Interview With George Kaplan---One Of Four Co-Editors Of A New Book
on Complex Systems Approaches To Problems In Population Health
We learned recently
about the publication of “Growing Inequality—Bridging Complex Systems,
Population Health, and Health Disparities” co-edited by George
Kaplan, Ana Diez-Roux, Carl Simon, and Sandro
Galea. We read very positive reviews about the book and invited
the University of Michigan’s George Kaplan to answer a few short
questions to help readers better understand how the book came to be
and its significance for epidemiologists. Below is the interview.
Creating the network you describe was unique. What were some of the
rewards for you as a participant/organizer?
Over many years – really many decades – I became more and more
convinced that we as epidemiologists needed to broaden our
perspectives in order to understand and improve population health for
all. While the search for independent causes – the often elusive magic
bullets – had worked in some cases, many of the problems that we were
interested in were more characterized by tangled webs of factors where
the search for a single cause was not realistic and often led to
definitions of problems that bore little resemblance to the real
Being able to bring
researchers who had experience in modeling and simulating complex
with those who knew
something about the biological, behavioral, social, spatial, and
institutional factors that produce disease outcomes in individuals and
populations was a rare opportunity to push the envelope, to move from
Occam’s quest for simplicity towards Einstein’s dictum that
“everything should be made as simple as possible, but not simpler.”
My colleagues and
I, assembled into the Network on Inequality, Complexity and Health, (NICH)
hoped to demonstrate the feasibility of embracing the complexity of
disease causation, rather than trying to eliminate it or control for
it. That in itself would have been reward enough, but the response we
got from other colleagues, particularly those in early stages of their
careers, indicating a thirst to learn more about these approaches was
the icing on the cake.
Did the interdisciplinary benefits that were anticipated or imagined
really come to pass? What were some of the challenges in working as a
network or through this network process?
All together we involved around 50 colleagues in this journey, with 18
coalescing into the membership of NICH and meeting for almost five
years. It was a very diverse group, representing perhaps well more
than a dozen disciplines (epidemiology, neurosciences, computer
science, economics, political science, public policy, mathematics,
communications, nutrition, law, education, medicine, psychology, child
While such a
diverse group brings great strength, it does bring with it certain
challenges. Many of the people from the health and social sciences had
no or little previous experience with complex systems modeling, and
several of the complex systems modelers in the network had little or
no experience working on health disparities or population health
topics. Thus, there was a continual process of providing background
information in all three areas, and of learning from each other. While
valuable, this was an iterative process that was time consuming but
essential. While not fundamentally different from my experience in
other interdisciplinary groups, there was the added challenge of
becoming comfortable with the logic, process of developing, and
understanding of the results of complex systems simulations.
interviews with each NICH members, there was great enthusiasm for the
network process. Typical were statements like,
“…[the diversity of
the network] allowed me to see and work on connections that I wouldn’t
have worked on otherwise,”
“…exposure to new
ideas was great, excellent people,”
expanded the scope and reach of what I do…invaluable colleagues,”
“I have moved to a
serious commitment to modeling health disparities and population
interest in life course issues,”
serious discussion of social determinants of health and health
disparities issues to my teaching of non-public health students,”
“….changes the way
that I look at things…enlarged my vision,”
No member indicated
in any way that they would have preferred less diversity in
backgrounds and methods, and most commented on the benefits of such
diversity to their thinking.
Finally, I would
say that over time the network members better understood and became
more comfortable with the use of in-silico/virtual/simulated
worlds in which to examine counter factuals that certainly were not
amenable to test in other ways.
Can you give a couple of examples of payoffs from the network process
and payoffs regarding any of the specific topics investigated?
The biggest payoffs were probably those that had to do with our
ability to stimulate network members and others into using complex
systems approaches in epidemiology and public health, and to
legitimize such efforts. There was dramatic increase in courses and
programs centered on using complex systems analytic approaches in a
number of departments and schools, and we attracted over 800
registrants, from 39 states and 19 countries, to a symposium at the
NIH Natcher Center that was centered around NICH and its approach.
It also became
clear that while there is the potential for considerable payoff, there
are considerable educational, training, and resource needs in order to
pursue the use of complex systems methods in epidemiology and other
areas of public health.
There were any interesting findings,
with some mentioned below . They are initial forays – much work needs
to be done. In fact, a rather small part of the funding for NICH
supported these efforts. They are more proof of concept than finished
work, as the purpose of the network was to demonstrate innovative
approaches not to generate complete bodies of work. You will note that
most of the approaches are considerably ‘upstream’ of most
epidemiologic studies. That reflected both the composition of
the NICH group, as well as a
rapidly evolved consensus that complex systems approaches could be
useful for approaching such ‘tangled’ problems.
et al. found that policies that modified the neighborhood educational,
physical activity, and nutritional environment were found to
dramatically reduce Black/White BMI disparities. But the effects took
considerable time, and each policy had a different time course.
Stange et al. found that
access to primary care actually increased the effects of specialty
care in treating specific diseases.
Yang et al. found that for
low SES populations, changing attitudes about walking was not
sufficient to change behavior unless the environments were changed as
Kreuger et al. showed that
the factors that affect preventive dental care utilization differ
according to SES – cost and access were the predominant factors for
low SES patients, while issues of trust impacted more on high SES
Kassman and Klasik simulated
the factors driving disparities in college enrollment and found that
affirmative action based on SES was not as effective as race-based
affirmative action in reducing disparities in enrollment.
Simon et al. modeled the
dynamics of the spread of crime and mass incarceration, and developed
a kind of “flight simulator” that allowed them to examine differential
effects of policies related to incarceration rates, recidivism etc.
Wolfson and Beall, created a detailed simulation to
help understand the different associations between income inequality
and health in the US and Canada, and were able to
assess onomic mobility than to the role of neighborhood racial and
income segregation. Their initial results suggest
that the different patterns in the US and Canada are more related to
differences in intergenerational socioeconomic
mobility than to neighborhood level factors.
Kumanyika et al. simulated various factors that might account for
differences in levels of physical activity between Black men and
women. They found an important role for the supportive behavior of
others in explaining the relatively lower levels of leisure-time
activity in Black women.
Boyce et al., found an
important role for classroom structure and teacher behavior in the
mental health consequences of classroom hierarchy.
Sample chapter and TOC: