policy research. My experience from inside this peer review
process fits precisely with the authors' external observations.
Most section members were behavioral researchers, with ex-
pertise in studying and modifying tobacco use behavior. This
area of study and intervention constitutes less than half of current research and practice in tobacco control: the larger remainder includes studies of relevant tobacco industry and government behavior, innovative interventions, and the development
and analysis of comprehensive government policies. To appreciate the range of concerns and methodologies reflected in state\x=req-\
of-the-art tobacco control work, one need only peruse any issue
of Tobacco Control: An International Journal (including the
special supplement on tobacco control research), any of the fre¬
quent issues of JAMA or of the American Journal of Public
Health devoted to tobacco control, or the proceedings of any of
the recent world conferences on tobacco and health.
Though all of the Behavioral and Public Health Research
Study Section members were thoughtful scholars learned in
their fields, they were simply not the professional peers of most
scholars working in the tobacco control arena. Aside from the
limited intersections of their behavioral interests with tobacco
control concerns, most had not contributed to the tobacco con¬
trol literature or participated in tobacco control conferences.
Unfortunately, when it comes to evaluating tobacco control
proposals, the irrelevance ofthe section members' experience
and expertise mattered. One year, a tobacco control proposal
reviewer announced that s/he did not believe in this type of
research and was accordingly giving it a very low rating. The
following year, the primary reviewer of a tobacco control
proposal succeeded in eliminating one study from the pro¬
posal with a simple observation to the effect that s/he didn't
like this type of study.
My impression both years was that subjective comments
like these were generally not offered and almost certainly did
not carry the day with respect to the many behavioral in¬
tervention proposals. As the sole section member with pro¬
fessional experience in tobacco control policy research, my
ability to respond to the criticisms of reviewers who don't do
this kind of research, and don't like it, in a way that would be
persuasive to a group of "peers" who also don't do this kind
of research and have very little interest in it, was very lim¬
ited. The result, however unintentional, was a substantial
degree of bias against most kinds of tobacco control research,
regardless of their quality.
Richard A. Daynard, JD, PhD
Northeastern University
Boston, Mass
1. Glantz SA, Bero LA. Inappropriate and
review. JAMA. 1994;272:114-116.
appropriate selection of 'peers' in grant
To the Editor.\p=m-\Intheir article, Drs Glantz and Bero1 suggest
that Glantz's initial research applications on tobacco control
policies received poor priority scores from two study sections
"because their members were not peers." They used this
limited experience to launch an unsubstantiated critique of
peer review policies. As a member of one of these study
sections (Agency for Health Care Policy and Research, Health
Services Research Dissemination [HSRD]), but not a primary or secondary reviewer of the application, I wish to point
out some serious flaws and objections that bias the article's
conclusions and undermine its credibility.
For the record, the reason the initial application that we
reviewed received a 92% priority score had nothing to do with
the composition and interests of the study section. The main
point, which the authors fail to mention, is that the methods
section was only a sketchy description of an interesting, but
complex, study. Incomplete protocols are rarely scored well
by any federal study section on any topic.
The article further implies that a letter accompanying the
revised application describing the "problems with the peer
review [the proposal] received" was related to the study
section's subsequent favorable review. In fact, the executive
secretary appropriately excluded this letter from the revised
application sent to the study section. This application fared
well the second time because the investigators provided more
details on their methods and responded in detail to the study
section's concerns described in its initial review.
The data on publication rates presented as evidence of inad¬
equate peer review are also severely biased. The case-control
analysis compared publication rates on "tobacco" by members
of a general health services research study section with pub¬
lications indexing "heart" by a cardiovascular study section.
Obviously, this is an inappropriate comparison group. Unlike
the disease- or organ-based emphases of some National Insti¬
tutes of Health study sections, HSRD emphasizes health
services/policy research and dissemination across a range of
disciplines, including consumer, clinician, and organizational
behaviors related to health. Taking the authors' arguments to
the extreme would require our study section to have experts
on every significant health hazard. Clearly, this is inappropri¬
ate, given the importance of representation of disciplines such
as the behavioral sciences, biostatistics, and health economics.
In fact, members of the study section were highly interested in
the subject matter of the application and expert in many of the
general domains it encompasses. Nevertheless, collateral re¬
viewers with specific expertise in tobacco control policy were
also invited to evaluate the application; they reached similar
conclusions regarding its potential significance pending sub¬
stantial revision and additional details.
In summary, the article's biased critique of current study
section review procedures (in the absence of any verifiable
data on these processes) calls into question the appropriate¬
ness of basing an evaluation of peer review on one interpre¬
tation of a single case.
Stephen B. Soumerai, ScD
Harvard Medical School
Boston, Mass
1. Glantz SA, Bero LA. Inappropriate and
review. JAMA. 1994;272:114-116.
appropriate selection of 'peers' in grant
In Reply.\p=m-\Thestatement by Dr Gruder and colleagues that tobacco policy applications were funded by the TRDRP at a higher
rate than applications as a whole is misleading. By our count, in
the last review cycle only two tobacco policy grants were funded
(out of a total of five submissions), and one of these was a grant
application submitted by Dr Bero, which had the policy components of the proposal deleted by the study section. This compares with 90 funded grants out of 391 applications for nonpolicy
grants. One of our points was that the inappropriate selection
of peers discouraged submission of policy grants.
We were, nonetheless, surprised that Gruder et al criticized our article, because he and his colleagues at TRDRP had
already taken affirmative steps to remedy the problems in
the peer review process we identified. In particular, last year
TRDRP convened a new study section to review tobacco
policy grants that was made up of appropriate reviewers.
Unfortunately, we do not yet know the outcome of this
process because the California legislature gutted the TRDRP.
Likewise, the National Cancer Institute also appears to
have taken our criticisms to heart, as opposed to the view
advanced by Dr Soumerai, by establishing a special study
section to review recently solicited proposals on tobacco policy
research, rather than relying on existing study sections.
Tobacco is the leading preventable cause of death in the
United States today, and public policy interventions are the
most effective way to deal with this problem. We are en-
Downloaded From: http://jama.jamanetwork.com/ by a Florida International University Medical Library User on 06/15/2015