Different
Views About Causality Clash In The International Journal Of
Epidemiology
December Marks
Final Issue Under Davey-Smith And Ebrahim
It is likely to be remembered and useful for longer
than the usual “shelf-life” of a journal issue. But then again, that
may not be so surprising. The International Journal of Epidemiology
under the co-editorship of George Davey Smith and Shah
Ebrahim has been anything but a typical journal. Its upcoming
December issue entitled “Causality in Epidemiology --The Final
Frontier”---will be the final one edited by Davey Smith and Ebrahim
and promises to lay bare the tensions and controversies about how the
work of epidemiologists can best serve public health.
In a journal already remarkable for the size of the
individual issues, The December issue will exceed even its own norms
by publishing what the co-editors have said “…will probably be the
largest single issue we have published and will provide readers with
an up to date and comprehensive review of schools of thought in
causality.”
Plan for the Issue
As of late November, the plan for the December issue is to publish
more than a dozen articles related to causality. Five of these papers,
letters, and commentaries can be grouped around a paper published
earlier this year by Jan Vandenbroucke, Alex
Broadbent,
and Neil Pearce entitled “Causality and causal inference in
epidemiology: the need for a pluralistic approach. A second group of
papers can be grouped around a synopsis by Tyler VanderWeele of
his book “Explanation in Causal Inference: Methods for Mediation and
Interaction” published last year. The IJE invited the synopsis,
commentaries on the book, and a response to the comments by
VanderWeele. Both groups of papers have now been published online by
the IJE but will appear in print only when the December issue appears.
The list of papers published to date is provided below.
Critics
Two papers published online, the first by Vandenbroucke and
colleagues, and the second more recent one by Nancy Krieger
and George Davey Smith entitled “The tale wagged by the DAG:
broadening the scope of causal inference and explanation for
epidemiology” offer a vigorous challenge to the potential outcomes
approach to causal inference in epidemiology.
Vandenbroucke and colleagues describe the growing
popularity of the “causal inference” movement in epidemiology and
explain their purpose for writing a critique as “…to forestall the
emergence of a ‘hardline’ methodological school within epidemiology,
one which we feel would damage the discipline if it became the
dominant paradigm.” Co-author Neil Pearce told the Monitor causal
inference is the wrong term to be used in describing modern methods in
epidemiology because inference requires
consideration
of a broad array of evidence from different sources. Also, the focus
on examining interventions causes investigators to lose sight of the
broader concept of the health of populations.
For Krieger and Davey-Smith, the motivation is “…to
strengthen epidemiological science and its capacity to contribute
usefully to the multi-sectoral work urgently needed to improve
population health and reduce, if not eliminate, health inequities.”
They fear that “these new ‘cutting edge’ methods will, by virtue of
their rule-bound nature, limit the scope of epidemiology and its
impact on the urgent real world problems of global population health."
High Stakes
Krieger and Davey Smith express concern about the
prominence of causal inference in epidemiology using counterfactual
and potential outcome reasoning. The stakes are high they assert
because epidemiology seeks to explain the determinants of health and
the answers provided or the causes pinpointed can make a big
difference. “There is no short cut for hard thinking about the
biological and social realities and processes that jointly create the
phenomena we epidemiologists seek to explain, always with an eye
towards producing knowledge that we and others can use to improve
population health, reduce preventable suffering, and we add, advance
health equity.”
While both sets of authors acknowledge the usefulness
of the potential outcomes approach in particular situations, both sets
of authors fear the same consequence, namely that it is too
restrictive in terms of the questions that can be asked, the studies
that can be designed, and the interpretations that can be rendered.
Both sets of authors call for a less restrictive approach that
considers many different types of questions and studies and considers
varying types of evidence in reaching a conclusion about causality.
Pluralism
In thinking about causality, Krieger and Davey-Smith
point to a framework called the “inference to the best explanation”
which they believe has greater potential to consider different types
of evidence, that is, to be more pluralistic in what it considers in
making inferences. Likewise, Vandenbroucke and colleagues end up
recommending a “pragmatic pluralism”.
Papers from the December issue of the IJE published early online:
Krieger N and Davey Smith G. The tale wagged by the DAG:
broadening the scope of causal inference and explanation for
epidemiology
https://tinyurl.com/hke5gcj
Vandenbroucke et al. Causality and causal inference in
epidemiology: the need for a pluralistic approach
https://tinyurl.com/hm5l2zz
VanderWeele et al. Letter to the Editor: Re: Causality
and causal inference in epidemiology: the need for a pluralistic
approach
https://tinyurl.com/jpgfjqf
Chiolero A. Letter to the Editor: Counterfactual and
interventionist approach to cure risk factor epidemiology
https://tinyurl.com/jhuhh8r
Schooling CM et al. Letter to the Editor: Causality and
causal inference in epidemiology: we need also to address causes of
effects
https://tinyurl.com/gwujvcr
Broadbent A et al. Letter to the Editor: Authors’ Reply
to: VanderWeele et al., Chiolero, and Schooling et al.
https://tinyurl.com/hqqt85o
VanderWeele TJ. Explanation in causal
inference: developments in mediation and interaction. Int. J.
Epidemiol. first published online November 17, 2016.
doi:10.1093/ije/dyw277
https://tinyurl.com/htgxbvv
Kaufman JS. The epidemiology of two
things considered together. Commentary on: Explanation in Causal
Inference: Developments in Mediation and Interaction, by Tyler J.
VanderWeele Int. J. Epidemiol. first published online November
17, 2016.
doi:10.1093/ije/dyw278
https://tinyurl.com/jy49mvh
Oakes JM and Naimi AI. Mediation, interaction,
interference for social epidemiology Int. J.
Epidemiol. first published online November 17, 2016.
doi:10.1093/ije/dyw279
https://tinyurl.com/gour5yw
Pearce N and Vandenbroucke JP.
Causation, mediation and explanation Int. J. Epidemiol. first
published online November 17, 2016
doi:10.1093/ije/dyw281.
https://tinyurl.com/jl9w7xf
VanderWeele TJ. The role of potential
outcomes thinking in assessing mediation and interaction Int. J.
Epidemiol. first published online November 17, 2016
doi:10.1093/ije/dyw280.
https://tinyurl.com/h3vr5yj
■
|