely to be
obscured. How can you characterize the relationship be- how complex systems react to interventions, and thus
they may also be useful for developing strategies that
tween lead exposure and adverse neurodevelopmental
have consistently positive impacts.36,37 These
outcomes without considering how psychosocial factors
methods can allow us to ask important novel
may generate potential confounding and other biases?68
questions. Would a combination of interventions
Also, what good is an educational intervention if the
work well? Do certain policies only have a high
child is still exposed to lead because the home is not
probability of success in specific contexts? Can
properly remediated or the source of exposure remains
multifactorial interventions or contingency algorithms
unidentified? Educational, social, medical, and
environmental interventions can fail or they can be generate better outcomes in these settings? Along
these lines, the Cancer Intervention and Surveillance
synergistic. Understanding the relationships between
Modeling Network (CISNET)77 has developed a
component causes from a variety of traditionally
simulation process that leverages clear modeling
separate fields can clarify the overall public health
assumptions and comparison of results from multiple
problem and intervention possibilities, and this
simulations to acquire convergent evidence that
principle has become a driving factor behind the
highlights putative causal factors. Importantly,
emerging concept of ‘‘exposome’’ research.69,70
whatever is learned from complex intervention
simulations and the careful observation of new
Example: moving from genes or environment, past
genes and environment, to genes with interventions, can be fed back into an evolving
environment. Studying genetic and environmental fac- knowledge base for guiding future research and
tors together can help us to avoid missing causal fac- interventions.78,79 Overall, broad transdisciplinary
 Translational Research
130 Ciesielski et al April 2017

Fig 2. Transdisciplinary approaches coordinate evidence to generate more useful knowledge. Cross-disciplinary
cooperation allows us to see how information from multiple sources can fit together to build our understanding of
health issues. ‘‘Science is built up with facts, as a house is with stones. But a collection of facts is no more a science
than a heap of stones is a house.’’ – Jules Henri Poincare,80 p. 127.
Translational Research
Volume 182 Ciesielski et al 131

Table III. Simple ways that epidemiologists can promote transdisciplinary translational research*
A Incorporate more non-epidemiology concepts and knowledge into epidemiology training
1) Accept students and postdocs who have diverse prior training outside of epidemiology
2) Include non-epidemiology experiences in doctoral training
(e.g. laboratory, clinical, or policy rotations)
3) Encourage epidemiology students to do postdoctoral training in complementary areas
(e.g. physiology, demography, public policy, computer science, regulatory agencies, etc.).
B Diversify traditional epidemiology working environments
1) Present and discuss epidemiologic research at disease-specific conferences.
2) Hire individuals who trained in more than one area, or have a unique background outside of epidemiology (these individuals will share
an overlapping vocabulary with people from a separate discipline thus expanding the ‘‘fluency’’ of the epidemiology department).
3) Explain and demonstrate the value-added of including epidemiologists in translational team science to the wider scientific
community.78,81
4) Invite basic researchers, clinicians, and policy experts to speak in epidemiology departments.
5) Develop proactive outreach mechanisms for embedding epidemiologists in clinical and basic science departments to promote
collaboration.
6) Create a multidisciplinary seminar/discussion series to promote information integration and collaboration by focusing on a specific
health issue at each meeting and advertising widely.
7) Offer small pilot funds for projects that access/integrate information from other fields to promote the development of proposals for
extramural funding.
8) Read papers from other disciplines that analyze large complex datasets to better harness diverse perspectives for crucial insights.92

*It is possible that structural changes in research institutions and funding sources might further promote transdisciplinary thinking and the success
of team science oriented researchers, but the small steps listed here are achievable in the near term and capable of informing potential next
steps. In addition, the epidemiologists that experiment with these small steps could serve as key resources in the development of large-scale
transdisciplinary efforts such as the NIH’s Clinical and Translational Science Awards program.93 Beyond exploring the specific actions proposed
here, the most important thing that we all can do to contribute to the conduct of effective transdisciplinary research is to ‘‘hold our knowledge
lightly’’91 and promote a culture of open-mindedness. This receptive yet objective perspective is the oil for transdisciplinary engines. Essentially,
it allows for discussions that illuminate crucial information from many disciplines to generate sound and comprehensive reasoning.

approaches are essential to better coordinate both our are demonstrating that when molecular biologists,
knowledge and efforts (Fig 2). pathologists, and epidemiologists collaborate, they can
evaluate molecular factors in new ways that permit the
EPIDEMIOLOGISTS CAN PLAY A KEY ROLE IN identification of etiologic subgroups and the
ADVANCING TRANSDISCIPLINARY APPROACHES physiologic mechanisms of disease.
Epidemiologists are well positioned to facilitate As a group, epidemiologists can further advance this
transdisciplinary translational research78,81 because approach by creating working environments that are
good epidemiology training provides a familiarity more open to (and capable of) cross-disciplinary con-
with a broad range of causal factors, and makes versation at all stages of research. This allows for better
practitioners aware of the ‘‘big picture’’. In fact, many integration and application of existing relevant informa-
epidemiologists are already at the forefront when it tion that can lead to more complete and useful knowl-
comes to advancing transdisciplinary research, and a edge,91 and also promotes the more efficient
variety of transdisciplines that depend on study design acquisition of new relevant information. In Table III,
and analytic principles from epidemiology have we list specific feasible strategies that can stimulate
already emerged. For example, epidemiologists are transdisciplinary thinking and create opportunities for
working with social scientists to understand the social intellectual crosspollination through better channels of
determinants of health (social epidemiology),82 and ep- communication.
idemiologists are cooperating with toxicologists to
identify chemical etiologic factors (environmental
epidemiology).83 These fields have even been further CONCLUSION
combined to allow transdisciplinary insights to flow In this article, we cite examples which demonstrate
from the consideration of social, ecological, and biolog- that transdisciplinary approaches can cultivate and vet
ical factors in infectious disease epidemiology.84,85 useful new strategies for dealing with complex health
These are just a few examples but we emphasize that challenges. These ensemble science methods can
epidemiology continues to spur synergy in new ways. develop whenever we make tangible efforts to improve
Among the newest epidemiology-based transdisciplines cross-disciplinary information exchange, and they
are epigenetic epidemiology86-88 and molecular allow us to streamline the development of effective
pathological epidemiology.89,90 These transdisciplines health interventions. Transdisciplinary approaches can
 Translational Research
132 Ciesielski et al April 2017

be as sophisticated as an international team of special- T. H. Ciesielski proposed the idea of writing an article
ists working together in a coordinated fashion, or they providing tangible examples of innovative methods in
can be as simple as talking more often with people epidemiology. The general format of the piece was
who have distinct training. The examples presented determined in a conference call with all authors. He
here are not intended to provide a blueprint for con- drafted the manuscript and all authors provided key
ducting transdisciplinary work. Instead, they serve to feedback and input (R. A. Hiatt, M. C. Aldrich, C. J.
(1) demonstrate what transdisciplinary insights can Marsit, and S. M. Williams). S. M. Williams is the se-
look like and (2) show that these insights can advance nior author and provided substantial editorial input.
translational research. Overall, we have observed that
being open to cross-disciplinary information exchange
is the defining feature of transdisciplinary translational REFERENCES
research, and it has tripartite utility. It helps us to 1. Hiatt RA, Sulsky S, Aldrich MC, et al. Promoting innovation and
explore widely, assess diversely, and intervene effec- creativity in epidemiology for the 21st century. Ann Epidemiol
tively in complex systems. 2013;23:452–4.
This review does not suggest that transdisciplinary 2. McKeown RE. Is epidemiology correcting its vision problem? A
collaboration is a panacea for health research challenges. perspective on our perspective: 2012 presidential address for Amer-
ican College of Epidemiology. Ann Epidemiol 2013;23:603–7.
However, it does suggest that transdisciplinary collabora- 3. Ness RB. Tools for innovative thinking in epidemiology. Am J Ep-
tion can help in some situations, and failing to enhance idemiol 2012;175:733–8.
cross-disciplinary communication and subsequent 4. Is Integrative Cancer Epidemiology the Next-Gen of Epidemi-
research approaches may slow down our progress. That ology?: Epidemiology and Genomics Research Program of the
National Cancer Institute’s Division of Cancer Control and Popu-
said, many barriers to the conduct of true transdisci-
lation Sciences, http://blog-epi.grants.cancer.gov/2013/08/13/
plinary/team science are ingrained in our research infra- integrative-cancer-epidemiology/; 2014. Accessed June 27, 2014.
structure, including promotion criteria, funding, training 5. Khoury MJ, Lam TK, Ioannidis JP, et al. Transforming epidemi-
strategies, and field-specific jargon,13,94-98 and these ology for 21st century medicine and public health. Cancer Epide-
barriers must be overcome if we are to realize the miol Biomarkers Prev 2013;22:508–16.
benefits of transdisciplinary research. Importantly, the 6. Spitz MR, Caporaso NE, Sellers TA. Integrative cancer epidemi-
ology–the next generation. Cancer Discov 2012;2:1087–90.
limitations of unidisciplinary research may not always 7. Lam TK, Spitz M, Schully SD, et al. ‘‘Drivers’’ of translational
be apparent, but if we utilize transdisciplinary cancer epidemiology in the 21st century: needs and opportunities.
approaches, those shortcomings are clarified and our Cancer Epidemiol Biomarkers Prev 2013;22:181–8.
ability to improve human health is enhanced. We 8. Wuchty S, Jones BF, Uzzi B. The increasing dominance of teams
cannot answer questions that we do not ask and we in production of knowledge. Science 2007;316:1036–9.
9. Lynch J. It’s not easy being interdisciplinary. Int J Epidemiol
cannot guide our actions with evidence that we do not 2006;35:1119–22.
assess, but in these tasks transdisciplinary perspectives 10. Frerichs RR. John Snow, http://www.ph.ucla.edu/epi/snow.
can expand the capabilities of epidemiologists and the html#PUMP%20HANDLE; 2014. Accessed June 23, 2014.
translational research community. 11. Rosenfield PL. The potential of transdisciplinary research for sus-
taining and extending linkages between the health and social sci-
ences. Social Sci Med 1992;35:1343–57.
12. Stokols D, Hall KL, Taylor BK, et al. The science of team science:
ACKNOWLEDGMENTS
overview of the field and introduction to the supplement. Am J
Conflicts of Interest: All authors have read the jour- Prev Med 2008;35:S77–89.
nal’s authorship agreement and the journal’s policy on 13. Hall KL, Vogel AL, Stipelman B, et al. A four-phase model of
disclosure of potential conflicts and have none to transdisciplinary team-based research: Goals, team processes,
and strategies. Translational Behav Med 2012;2:415–30.
declare. 14. Vineis P, Kriebel D. Causal models in epidemiology: past inheri-
This work was supported by the National Institute of tance and genetic future. Environ Health 2006;5:21.
Mental Health at the National Institutes of Health (NIH- 15. Hennekens CH, Buring JE. Chapter 2 design strategies in epide-
NIMH R01MH094609; UO1ES019457); the National miologic research. In: Mayrent SL, ed. Epidemiology in Medicine,
1. Philadelphia, PA: Lippincott Williams & Wilkins, 1987:16–29.
Institute of Environmental Health Sciences at the
16. Rothman KJ. Causes. Am J Epidemiol 1976;104:587–92.
National Institutes of Health (NIH-NIEHS 17. Ritchie MD, Hahn LW, Roodi N, et al. Multifactor-dimensionality
R01ES022223); March of Dimes Ohio Collaborative reduction reveals high-order interactions among estrogen-
for the Prevention of Preterm Birth, the National Insti- metabolism genes in sporadic breast cancer. Am J Hum Genet
tutes of Health (NIH P20 GM103534), and the National 2001;69:138–47.
Cancer Institute at the National Institutes of Health 18. Moore JH. 2014. Available at: http://www.epistasis.org/index-3.
html. Accessed June 25, 2014.
(K07 CA172294) to [M. C. Aldrich]. The funders had 19. Moore JH. Computational analysis of gene-gene interactions us-
no involvement in the writing of the manuscript or the ing multifactor dimensionality reduction. Expert Rev Mol Diagn
decision to submit it for publication. 2004;4:795–803.
Translational Research
Volume 182 Ciesielski et al 133

20. Park HW, Shin ES, Lee JE, et al. Multilocus analysis of atopy in 40. Marshall BD, Galea S. Formalizing the role of agent-based
Korean children using multifactor-dimensionality reduction. Tho- modeling in causal inference and epidemiology. Am J Epidemiol
rax 2007;62:265–9. 2015;181:92–9.
21. Collins RL, Hu T, Wejse C, et al. Multifactor dimensionality reduc- 41. Hernan MA. Invited commentary: agent-based models for causal
tion reveals a three-locus epistatic interaction associated with sus- inference-reweighting data and theory in epidemiology. Am J Ep-
ceptibility to pulmonary tuberculosis. BioData Min 2013;6:4. idemiol 2015;181:103–5.
22. Afzal S, Gusella M, Jensen SA, et al. The association of polymor- 42. Diez Roux AV. Invited commentary: the virtual epidemiologist-
phisms in 5-fluorouracil metabolism genes with outcome in adju- promise and peril. Am J Epidemiol 2015;181:100–2.
vant treatment of colorectal cancer. Pharmacogenomics 2011;12: 43. Marshall BD, Galea S. Marshall and Galea respond to ‘‘data the-
1257–67. ory in epidemiology’’. Am J Epidemiol 2015;181:106–7.
23. Bastone L, Reilly M, Rader DJ, et al. MDR and PRP: a compari- 44. Howards PP, Schisterman EF, Poole C, et al. ‘‘Toward a clearer
son of methods for high-order genotype-phenotype associations. definition of confounding’’ revisited with directed acyclic graphs.
Hum Hered 2004;58:82–92. Am J Epidemiol 2012;176:506–11.
24. Ritchie MD, Hahn LW, Moore JH. Power of multifactor dimen- 45. Hernandez-Diaz S, Schisterman EF, Hernan MA. The birth weight
sionality reduction for detecting gene-gene interactions in the ‘‘paradox’’ uncovered? Am J Epidemiol 2006;164:1115–20.
presence of genotyping error, missing data, phenocopy, and ge- 46. Greenland S, Pearl J, Robins JM. Causal diagrams for epidemio-
netic heterogeneity. Genet Epidemiol 2003;24:150–7. logic research. Epidemiology 1999;10:37–48.
25. Gui J, Moore JH, Williams SM, et al. A simple and computation- 47. Khoury MJ, Ioannidis JP. Medicine. Big data meets public health.
ally efficient approach to multifactor dimensionality reduction Science 2014;346:1054–5.
analysis of gene-gene interactions for quantitative traits. PLoS 48. Kraft P. Curses–winner’s and otherwise–in genetic epidemiology.
One 2013;8:e66545. Epidemiology 2008;19:649–51. discussion 57–8.
26. Lou XY, Chen GB, Yan L, et al. A generalized combinatorial 49. Kraft P, Zeggini E, Ioannidis JP. Replication in genome-wide as-
approach for detecting gene-by-gene and gene-by-environment sociation studies. Stat Sci 2009;24:561–73.
interactions with application to nicotine dependence. Am J Hum 50. Vineis P, Brennan P, Canzian F, et al. Expectations and challenges
Genet 2007;80:1125–37. stemming from genome-wide association studies. Mutagenesis
27. Heidema AG, Feskens EJ, Doevendans PA, et al. Analysis of mul- 2008;23:439–44.
tiple SNPs in genetic association studies: comparison of three 51. Williams SM, Canter JA, Crawford DC, et al. Problems with
multi-locus methods to prioritize and select SNPs. Genet Epide- genome-wide association studies. Science 2007;316:1840–2.
miol 2007;31:910–21. 52. Greene CS, Penrod NM, Williams SM, et al. Failure to replicate a
28. Andrew AS, Karagas MR, Nelson HH, et al. DNA repair polymor- genetic association may provide important clues about genetic ar-
phisms modify bladder cancer risk: a multi-factor analytic strat- chitecture. PLoS One 2009;4:e5639.
egy. Hum Hered 2008;65:105–18. 53. Williams SM, Haines JL. Correcting away the hidden heritability.
29. Ihsan R, Chauhan PS, Mishra AK, et al. Multiple analytical ap- Ann Hum Genet 2011;75:348–50.
proaches reveal distinct gene-environment interactions in smokers 54. Ciesielski TH, Pendergrass SA, White MJ, et al. Diverse conver-
and non smokers in lung cancer. PLoS One 2011;6:e29431. gent evidence in the genetic analysis of complex disease: coordi-
30. Su MW, Tung KY, Liang PH, et al. Gene-gene and gene- nating omic, informatic, and experimental evidence to better
environmental interactions of childhood asthma: a multifactor identify and validate risk factors. BioData Min 2014;7:10.
dimension reduction approach. PLoS One 2012;7:e30694. 55. Jallow M, Teo YY, Small KS, et al. Genome-wide and fine-
31. Hu T, Sinnott-Armstrong NA, Kiralis JW, et al. Characterizing ge- resolution association analysis of malaria in West Africa. Nat
netic interactions in human disease association studies using sta- Genet 2009;41:657–65.
tistical epistasis networks. BMC Bioinformatics 2011;12:364. 56. Timmann C, Thye T, Vens M, et al. Genome-wide association
32. Hu T, Chen Y, Kiralis JW, et al. An information-gain approach to study indicates two novel resistance loci for severe malaria. Na-
detecting three-way epistatic interactions in genetic association ture 2012;489:443–6.
studies. J Am Med Inform Assoc 2013;20:630–6. 57. Saxena R, Voight BF, Lyssenko V, et al. Genome-wide association
33. Yee J, Kwon MS, Park T, et al. A modified entropy-based analysis identifies loci for type 2 diabetes and triglyceride levels.
approach for identifying gene-gene interactions in case-control Science 2007;316:1331–6.
study. PLoS One 2013;8:e69321. 58. Scott LJ, Mohlke KL, Bonnycastle LL, et al. A genome-wide as-
34. Moore JH, Lari RC, Hill D, et al. Human microbiome visualiza- sociation study of type 2 diabetes in Finns detects multiple suscep-
tion using 3D technology. Pac Symp Biocomput 2011;154–64. tibility variants. Science 2007;316:1341–5.
35. Hu T, Chen Y, Kiralis JW, et al. ViSEN: methodology and soft- 59. Zeggini E, Weedon MN, Lindgren CM, et al. Replication of
ware for visualization of statistical epistasis networks. Genet Epi- genome-wide association signals in UK samples reveals risk
demiol 2013;37:283–5. loci for type 2 diabetes. Science 2007;316:1336–41.
36. Galea S, Riddle M, Kaplan GA. Causal thinking and complex sys- 60. IOM. Bridging the Evidence Gap in Obesity Prevention: A
tem approaches in epidemiology. Int J Epidemiol 2010;39: Framework to Inform Decision Making. Washington DC: Na-
97–106. tional Academy of Sciences, 2010.
37. Auchincloss AH, Diez Roux AV. A new tool for epidemiology: the 61. Chatterji M, Green LW, Kumanyika S. L.E.A.D.: a framework for
usefulness of dynamic-agent models in understanding place ef- evidence gathering and use for the prevention of obesity and other
fects on health. Am J Epidemiol 2008;168:1–8. complex public health problems. Health Educ Behav 2014;41:
38. Diez Roux AV. Integrating social and biologic factors in health 85–99.
research: a systems view. Ann Epidemiol 2007;17:569–74. 62. The Cochrane Collaboration. Available at: http://www.cochrane.
39. Hiatt RA, Porco TC, Liu F, et al. A multi-level model of postmen- org/. Accessed September 8, 2014.
opausal breast cancer incidence. Cancer Epidemiol Biomarkers 63. Hemila H, Chalker E. Vitamin C for preventing and treating the
Prev 2014;23:2078–92. common cold. Cochrane Database Syst Rev 2013;1:CD000980.
 Translational Research
134 Ciesielski et al April 2017

64. Hemil€a H, Chalker E. Vitamin C for preventing and treating the com- 81. Hiatt R, Samet J, Ness RB. The role of the epidemiologist in clin-
mon cold. Hoboken, NJ: John Wiley & Sons, Ltd. Available at: ical and translational science. Ann Epidemiol 2006;16:409–10.
http://onlinelibrary.wiley.com/doi/10.1002/14651858.CD000980. 82. Berkman LF. Social epidemiology: social determinants of health
pub4/abstract; 2015. Accessed March 17, 2015. in the United States: are we losing ground? Annu Rev Public
65. IARC Monographs on the Evaluation of Carcinogenic Risks to Health 2009;30:27–41.
Humans. Available at: http://monographs.iarc.fr/; 2014. Accessed 83. Landrigan PJ. Children’s environmental health: a brief history.
September 8, 2014. Acad Pediatr 2016;16:1–9.
66. Rooney AA, Boyles AL, Wolfe MS, et al. Systematic review and 84. Gurtler RE, Yadon ZE. Eco-bio-social research on community-
evidence integration for literature-based environmental health sci- based approaches for Chagas disease vector control in Latin
ence assessments. Environ Health Perspect 2014;122:711–8. America. Trans R Soc Trop Med Hyg 2015;109:91–8.
67. Databases, Searches & Other Resources: The National Toxi- 85. Gonzalez JP, Lambert G, Legand A, et al. Toward a transdisci-
cology Program. U.S. Department of Health and Human Services. plinary understanding and a global control of emerging infectious
Available at: http://ntp.niehs.nih.gov/results/dbsearch/index. diseases. J Infect Dev Ctries 2011;5:903–5.
html; 2014. Accessed September 24, 2014. 86. Barrow TM, Michels KB. Epigenetic epidemiology of cancer.
68. Bellinger DC. Lead neurotoxicity and socioeconomic status: Biochem Biophy Res Commun 2014;455:70–83.
conceptual and analytical issues. Neurotoxicology 2008;29: 87. Marsit CJ. Influence of environmental exposure on human epige-
828–32. netic regulation. J Exp Biol 2015;218:71–9.
69. Wild CP. Future research perspectives on environment and health: 88. Ciesielski TH, Marsit CJ, Williams SM. Maternal psychiatric dis-
the requirement for a more expansive concept of translational can- ease and epigenetic evidence suggest a common biology for poor
cer research. Environ Health 2011;10 Suppl 1:S15. fetal growth. BMC Pregnancy Childbirth 2015;15:192.
70. Rappaport SM. Implications of the exposome for exposure sci- 89. Ogino S, Campbell PT, Nishihara R, et al. Proceedings of the sec-
ence. J Expo Sci Environ Epidemiol 2011;21:5–9. ond international molecular pathological epidemiology (MPE)
71. Igl W, Johansson A, Wilson JF, et al. Modeling of environmental meeting. Cancer Causes Control 2015;26:959–72.
effects in genome-wide association studies identifies SLC2A2 and 90. Nishi A, Milner DA Jr, Giovannucci EL, et al. Integration of mo-
HP as novel loci influencing serum cholesterol levels. PLoS Genet lecular pathology, epidemiology and social science for global pre-
2010;6:e1000798. cision medicine. Expert Rev Mol Diagn 2016;16:11–23.
72. Ho MM, Yoganathan P, Chu KY, et al. Diabetes genes identified 91. Hiatt RA. Invited commentary: driving for further evolution. Am J
by genome-wide association studies are regulated in mice by Epidemiol 2015;181:459–62.
nutritional factors in metabolically relevant tissues and by glucose 92. Huang X, Bruce B, Buchan A, et al. No-boundary thinking in bio-
concentrations in islets. BMC Genet 2013;14:10. informatics research. BioData Min 2013;6:19.
73. Lusk CM, Dyson G, Clark AG, et al. Validated context-dependent 93. Clinical and Translational Science Awards, https://www.
associations of coronary heart disease risk with genotype variation ctsacentral.org/; 2014. Accessed June 23, 2014.
in the chromosome 9p21 region: the atherosclerosis risk in com- 94. Vogel AL, Stipelman BA, Hall KL, et al. Pioneering the transdis-
munities study. Hum Genet 2014;133:1105–16. ciplinary team science approach: lessons learned from National
74. Scriver CR. The PAH gene, phenylketonuria, and a paradigm Cancer Institute Grantees. J Transl Med Epidemiol 2014;2.
shift. Hum Mutat 2007;28:831–45. 95. Enhancing the effectiveness of team science. Committee on the
75. Casey L. Caring for children with phenylketonuria. Can Fam Science of Team Science. Board on Behavioral, Cognitive, and
Physician 2013;59:837–40. Sensory Sciences. Division of Behavioral and Social Sciences
76. Robert M, Rocha JC, van Rijn M, et al. Micronutrient status in and Education. National Research Council. Cooke NJ, Hilton
phenylketonuria. Mol Genet Metab 2013;110 Suppl:S6–17. ML, editors. Washington DC: 2015 by the National Academy of
77. Cancer Intervention and Surveillance Modeling Network, http:// Sciences; 2015.
cisnet.cancer.gov/; 2014. Accessed June 23, 2014. 96. Kemp SP, Nurius PS. Preparing emerging Doctoral Scholars for
78. Hiatt RA. Epidemiology: key to translational, team, and transdis- transdisciplinary research: a developmental approach. J Teach
ciplinary science. Ann Epidemiol 2008;18:859–61. Soc Work 2015;35:131–50.
79. Cancer Control Framework and Synthesis Rationale, http:// 97. Lawlor EF, Kreuter MW, Sebert-Kuhlmann AK, et al. Methodo-
cancercontrol.cancer.gov/od/about.html; 2014. Accessed June logical innovations in public health education: transdisciplinary
23, 2014. problem solving. Am J Public Health 2015;105 Suppl 1:S99–103.
80. Poincare H. The Foundations of Science: Science and Hypothesis, 98. Czajkowski SM, Lynch MR, Hall KL, et al. Transdisciplinary
The Value of Science, Science and Method. In: Cattell JM, ed. translational behavioral (TDTB) research: opportunities, barriers,
New York and Garrison NY: The Science Press, 1913:127. and innovations. Translational Behav Med 2016;6:32–43.