|
Epi Wit & Wisdom Articles
ACE Attendees Reject Motion That
Risk Factor Epidemiology is Bad For the Field
Some Say These Are Exciting
Times For Epidemiology
Following what one attendee
called a “masterful” job of rebuttal by Epidemiology editor Ken
Rothman, an estimated 300 attendees at the recent American College of
Epidemiology (ACE) meeting voted approximately 2 to 1 against a motion
proposed in an Oxford-style debate that risk factor epidemiology is
placing epidemiology at risk.
The motion could have been
interpreted as a debate on the limitations or inadequacies of risk
factor epidemiology, or alternatively, as a debate on the view that
risk factor epidemiology is just plain bad for the field of
epidemiology.
In his rebuttal, Rothman
successfully interpreted the motion for the audience as meaning that
risk factor epidemiology is harmful to the field of epidemiology and
he asked people to vote with that interpretation of the motion in
mind. NIEHS epidemiologist Marilyn Tseng noted, “I thought that it all
came down to semantics...It’s hard to be an epidemiologist and vote
that what most of us are doing is actually harmful to epidemiology.”
The debate was moderated by New
England Research Institute’s John McKinlay, who termed the debate “a
big success.” Speaking for the affirmative were AJPH editor Mervyn
Susser and Harvard’s Lisa Berkman. Susser argued that the risk factor
paradigm cannot meet the demands of new global patterns of disease,
and that it needs to manage multilevel analyses, especially given
developments in molecular epidemiology and in communications systems.
He believes we are near displacement of the risk factor paradigm.
Berkman argued that by focusing on individual risk factors and their
independent effects, risk factor epidemiology tends to look at
proximate (“downstream”) rather than distal (“upstream”) causes of
disease. It ignores the social context in which risk factors emerge
and only looks at risk factors as attributes of individuals. Berkman
called for a more “contextualized” approach to epidemiology that can
recognize pathways, multi-level influences, and correlated risks. She
pointed out that clinical trials/behavioral interventions have largely
been disappointing, and that the largest changes in risk factors have
been the result of widespread secular changes that we do not really
understand. She suggested that we do not really know enough about
social forces that have important influences on risk factors and on
health, probably because we are not closely allied with fields that
deal with broader social forces.
Speaking for the negative were
UNC’s David Savitz and Epidemiology editor Ken Rothman. Savitz argued
that the focusing on proximal causes is important to develop
interventions, and that risk factors are studied as isolated causes in
order to achieve “bits of understanding.” Overall, he thought that
while epidemiology could move into a new realm in looking at causes,
the risk factor paradigm remains viable and valuable. Rothman began by
making statements that epidemiology responds to the times rather than
being shaped by intent, and that risk factor epidemiology studies
causal mechanisms. He argued initially that “upstream” factors should
not be studied at the expense of more proximal factors and that causal
pathways must be elucidated before we can think about changing
society. Furthermore, even when we understand pathways and have
choices, it does not always follow that the best place to intervene is
“upstream.” As an example, he cited the relationship between poverty
and infant mortality. It may be more feasible to correct the problem
of vitamin A deficiency to reduce infant mortality than to correct the
problem of poverty.
According to McKinlay, the
audience lined up quickly to speak from the floor. Among the points
made in favor of risk factor epidemiology were that macroeconomics
cannot predict well and epidemiology should not go down the same path
to become “macroepidemiology,” and that risk factor epidemiology
should not be discarded, despite its limitations. One participant
noted that if we had taken a broader view, we might have been able to
predict the re-emergence of infectious diseases and that generally we
should put ourselves in a position to be proactive and not just
reactive vis à vis disease. One observer commented on the need for
other disciplines to inform research and gave the example of
bronchitis reduction in England because of “upstream” changes in laws
about fires in homes and not because people wore “downstream” scarves
and masks.
In response, Susser agreed that
risk factor epidemiology is viable and that we cannot do science
without being reductionist. But it is an inadequate technique, he
said, to manage multi-level analyses. We do not have the techniques,
but we can get them, he stated.
Berkman noted three sources of
dissatisfaction with the current practice of epidemiology—1) methods
are ill-equipped to sort mechanisms from confounding; 2) behavioral
interventions have not been successful, as opposed to social changes
(e.g., taxation), but we have hardly an inkling about social forces;
and 3) the causes of disease themselves come and go, but the same
people are consistently at risk—the poor. During the debate, some
speakers stated that epidemiologists working at different points along
the causal stream may actually have different values with “upstream”
scientists more motivated by social justice and “downstream”
scientists more driven by scientific values.
In closing, Rothman agreed with
Susser and Berkman that we do not have the methods, and he agreed that
we should have been able to anticipate changes in health and disease
patterns. As mentioned above, he disagreed that risk factor
epidemiology is doing harm. He argued that the proposition should have
been— risk factor epidemiology is pacing epidemiology briskly.
The negatives defeated the
motion. According to Tseng, “In the end, I think the debate settled
the issue of whether or not risk factor epidemiology should be thrown
out (the decision was no), but it did not settle the issue of whether
or not risk factor epidemiology and its corresponding paradigm are
sufficient to overcome the criticisms of epidemiology that have been
raised in recent years.” According to McKinlay, there were no final
answers provided during the debate but there was a sense in the
audience that this is an exciting time in epidemiology, a time of
ferment. He sees the debate as a reflection of the fact that the
profession is in a healthy state. Tseng agrees that “these
disagreements inspire some hope that epidemiology can be doing a
better job at promoting public health than it has been, and that we
might have the motivation and the creativity to guide the field in
another direction, even if it is just another ‘tributary’ of the
field.”
Published October 1997 v
|
