An Interview With Katherine Keyes, Co-Author With Sandro Galea Of
New Book “Population Health Science”
At last month’s
one-day conference sponsored by the International Journal of
Epidemiology, Katherine Keyes of Columbia University presented
a talk entitled, “Why Does Epidemiology Matter?” Quite a
thought-provoking question to discuss in a room full of
epidemiologists.
Weaving through the ongoing debate about different
approaches to causality in epidemiological research, Keyes ultimately
concluded that she was not sure of the answer to her own question.
She noted “the pages of our journal are beginning to be filled with
increasingly cantankerous debates in the field about who gets to
define what a causal question is and the very role of science in our
framing of study design and analysis, our research findings are
increasingly questioned, and our funders are increasingly questioning
our very relevance.”
Keyes became uncomfortable with her role in
epidemiology and decided that what she cares about and wants to work
on may not be definable in terms of epidemiology.
This conclusion created the motivation for her new
book: Population Health Science. Co-written with
Sandro Galea, now Dean of the Boston University School of Public
Health, the text is meant to “articulate and expand on principles that
allow readers to engage in population health thinking, and to have a
platform and reservoir of theory from which to draw when formulating
questions of relevance to population health.” (See related article on
the Principles of Population Health in this issue.)
Citing major past public health wins achieved through
clean-water initiatives, vaccination campaigns and smoking cessation
efforts, the authors note the landscape of disease is evolving. “We
are faced increasingly with challenging chronic and acute diseases for
which no clear answers are apparent,” they note. (1) They lament that
the methods currently in use in epidemiology research fall short in
delivering impactful results on population health. We interviewed
Keyes to learn more about what motivated her to write the book, the
future relationship between epidemiology and population health
science, and the role of population health science in public health.
Galea concurred with all of the views expressed in the interview.
The Interview
EM:
You’ve stated this in
your talk, but can you briefly describe your take on the state of
epidemiology that motivated you to write a book on population health
science?
Keyes:
Epidemiology has
always been a discipline that draws a diverse array of scholars with
varied interests and backgrounds. There has also always been a
discomfort about how to define epidemiology - it is a collection of
methods, a set of research agendas, an approach to study design and
analytic approaches, all of the above, or none of the above? If you
ask 100 epidemiologists to define what they do, you will undoubtedly
arrive at many different answers. Thus there has been an interest in
defining epidemiology, and one predominant voice in the field has been
defining epidemiology in terms of the policy relevance of the
questions, the ability of the analytic approaches to generate
estimates that can be interpreted as well-defined causal estimates.
Such voices have advanced our field in important ways, and have forced
us to confront the reality that many of our traditional analytic
approaches are insufficient, and confront the assumptions that are
embedded in the traditional design and analysis tools that we use.
That being said, as epidemiology is defining itself as
a field in new and specific ways, there are scientists who
increasingly do not fit the mold of the epidemiologist, those who seek
to have a broad and cross-disciplinary research program that focuses
questions of inequality and social justice, assessing the world’s most
significant disease burdens, using theoretical frameworks and tools
that are not necessarily covered in standard epidemiological texts. We
wrote Population Health Science with those scholars in mind, in an
attempt to begin aggregating and presenting theoretical and
methodological ideas about population distributions of health and
disease and what we might be able to learn about them. We draw on
scholars not only from epidemiology but preventive medicine,
philosophy, social science, and health policy in an attempt to begin
formalizing what a discipline of population health science might look
like, and what foundational principles should guide our work.
Conflict with Epi?
EM:
In the introductory
chapter of the book, you describe population health science as
“quantitative underpinning” and “the scaffolding of a broader agenda
for both preventive medicine ... and public health.” Epidemiologists
have long stated that epidemiology is the basic science of public
health. Does population health science supplant epidemiology? How do
the two coexist moving forward?
Keyes: Yes, we see population health science and epidemiology
as correlated and coexisting. Whereas epidemiology is increasingly
being defined by a particular set of methods that answer a particular
set of questions, we see population health science as a broader
framework of ideas and methods that can be used for a diverse array of
research questions - some of those questions may be ‘epidemiological’
but some may not. Thus, we do not see these disciplines as competing,
but hopefully, as complementary and reinforcing each other.
EM: Can you expand on your
statement that population health science is less a collection of
methods and more an articulation of values?
Keyes: We have disciplines that have developed and described
methods - in epidemiology, biostatistics, economics, health policy,
psychology, and many other disciplines. We see Population Health
Science as a way to aggregate theory and ideas in ways that
bring aligned researchers together towards common goals and research
agendas, using a wide variety of methods and triangulations of
evidence. Our book is not a textbook of methods that one can use to be
a population health scientist, but the ideas and theories from
generations of scholars who have laid the foundation for the formation
of the discipline.
Reception
EM:
What kind of reception
are you getting to the new book and its ideas? Are epidemiologists
seeing it positively? Some we know may be seeking to change the
emphasis in epidemiology itself rather than trying to emphasize a new
discipline. Why did you choose differently?
Keyes: Thus far the book
has had a great reception - our colleagues from both epidemiology and
aligned disciplines have had useful feedback and comments, and the
book is beginning to be used in various classes, which was our
ultimate goal. Part of the reason that we decided to work on ideas
for invigorating a discipline of Population Health Science is that it
is among the buzzwords that we keep hearing about in the public health
community.
There are now degree programs, classes, and other
academic programming labeled as “Population Health Science”, but there
was no formalized agenda, in our view, of what Population Health
Science is, how it might be distinct from public health and
epidemiology, what principles guide the work, and what questions fall
under its umbrella. Our book was our attempt at a first pass on
addressing those topics.
EM:
As a pragmatic
science, do you see population health scientists engaged in
translating their science or advocating for solutions to population
health problems? Will population health scientists be more “applied”
than epidemiologists?
Keyes: I do not see
population health science as more applied than epidemiology, which is
already a very applied discipline that often explicitly recommends
research that is directly policy relevant. Population Health Science,
to me, is agnostic to direct policy relevance. Of course, our work
should always have in mind practical ways of improving population
health, but we should not privilege those research questions with
direct policy relevance, as such approaches can often be limiting. For
example, it is often difficult to envision direct policy applications
of the study of inequality, but that does not make inequality less
worthwhile to study. Perhaps the estimates that we obtain from studies
of inequality should be interpreted with more caution as they might
not be direct causal estimates, but inequality is well worth studying,
both theoretically and empirically.
EM:
In your words,
“population health scientists are charged with considering the causes
of [health and disease] distributions, rather than the causes of
cases.” How distinct are the underlying causes of distributions rather
than of cases?
Keyes: It would depend on
the outcome under investigation. As an example we use in the book,
consider intelligence. Within a high socio-economic population with
high quality education, supportive learning environments, and
excellent nutrition and health care, why one person is high IQ and
another person low IQ might be entirely determined by inheritance and
genetic lottery. But in a different environment in which goods and
services are unequally distributed, the reasons why one person has
high IQ and another does not might not be linked to genes. As such,
the determinants of the population mean and standard deviation of a
distribution of health may have different predictors than the
determinants of being high or low on the distribution in any
particular population.
EM:
Principle #8
recognizes that improvements to overall population health may be a
disadvantage to some groups. Can population health science help to
choose between equity and efficiency in these cases?
Keyes: In my view, no.
When there are tradeoffs between equity and efficiency (and it is
important to note that often times the two can be aligned), whether
equity or efficiency is the best approach will depend on the values of
the decision makers, rather than on empirical evidence alone. As an
example that is often cited, age has been used as a factor in decision
making for organ donation. Privileging the young is efficient but
would not be equitable, as it discriminates based on age. The answer
for what to do in that situation is a matter of values. The same
issues arise when we discuss distributing health care resources at a
population level, should we privilege preventive medicine and care for
the young, or treatment and palliative care? Of course, this is a
straw man argument as we would want to do both, yet in a world of
finite resources, some difficult decisions need to be made.
Principles
EM:
Your principles of
population health are succinctly stated. Can you expand a bit on how
you see them as being useful in doing population health science?
Keyes: I think that
remains to be seen. The principles are not prescriptive of a
particular research agenda, methods, or study design, rather guiding
principles in how to outline our research program and choose topics of
importance. I see them being used in training, to help young scholars
understand the history of our field and the principles to keep in mind
as they progress in their research program.
EM: This book and your
previous book were both co-written with Sandro Galea. How did your
collaborative writing partnership come about? Or what do you feel is
the biggest advantage gained in collaborative writing?
Keyes:
I began working with Sandro when he joined the department of
epidemiology at Columbia University as chair as I was finishing my
doctoral degree and beginning a post-doc at Columbia. Very quickly we
saw that our interests and ideas were aligned and we began meeting
regularly to talk about how to teach, think and write about
epidemiology as a discipline. These discussions led to our first
textbook, Epidemiology Matters, but throughout the writing of that
book we realized that many of the ideas that guide our personal
research agendas did not fit the subject matter of an introductory
epidemiology text.
When we finished Epidemiology Matters we realized that
we had another book to write which was very different in frame and
content, and very quickly began writing the second book. I think our
collaborative arrangement works because Sandro and I are both
enthusiastic scholars. We love new and challenging ideas, love working
on a wide variety of projects and topics, and are not dissuaded by
setbacks or criticism. I think both Sandro and I also insist on
thinking about larger questions in our discipline beyond the next
grant or the next paper, and these inquiries naturally led us to
collaborate.
■
|