|
Epi Wit & Wisdom Articles
Pittsburgh Epidemiology Chairman
Links Survival of Epidemiology to Focus on Public Health Problems
Criticizes “Circular
Epidemiology”
“The Irresponsibility of Good
Epidemiologists in Public Health” This was not the first and probably
will not be the last provocative statement will see coming from Lewis
Kuller, chairman of the Department of Epidemiology at the University
of Pittsburgh. On this occasion, the attention-grabbing title came
from a seminar he gave at the Johns Hopkins Depart-ment of
Epidemiology in December. According to Kuller, epidemiology as
currently practiced has the tendency of becoming circular, that is, of
doing the same things over and over again (“circular epidemiology”),
rather than progressing from descriptive to analytic to experimental
studies. In contrast, “the heroes of epidemiology,” says Kuller, “have
identified a problem (descriptive epidemiology); developed methods to
test specific hypotheses (observational epidemiology); and then
experimented (clinical trials) to prove or disprove a hypothesis, and
then applied good public health and preventive medicine strategies to
utilize the information that they have acquired to reduce morbidity
and mortality.”
Circular Epidemiology
What is “circular epidemiology?”
Kuller defines circular epidemiology as the continuation of specific
types of epidemiological studies beyond the point of reasonable doubt
of the true existence of an important association or lack of it. Put
another way, circular epidemiology occurs when the number of studies
of the consistency of an association becomes extreme.
According to Kuller, “good
epidemiology journals are filled with very well done epidemiological
studies that unfortunately are repetitions of the obvious or
variations on a theme.” Kuller filled his presentation at Johns
Hopkins with examples of circular epidemiology such as studies of
weight gain as a risk factor for diabetes, a high level of LDLc as a
risk factor for heart attack, early age at first pregnancy associated
with a reduced risk of breast cancer, and, of course, everyone’s
favorite example in epidemiology, smoking and lung cancer.
Causes
Why does epidemiology have this
tendency to stall in place? Kuller advanced several potential
explanations. First, epidemiology is sub-categorized into
sub-specialties. Methodologic “truths” discovered by practitioners in
one area might be rediscovered by epidemiologists in another area.
Second, some researchers do not understand the underlying biologic
principles of the conditions being studied. Some epidemiologists even
go so far as to say that such understanding is unnecessary. What
happens is that previously described inappropriate methods are used
and hypotheses are considered proven or unproved without regard to
biological plausibility and logic, thus leading to more inappropriate
studies. Third, epidemiology lacks a systematic approach to the
acquisition of new knowledge to reach a public health goal. Put
another way, goal-directed behavior is not a requirement of good
epidemiology studies. Fourth, funding support for epidemiologic
research is biased toward the continued study of already proven
hypotheses because a new hypothesis which lacks substantial prior data
is unlikely to be successful in terms of peer review. There is a bias
towards simplicity and towards an already proven hypothesis among
those responsible for research support decisions. This “band-wagon
syndrome” says “If I have already done it, then I will support other
people doing it.” Fifth, there is also a publication bias in favor of
already proven hypotheses. There are no clear stopping rules--when do
you stop publishing papers showing that smoking causes lung cancer?
that blood cholesterol levels predict heart attacks? that being less
educated and having a low socio-economic status is bad for your
health? Sixth, epidemiologists forget the generalizability of causal
associations. Studies are replicated in different age, race and sex
groups when such replication is unnecessary. According to Kuller, “we
forget that one of the false arguments for the Tuskegee experiment was
that treatment of syphilis with penicillin would be different between
blacks and whites.”
The consequences of “circular
epidemiology” include wasted resources that could be better used to
move hypotheses forward on the continuum from descriptive studies to
clinical trials. Also, unnecessary replication “only delays the
implementation of good public health practices and often is a
detriment rather than a help to the populations at risk,” according to
Kuller.
Solutions
He believes the antidote for
“circular epidemiology” is for epidemiology to truly dedicate itself
to being the basic science of public health and preventive medicine.
Just as public health needs epidemiology to create sound science-based
programs, so epidemiology needs public health goals to remain vital
and relevant, and to earn the financial support it needs for research.
Contrary to the view that epidemiologists should present their
findings in as detached a manner as possible and refrain from
discussing the potential implications of their work, Kuller argued
that defining the implications of an epidemiologic study in
publications should be encouraged not discouraged. Says Kuller,
“epidemiology papers should clearly describe hypotheses and
implications of results in terms of direction of future studies and
potential importance of results for the practice of public health,
prevention and clinical medicine...the role of the current study in
the evolving process of epidemiology from descriptive to experimental
epidemiology should be clarified.”
Kuller concluded his remarks by
saying...“the future of epidemiology is very bright if we continue to
stress that it is an important basic science of preventive medicine
and public health...we should continually monitor the successes and
failures of epidemiology studies in improving the ‘public health’.
Published March 1999 v
|
