Monday, 20 October 2014

Heart miracles: Is the truth emerging?

If there is one pseudo-scientific claim that illustrates the credulity of the media and the duplicity of the public health movement better than any other, it is the idea that smoking bans lead to dramatic reductions in heart attack incidence.

It is now ten years since the British Medical Journal published Stanton Glantz's notorious 'Helena Miracle' study which claimed that the heart attack rate fell by 40 per cent after a small town in Montana banned smoking in pubs and restaurants. Numerous copycat studies followed, typically involving thinly populated towns and regions which, because of the small number of heart attacks that take place each month, are given to large fluctuations in hospital admissions.

From the outset, the most plausible explanation for the heart miracle phenomenon was that activist-researchers were scouring hospital records for unusual declines in heart attack admissions that roughly coincided with 'smokefree' laws. With so many smoking bans being enacted, it was inevitable that they would coincide with a blip in admissions now and again.

But when whole nations bring in smoking bans, the rate of decline has typically been zero or in the low single digits, ie. in line with the long term trend. (The most notable exception was a study of Scotland which claimed a 17% decline—a finding that is totally inconsistent with official NHS data.)

Having written about this for the five years, I was pleased to see some sanity rear its head in the American Journal of Medicine in January. A study by Basel et al.—which I have only just become aware of it thanks to Klaus in Denmark—looks at rates of acute myocardial infarction (heart attacks) in Colorado after a statewide smoking ban went into effect in 2006. This is of particular interest since two widely touted heart miracle studies have involved small pockets of Colorado. A 2006 study of Pueblo, Colorado claimed that there was a 27% decline in heart attacks when it went 'smokefree' in 2003 and a 2006 study of Greeley, a small town in Colorado, also claimed a 27% decline.

The researchers looked at the data for the whole of Colorado before and after its strict statewide smoking ban came into force. They looked first at total admissions for acute myocardial infarction and then they excluded the eleven towns and counties that already had smoking bans in place. In both instances, they found no effect from the ban.

We did not observe a significant decrease in acute myocardial infarction hospitalization rates in Colorado after enactment of a comprehensive statewide smoking ordinance. Even after removal of geographic regions where preexisting smoking ordinances were under enforcement, no statistically significant reduction in acute myocardial infarction hospitalizations was detectable. This contrasts with a number of prior studies, including local smoking ordinance studies in Pueblo and Greeley, Colorado, and adds to a growing literature that the cardioprotective effect of smoking bans may be less than initially suggested.

This finding is important and telling, but the study is also worth reading for its discussion of the existing literature. It is clear that heart miracles are confined to small, obscure towns in a way that can only be described as suspicious. (I have inserted hyperlinks to each study mentioned below.)

Overall, a review of published research shows that acute myocardial infarction RR reduction appears inversely related to sample size. For example, small studies in Bowling Green, Ohio, and Helena, Montana, found dramatic RR reductions (39% and 40%, respectively) but also had few acute myocardial infarction counts (58 acute myocardial infarctions in Bowling Green, 64 acute myocardial infarctions in Helena) and relatively small study populations (30,052 and 68,140, respectively). Studies in Greeley and Pueblo, Colorado, and Graubünden, Switzerland, found less dramatic RR reductions (27%, 27%, and 22%, respectively), corresponding to somewhat larger study populations (∼86,000, 147,751, and 188,000, respectively).

As the authors note, these large declines in small communities (which are not just implausible, but mathematically impossible), contrast sharply with evidence from large communities and whole nations. national study used Medicare Provider Analysis and Review files and national death records; a nonsignificant reduction in acute myocardial infarction-related (RR, −4.1; 95% CI, −9.4 to 1.3) and all-cause (RR, −0.7, 95% CI, −2 to 0.6) mortality was observed 1 year after smoking ordinance enactment. In this study, researchers evaluated all possible pairs of ordinance and nonordinance hospitals and recorded the change in acute myocardial infarction incidence post-ordinance. They found that RR reductions of 10% or greater were common, but that RR increases of 10% or greater were equally as common; taken in aggregate, the mean was near zero.

Another study examined 74 cities geographically distributed across the United States that were affected by smoke-free legislation. Individual cities showed wide variation in acute myocardial infarction incidence after ordinance enactment, with risk ratios ranging from −36% to +54%; however, the mean risk ratio for the 74 cities was 0.97 (95% CI, 0.96-1.02).

... A study performed in Christchurch, New Zealand after a countrywide smoke-free ordinance, found a 0% RR reduction in acute myocardial infarction with an approximate population size of 350,000. Countrywide studies with larger population bases provide concordant findings. In England, a 2.4% RR reduction was observed (population of 50 million). In Italy, a 4% RR reduction was observed (population of 58 million). In France, a 0% RR reduction was observed (population of 63 million). Finally, in a study examining the US Medicare population in states with a smoke-free ordinance versus those without, a 0% RR reduction was demonstrated (population of 30 million).

In the case of the English study, the heart attack rate fell at exactly the same rate after the smoking ban as it had been doing before the smoking ban. After dressing this up with some superficial computer modelling, Anna Gilmore—for it was her—relied on nothing more than a post hoc ergo propter hoc assumption. A similar claim, though never published, was made about Wales.

The authors attribute much of the heart miracle phenomenon to publication bias. That is likely to be a part of it, although I think that researcher bias and selection bias played more of a part.

These analyses support the hypothesis that small study populations may be more likely to find dramatic changes in acute myocardial infarction incidence, whereas increasing the study sample size attenuates the magnitude of the reduction. Also, review of the studies in aggregate reveals data asymmetry that suggests the potential for publication bias or heterogeneity not entirely explained by a random-effects meta-analysis. The presence of publication bias may explain why small sample size studies have tended to report large decreases in acute myocardial infarction incidence, whereas relatively few small sample studies have shown no effect.

The whole heart miracle scam has, in my view, been built on two simple tricks:

Firstly, dredging the data for any town that saw a large decline (in percentage terms) in heart attacks at around the time of a smoking ban. Nobody decided to do a study of Helena, Montana or Bowling Green, Ohio before the bans took place. The decision to focus on such obscure places came about only once it was clear that they were anomalous (not unlike Derren Brown's horse-racing trick). They were then presented to the media with the implication that they had been randomly selected.

Secondly, although less frequent, studies of larger populations have portrayed rather small declines in the heart attack rate as being the result of a smoking ban, without acknowledging that that there had been a secular decline of the same magnitude long before the ban was enacted. As the authors of the above study note, the secular decline is simply ignored in such cases.

That's really all there is to it. The 'public health' lobby has been selling this lemon to the public for ten years while describing sceptics, such as Michael Blastland (the creator of BBC's excellent More or Less series), as 'denialists' and 'dissidents'. The American Journal of Medicine study won't be enough to set the record straight in the public's mind—it received no media coverage, naturally—but it is further ammunition for those who do not believe in the 'noble lie'.

1 comment:

Chris Oakley said...

This will receive no media coverage for the same reason that the junk you reference has. The media is populated by generalists who are utterly clueless and rely on press releases for "news". Every piece of "research" supporting a public health industry claim, however dishonest and ridiculous is supported by a barrage of press releases. Every study that exposes the ridiculous nature of public health industry claims is ignored. The net effect is that the public is grossly misled. The BBC is especially guilty of "evidence by press release". Over 60% of the UK public relies on the BBC for news. There is no excuse for misleading them. Seeing through the public health industry charade is not that hard. The BBC prefers not to.