Thursday 3 November 2011

No miracle in Oregon

Jacob Grier has written an article in The Oregonian looking at the state's heart attack rate following its smoking ban of 2009. It will no come as no surprise to regular readers that the dramatic fall in heart attacks predicted by ban supporters has failed to materialise.

Two years ago, Ty Gluckman cited research from Americans for Nonsmokers' Rights founder Stanton Glantz showing that smoking bans led to a 17% fall in hospital admissions for acute myocardial infarction and concluded that "it's highly likely that Oregon's heart attack rates are already dropping as we near the law's one-year anniversary." He said:

If we reduce the number of acute heart attacks by 17 percent, there will be at least 1,100 fewer hospital admissions in Oregon in just one year. At a cost of more than $35,500 per admission, the savings will be substantial: $40 million.

"Highly likely" is right. Oregon's heart attack rate has been dropping for years. In the year before the ban, it fell by 6.67%. As Grier shows in his article, in the years since the ban, it has fallen by 7.21% and 3.11%—very much in line with the long-term trend.

Taken from http://www.jacobgrier.com/blog/archives/4887.html


Once again, routine hospital admissions data provide no evidence of a heart miracle. This is not surprising since the 17% figure comes from a meta-analysis of  studies which are riddled with data-dredging, retrospective cherry-picking and blatant researcher bias. That the myth of the smoking ban heart miracle has travelled the world and informed policy for the best part of a decade is a scandal that would be front page news in any other field of science.

As Grier notes, the Oregon non-event adds to the weight of evidence taken from countries around the world—including three peer-reviewed studies focusing on the USA—showing no positive effect from smoking bans on the heart attack rate.

Critics have dubbed these fantastical results "heart miracles," and, like most miracles, they proved too good to be true. When larger populations are examined, the effect diminishes or disappears entirely. The most extensive study to date was recently conducted by the RAND Corp., with a data set of more than 670,000 heart attack admissions from 26 states over a period of 11 years. It concludes that "smoking bans are not associated with statistically significant short-term declines in mortality or hospital admissions" for heart attacks.

The RAND study also explains how publication bias in favor of results and the large variations in smaller samples have combined to produce the illusion that bans are effective.

Gluckman has responded in the same newspaper, acknowledging that the figures Grier presents are correct while downplaying their significance. His defence is that many venues were smoke-free before the 2009 ban and so the effect of the state-wide legislation was less pronounced that it might have been. (This is the same excuse given by Anna Gilmore and colleagues (without evidence) when they failed to replicate the 17% drop claimed in the notorious Scotland heart miracle study.) It is strange that Gluckman did not mention this limitation when he was raving about the $40 million saving the state would make when it reduced heart attacks by 17%.

Now, at the eleventh hour, Gluckman concedes that heart disease is multi-factorial. He notes that major causes such as obesity may cloud the results. Indeed they could. It was always massively implausible that one minor variable—secondhand smoke exposure in bars—could have a large enough effect to show up in aggregate data, but this is the lemon that has been sold to the public around the world. We were told unequivocally that smoking bans produced declines in the heart attack rate of 17%, 40%, 50% and higher. Hospital admissions data have consistently shown this to be a fantasy.

Smoking ban campaigners have little choice but to back-track and make excuses in the face of reality, but there are two important things to remember.

The first is that this scam has only been exposed because it relied on data that are accessible to the public. The public will never be allowed to see the raw data behind the vast majority of what passes for science in journals like Tobacco Control. Light is the best disinfectant, but the bulk of the anti-smoking movement's policy-based evidence remains hidden in dark corners.

The second is that, although the fiction of the heart miracle is now virtually impossible to maintain, studies making impossible claims were published in peer-reviewed journals, including respected organs such as the NEJM and the BMJ for several years. Several meta-analyses (including two written by Stanton Glantz) have uncritically perpetuated the myth. And yet, whenever hospital admissions data are publicly available, they strongly indicate absolutely no impact on heart attacks from smoking bans.

There are lessons here about the perils of publication bias, advocacy-led science and the limitations of ecological studies. But in the end, it comes down to one simple fact — we have been lied to.

9 comments:

Anonymous said...

And lied to and lied to and lied to ad infinitum. When the truth is revealed, as it will be, the medical profession stands to experience the most far reaching revision of its role in society imaginable. It will be the just reward for calling yourself a scientist when what you have been is a paid shill for vested interests. Role on the future.

Carl V Phillips said...

Nice summary, Chris. And I commend especially your point about credibility that I will amplify a bit into: If this is what they claim about something that we obviously will be able to get better data on and debunk in the near future, think about the liberties they take with the truth when they know they can keep all the data hidden.

One correction: These were not ecological studies. Rather, they are best characterized as natural experiments. The structure of the data was never the problem -- it was the intentionally biased selection of which cases to publish and obfuscating statistical models. I would suggest removing "ecological studies" from your list of what this story condemns and adding "the fetishized peer review process".

Christopher Snowdon said...

Carl,

Thanks for your comment, which is insightful as ever. Re: ecological studies, most of these efforts are just natural experiments but a small number of them have bothered to use a control group eg. Pell compared Scotland with England (albeit using a totally different definition of a heart attack). When two or more countries (or states) are compared, that could be described as an ecological study, no?

Anonymous said...

If these studies show a lower heart attack rate due to less SHS/ETS exposure, why do they never show there being fewer never-smokers having heart attacks?

How can these reseachers ever control for the heart attack victoms that never go to the hospital?

About 1/2th of heart attack deaths occur outside of hospitals.

Gary K.

Anonymous said...

Graphology!!!

Also, the rates are all per 100,000.
About 200/100,000 in 2000 and about 140/100,000 in 2010.

That looks a lot bigger than saying the rate dropped from about 2/10ths of 1% all of the way down to about 1.4/10ths of 1%.

Gary K.

Anonymous said...

A few actual numbers done with a little research.

In 2008 there were about 5,895 admissions for heart attacks.

In 2009 there were about 5,530 admissions for heart attacks.

In 2010 there were about 5,364 admissions for heart attacks.

Sooo, in 2010, the antis are claiming that bans prevented a whole(whopping) 166 admissions for heart attack out of a population of about 2.85 million adults.

Gary K.

nisakiman said...

The problem we have here is that when Glantz et al put out their outlandish claims, the MSM lap it up, and make big front-page spreads in their respective publications about it.

However, when those claims are found to be false, the news is lucky to attract two column inches tucked away inside. Mostly, it will be ignored altogether.

It's no wonder that people believe all this crap. It's only on blogs like this that the truth ever sees the light of day, and with all due respect, Chris, not very many people read blogs like this, so the vast majority remain in ignorance of the facts.

Anonymous said...

And we'll never see a newspaper article that reveals the lie.

Becky Johnson said...

Here is another study where the complete set of admission data disprove the thesis of the study: that administering seasonal flu vaccines to minors results in fewer hospital admissions. This Mayo Clinic study showed vaccinated children were three times more likely to be admitted to the hospital later on...for ASTHMA!

(and they blame THIS on 2nd hand smoke too)

see:http://beckyjohnsononewomantalking.blogspot.com/2009/12/mayo-clinic-study-shows-children.html