|
Historical Interviews
[Editor’s note:
Way back in 1990 we included mention of Ken Rothman’s 1981 article in
the NEJM on the Rise and Fall of Epidemiology in our recap of events
for the decade 1980 - 1989. At the time, we said that perhaps we would
invite him to write a revised history of the period 1980 - 2000 in
anticipation that he might want to make revisions at the halfway mark.
Well, we did not quite get around to extending that invitation way
back then, but instead asked him in May 2000 if he would agree to be
interviewed some 20 years after the fact and give us the benefits of
his hindsight. Ken graciously agreed, and the result is the interview
published this month as part of our 20th anniversary issue. It
provides a view on epidemiology yesterday, today and tomorrow and
makes for stimulating reading from one of epidemiology’s most
provocative thinkers.]
Epi Monitor: In
1981, you published a widely-read tongue-in-cheek scenario of the
future of epidemiology in the New England Journal of Medicine entitled
the “Rise and Fall of Epidemiology 1950 - 2000.” It has been almost 20
years since that piece appeared, and no one else has ever written
anything quite like it to our knowledge. It painted a picture of what
the future landscape of epidemiology would look like. What event or
incident triggered you to write the article in the first place?
Rothman: I
wrote the article out of a sense of frustration with growing
bureaucratic obstacles in epidemiology. My main motivation was to
embed in an amusing essay a warning about what might happen to
epidemiology if the bureaucratic obstacles continued to mount.
Epi Monitor:
You did a couple of things in this article, including writing your
abridged version of the history of epidemiology from its establishment
by John Graunt in 1662 to 1980, the time your
article was written. Has anything changed in your thinking or research
over the years that would make you describe the history of
epidemiology or any of its periods differently today?
Rothman:
Abridged is an understatement. The few introductory paragraphs were
barely a caricature of the history of epidemiology. It jumped from
Graunt to Farr and Snow to the mid-twentieth century, omitting
everything else. I intended it only to give a reader a sense that the
rise of epidemiology was mostly recent and steep, and perhaps to
educate clinicians a little about the contributions of these British
pioneers. There is much that was left out, such as the work of
Pierre Louis, Major Greenwood and
Wade Hampton Frost, to name just a few.
Epi Monitor:
What about the “boom period” for epidemiology, from 1950 - 1980?
Would you want to characterize that period any differently or say
anything different about it now?
Rothman: The
boom may have started around 1950, after the second world war, but I’m
happy to say that it didn’t end in 1980. It’s continuing today.
Epi Monitor:
Your scenario painted a fairly bleak picture of where epidemiology
would end up in 2000. Forgetting for a moment the details of the
scenario you described, what was the main concern back then that drove
you to write this piece about the state of epidemiology?
Rothman: Some
readers of the essay may have taken the message in the paper’s title
too literally. The paper was not intended to predict the path of
epidemiology over the next 20 years any more than George
Orwell intended his satiric novel 1984 as a prediction of
what the world would be like in 1984. It was just a vision of the grim
possibilities that might await us if the rapidly escalating concerns
over privacy and bureaucratic interference in epidemiologic research
continued along the path that they had taken during the 1970s. This
warning was interpreted by some, particularly those who read little
beyond the paper’s title, to mean that I was pessimistic about the
future of the field. That was not the case.
There was a confluence of events
that motivated the paper. I had spent nine months with an approved and
funded research project kept on hold while some government official in
Washington had the study proposal in his or her inbox. Under a new
regulation all approved government-sponsored research contracts had to
get further approval from the Office of Management and Budget. After a
long wait the questionnaire was approved, but only with numerous
changes that made little sense and compromised the value of the
information. There was no dialogue with me, and the official had no
appreciation for epidemiology.
Around the same time, new and
much more stringent privacy rules and the increased truculence of
hospital IRB review boards slowed epidemiologic research considerably.
IRBs were influenced by the scandalous Tuskegee trial and some other
outrageous ethical lapses that had been in the news. But the IRBs
overcompensated by making demands such as written consent from dead
people to look at their medical chart, or informed consent for anyone
whose medical record might even be screened to determine eligibility
for a study. These demands were imposed even on record-review studies
that never had any patient contact, and which would never report
individual level information. One of my colleagues, whom I referred to
in the essay with a pseudonym, spent nearly a year of his career
getting a batch of IRB approvals for a multi-institution case-control
study.
On another front, the federal
government was busy creating guidelines on how to conduct
epidemiologic research. Epidemiologists were uncertain how the
guidelines might be applied and the extent to which they would affect
our independence as scientists, but they set an ominous precedent for
bureaucrats to intrude into scientific research. Those of us who
believed that imagination and creativity ought to play a key role in
study design were concerned about the implications of this new
development.
Epi Monitor:
The issues you put your finger on back in 1980--IRB oversight and
government regulation of research are vitally important issues for
both society and epidemiologists. Has the burden for epidemiologists
of being overseen and regulated diminished, increased or stayed at the
same level as that in 1980?
Rothman: Last
week I received a copy of the BMJ with an article describing the IRB
approval process for a multi-center study in the UK. The article
showed a photograph of the project staff with the 60,000 sheets of
paper that they had to submit to 176 centers to obtain IRB approval.
They wrote that photocopying time alone was more than 50 hours. This
nightmarish report is worse than any prediction in my essay.
Fortunately, the situation is not that bad for every study. In many
places IRB approval has become more streamlined for epidemiologists,
as IRBs become more sophisticated. On the other hand, every time new
people join an IRB, the education process must begin again. Here’s an
example of the problem: Rigorous informed consent for a randomized
trial is an essential ethical protection. But even if getting consent
leads to the loss of many potential participants in a trial, the loss
of participants before randomization will not affect the validity of
the randomized comparison among those who do volunteer. On the other
hand, if informed consent weeds out half of the participants in a
case-control study, the validity of the study may be seriously
impaired. Every new IRB member needs to be educated about design
issues such as the different implications for validity when people
decline to participate in a trial and when people decline to
participate in a case-control study.
Epi Monitor: If
the burden of being overseen and regulated already seemed onerous back
in 1980, and if we assume the burden has increased since then, why has
this large and growing burden not led to the kinds of outcomes you
tried to warn us about in 1980, and which presumably you would want to
warn us about again today if the scenarios you described have not yet
come to pass?
Rothman:
Actually, I’m not sure that regulation has increased dramatically
since 1980. Some things may have gotten out of hand---the BMJ article
gives one example---but in other ways there seems to have been a
leveling off below the intolerable threshold.
Epi Monitor: In
our opinion, the burden has increased and continues to grow. Witness
new privacy regulations, a new data sharing law, and the tightening of
IRB regulations. It is fascinating to wonder why the scenario you
imagined and which seemed so plausible has not occurred. Are
epidemiologists more resilient than you thought? Is the burden not as
burdensome as you thought? Is the burden real, but we are ignoring it
or getting around it somehow? Have epidemiologists changed their
thinking to now see the burden in a more favorable light, i.e. more
necessary or worthwhile?
Rothman: It is
possible that bureaucratic intrusion has increased, and that we have
just become more accepting of it. I am less in a position to know
about the red-tape epidemiologists face now than I was 20 years ago,
when I spent my time mainly doing research. Though I still do
research now I devote much of my time to writing and being Editor of
Epidemiology, and I don’t have the same perspective I did then. We
should also keep in mind that 20 years is a long time and people are
highly adaptable. It would be interesting to contemplate transplanting
a researcher suddenly from 1980 into the present and get a reaction.
Epi Monitor:
What about the burden for epidemiologists of having to live within the
bounds of various guidelines and good epidemiology practices
recommended by professional organizations or other groups? Has this
burden increased, decreased, or stayed the same as in 1980?
Rothman:
Fortunately, the guidelines for the conduct of research that have
appeared have been innocuous, even bordering on useful. As far as I
know they have been issued delicately, more as suggestions than
impositions. When they are not dogmatic, guidelines for scientific
work can be helpful, although they will probably never replace good
textbooks.
Epi Monitor:
Admittedly, your scenario for epidemiology in 2000 was
tongue-in-cheek, but still it is fascinating to reread your
descriptions of what would come to pass by 2000 to see how many things
have or have not actually occurred. For example, you “predicted” that
academic base of epidemiology would shrink. Also, you “predicted”
that the American College of Epidemiology would become the
professional union for epidemiologists. What do you say about these
today since neither one describes the state of affairs in epidemiology
today?
Rothman: The
academic base is larger than it was, but I suspect that epidemiologic
activities outside of academia have grown even faster. Industry and
government have acquired their own epidemiologic strongholds. The
pharmaceutical and health-care industries employ armies of
epidemiologists nowadays. I have even met lawyers who have studied
epidemiology. I am surprised how large epidemiology has become. It is
impressive that all the students that we train can manage to find
positions even as their numbers continue to increase. The growth in
epidemiology is most striking in western Europe and is beginning to be
a global phenomenon. The epidemiology profession still has academia as
its base, but there is plenty of activity outside it, as you yourself
know well.
As for the American College of
Epidemiology, in 1980 it was new and its direction was unclear. One of
the arguments for starting the College was to gain recognition, that
is, membership in the College for non-physician government
epidemiologists who were being slighted in pay and promotion. The
College was intended to play the role of a credentialing body. But
many epidemiologists at the time thought there might be a better way
to fix a job classification problem for federal employees than to
create a new professional organization. The SER held a debate about
whether starting the College was a good idea. I didn’t participate,
but I attended a packed session at an SER meeting and listened to the
arguments, pro and con.
All that is ancient history in
our profession. Today it is clear that the College has evolved to fill
an important role for epidemiologists. None of the divisiveness that
accompanied its founding has persisted. Rather than concentrating on
credentials, the College has become a high-quality scientific society
as well as the professional activist arm of epidemiology, stepping in
with authoritative lobbying and supplying the weight of considered
opinion on the very kinds of issues that motivated my article 20 years
ago. Those who asked why we needed another North American organization
beyond the SER found that a growing discipline had plenty of room to
accommodate both. Personally I am happy to be affiliated with both the
College and the SER, and I think both are vital organizations for our
profession.
Epi Monitor:
Having established your credentials as someone who thinks about the
threats to epidemiology, what are the most important threats that you
see today for epidemiology over the next 20 years?
Rothman: Our
most important threats for the next 20 years: HIV and emerging
diseases; environmental deterioration; the health effects of
overpopulation, poverty and war; smoking; and behavior-related
diseases. All these of course are not threats to epidemiologists, but
threats to public health. They are challenges for epidemiologists. As
for the original concerns that prompted my paper 20 years ago, the
potential for trouble is still with us. For example, there are
increasing and legitimate concerns about privacy in the Internet era,
concerns that may lead to a reaction that will impede epidemiologic
access to crucial data.
Epi Monitor:
For every existing threat that could cause a downfall in epidemiology
there probably is an equal number of opportunities which could advance
the field to new heights. What opportunities do you see as most
promising for the field over the next twenty years?
Rothman: Our
work evolves. An article in Science last week described how the entire
country of Denmark is becoming an epidemiologic laboratory. Registries
there link exposures with diseases to make an epidemiologist’s utopia.
Though it is not without its own bureaucratic problems, Denmark is an
example of an opportunity realized. Other countries in northern Europe
have had similar successes. In the US, computerization of large
patient populations along with the growth of prepaid health plans has
opened new avenues for epidemiologic research, overcoming many
hurdles. We have also seen the advent of enterprise epidemiology: the
large cohort study of a population followed for decades. Such studies
generate treasures of data on a range of problems that the study
designers cannot even imagine. And the biggest opportunity waiting to
be mined for epidemiologic research: the internet. It poses problems,
but also presents opportunities to track people easily and cheaply. I
am optimistic that the future of the profession is bright. I’m not
predicting a fall.
Epi Monitor:
Thank you for revisiting the state of epidemiology and for giving us
your views about epidemiology then and now. As we celebrate our 20th
birthday at the Epi Monitor, we also expect a bright future for
epidemiology and look forward to chronicling these positive
developments for our epidemiology colleagues worldwide. Let’s plan to
talk again in 2020 to see how close we come to getting it right!
Rothman: It
would be delightful to look back again with you in 20 more years. One
thing I haven’t mentioned is that in the past 20 years the Epi Monitor
has become part of our professional landscape, an interesting and
useful resource. I know the commitment that it takes to keep it going,
and I think you ought to be commended for the effort.
Epi Monitor:
Thank you. What a great statement to hear on our 20th birthday.
Published June 2000
|
