In June of 2002 a smoking ban was implemented in the town of Helena,
Montana. Six months later the ban was lifted due to a judicial ruling.
In April, 2003, two researchers made an astounding claim:
the ban had lead to an immediate 60% drop in heart attacks. They further
claimed that when the ban was lifted the number of heart attacks returned
to normal levels.
This study has received widespread distribution
and acceptance, and has been used to"prove" that bans
improve public health. But it doesn't take much digging to see
that the study was a complete fraud; a shameful exercise in junk science.
Fact: In 2002 the city
of Helena, Mt. passed a ban on smoking for all business. Six months
later the ban was overturned by a judge, and smoking was once again
allowed in public places.
Fact: In 2003 Dr. Richard
P. Sargent and Dr. Robert M. Shepard announced to the press they had
done a study which showed a 60% decrease in acute myocardial infarctions
(heart attacks) during the period the ban was in place.
are usually announced when a study is published. "Science
by Press Conference" is not highly regarded in the scientific
Fact: When the study was published in the British Medical
Journal, a year later, they claimed a 40% reduction in heart attacks.
Why did they give one number to the press, and an entirely different,
much lower number to the BMJ?
Their original PowerPoint presentation is
available here (pdf).
Note the 60% claim on the last page.
Fact: Sargent and Shepard
enlisted the aid of Dr. Stanton Glantz (whose doctorate is in applied
mechanics, not medicine or biology), an active and
prominent leader of the anti-smoker movement in the US.
Fact: BMJs web site includes
a "Rapid Responses" section,
where respondents can comment on a study. Dr. Glantz posted there,
ignoring all of the criticisms and questions that had been
posted. He said their study "reported
a 42% (95% CI 1% to 79%) statistically significant drop in hospital
admissions for acute myocardial infarctions (AMI)"
Fact: The confidence interval
cited by Glantz does not appear any where in the original study.
Studies with wide confidence intervals are
suspect. Studies where the CI is close to 1.0 are very suspect. (See
Statistics 101 for more details.) Assuming the CI Glantz cites is accurate,
one less heart attack or one misdiagnosis could have pushed the lower
boundary of the CI below 1.0, which would have made the results statistically
Fact: The researchers
claimed that the isolated community, which resulted in a very small
sample size, was a positive factor.
Fact: In statistical studies,
large sample sizes across various populations are usually considered
Fact: Compliance with SHS laws
is an important factor in SHS exposure.
Fact: In the study the
researchers reported, "We did not make any direct observations
to measure how much exposure to secondhand smoke was reduced during
the months when the law was in force. We do not know the prevalence
of smoking in venues covered by ban, though the city-county health
department reported that all but two businesses complied."
Fact: On April 7, 2003,
Sargent and Shepard attended a meeting of the Montana Senate where
they heard testimony from business owner Laura Fix that smokers were
driving outside town to patronize venues that allowed smoking. The
minutes from this meeting are available here.
Fact: At the same meeting
Rich Miller of the Gaming Industries Association, testified that three
of Helena's five casinos never complied with the ban.
Fact: It was well known
that some of the casinos were ignoring the
ban, and many bars (at least nine according to this article) were
not in compliance.
Fact Worth Repeating: Sargent
and Shepard were present at the meeting where the widespread non-compliance
was discussed, yet still claimed, in their study, that only two business
were non-compliant. At the same meeting they learned that smokers were
making the short trip out of town to smoking venues, but ignored that
fact in their study.
Fact: 38% of the patients
in this study were smokers, 29% were ex-smokers, and 33% were life
long non-smokers. This was mentioned in the study, but was not taken
into account or used to adjust the final numbers.
Fact Worth Repeating: Only a third of the patients in this study were
Fact: Although the study claimed to be about the effects of second
hand smoke, the researchers did not interview a single patient about
their SHS exposure.
Not one! If they wanted accurate information, wouldn't they
have interviewed as many patients as possible? There were only
40 of them.
Fact: A similar dip in
heart attacks occurred in 1998. The researchers ignored this
Fact: No confounders were considered, checked for, or used to adjust
In any legitimate statistical study, confounders must be accounted
for and used to adjust the data. (See Statistics
101 for more information.)
In this case, confounders would have included the patients smoking
status (smoker, ex-smoker, never smoker) exposure to any conditions,
substances, physical activity or environments that could cause or contribute
to heart attacks, and recent changes to those exposures.
Fact: The authors of this study claimed they had no competing interests.
Fact: Richard Sargent
and Robert Shepard were anti-tobacco activists before conducting this
study. When asked about this Shepard said, "They
can say what they like. The bottom line is, the data speaks for itself.
We don't have to apologize for our particular bias." (Emphases
Fact: Stanton Glantz is
one of the leading anti-tobacco and anti-smoker activists in the United
States. He has brought millions of dollars of anti-smoking grant money
to the University of California, San Francisco, which eventually awarded
him a chair as Professor of Medicine despite his lack of a medical
degree. He is the founder of Americans for Non Smokers Rights, Smoke
Free Movies and has personally received hundreds of thousands
of dollars in grants for anti-tobacco studies. In addition, he has
generated tens of millons of dollars in anti-smoking grants for his
Fact Which Shows Fraud: Although
the researchers claimed heart attacks dropped for the six months
the ban was in place, the only drop was during the first three months. For
the last three months the rate returned to normal. This can be seen
in their own chart.
Fact Worth Repeating: "We
did not make any direct observations to measure how much exposure to
secondhand smoke was reduced during the months when the law was in
Fact: All of the funding
for this study came from anti-smoking organizations.
Fact: The American Cancer
Society awarded Sargent Shephard the Ted Marrs Award, which recognizes
individuals that consistently demonstrate excellence in the area of
public issues. They cited the 60% claim from their original press conference,
not the 40% claim made in the actual study.
Responses to the Study
The following comments
about the study were published in the BMJs Rapid
Response page. (Note: Some of the original comments were quite long,
and have been edited for space reasons. You can see the responses
in their entirety here.)
Quick Hitts Blog.
Listen to the
Quick Hitts Podcast.
These numbers are far too
small. The study surveyed at most 40 admissions across a six month,
or 26 week period. So, even an increase/decrease of one per week
makes a huge impact. It is also claimed that the increase of 5.6 in
outside Helena admissions was insignificant, however it is a near 50%
- Ben Hirsch, Research Assistant
University of Massachusetts
The attempt to make claims about the effects of smoking bans based
on this very weak ecologic study raises disturbing questions about
our ability to distinguish between sound science and wishful thinking.
- Geoffrey C Kabat
This is, in my opinion, gross misrepresentation designed to provide maximal public impact in furthering the biased and unscientific opinions of these authors. . . This so called study does not even come close to meeting the basic criteria of a properly executed scientific study.
- Henry F Mizgala
Emeritus Professor of Medicine
of British Columbia
Smoke-free workplace legislation protects workers
and the public from cardiac, respiratory, and cancer risks associated
with second-hand smoke and facilitates cessation among smokers. However,
unrealistic expectations can hinder efforts to expand smoke-free workplace
laws. The 40% decline in acute myocardial infarctions (AMIs) associated
with a smoke-free ordinance in Montana is neither biologically nor
- Farzad Mostashari, MD MSPH,
NYC Department of Health and Mental Hygiene
We previously presented data showing
that the Helena observations are consistent with random variation because
of the small number of observations on which they are based
Professor, Department of Medicine and Endowed Chair,
Tobacco Harm Reduction
The entire study is available here.
The Smokers Club Inc. has gathered hundreds of articles about the
Jacob Sullum responds to
the preliminary announcement, and analyzes the numbers.
Michael Siegel, a tobacco control advocate, analyzes the study.
Steven Milloy tries, unsuccessfully, to get more information out of the authors.
An interview with Dr. Sargent
What happens when you examine the effects of bans on large populations?
study did just that. This is
a preliminary version of their study. Trying
to get it published has
been, shall we say, an adventure.
Many thanks to Michael McFadden (author of Dissecting
Anti-Smokers Brains) and Dave
Kuneman who provided an enormous amount of help with this page.
They provided many of the facts and did the research that made this