Monday 21 January 2013

The English asthma miracle

The BBC has reported the latest heart-warming news about the English smoking ban...

There was a sharp fall in the number of children admitted to hospital with severe asthma after smoke-free legislation was introduced in England, say researchers.

A study showed a 12% drop in the first year after the law to stop smoking in enclosed public places came into force.

The implication is that this is the result of reduced secondhand smoke exposure. Of course, the smoking ban mainly affected places where children don't go, ie. workplaces, pubs and clubs, so the authors suggest that the smoking ban inspired people to make their own homes "smokefree" of their own volition.

The lead researcher, Prof Christopher Millett, said the legislation has prompted unexpected, but very welcome, changes in behaviour.

"We increasingly think it's because people are adopting smoke-free homes when these smoke-free laws are introduced and this is because they see the benefits of smoke-free laws in public places such as restaurants and they increasingly want to adopt them in their home.

"This benefits children because they're less likely to be exposed to second hand smoke."

Regular readers will know that studies of this sort have a tendency to be refuted by publicly available hospital admissions data. Actually, that's an understatement. In every single instance when the hospital admissions data are available, studies of this sort have not stood up. I'll discuss the mind-boggling methodology of this study (of which Stanton Glantz is an author) tomorrow, but for now here's the data from England's Hospital Episode Statistics which do not appear to support the claims being made.

Shown below are the number of hospital admissions for asthma (ICD code J45) since 2000/01. The figures shown are for the 0-14 year age group studied in the new study. The smoking ban began in July 2007. Click to enlarge.

Out of curiosity, I also looked at the number of asthma admissions for all age groups.

Regarding emergency admissions, it worth recalling the announcement from Asthma UK in 2010...

Asthma UK said the number of emergency admissions had remained unchanged for a decade

They were referring to Scotland in that instance, but that did not prevent claims of a Scottish asthma miracle being made. It may be that the situation in England is radically different from that in Scotland, and it may be that the trend in emergency admissions data differs radically from that of finished consultations. But, as we shall see tomorrow, it isn't.

(Please note that this post originally said the admissions were consultations. The graphs still do. In fact, the 0-14 year old graph shows admissions and the all ages graph shows emergency admissions. Thanks to BrianB for pointing out the error.)


ConceptTrader said...

When laying in bed early this morning with RADIO 5 on, I heard this report. The first thing that came into my head was how could this be the case when children do not go to all night house parties,clubs and pubs....Of course as I m sure you will show tomorrow, the claims are false.

Alan Buckingham said...

Here's the article:

The paper does not pass the most basic test of face validity: If there is such a powerful link, how can it be that between about 1960 and 2000, when rates of smoking were decresing, rates of asthma were increasing.

BrianB said...

Chris, it looks like you downloaded the Outpatient data from HES (consultations are outpatient visits).

Go to the Inpatient: Primary Diagnosis tables, and you will find age-specific numbers.

I downloaded the "4 character" tables so that I could look at "severe" asthma admissions (which is what all media reports of this study quote - somust be part of the press release). You can see that "severe" asthma (Status Asthmaticus) only accounts for about 10% of "all" asthma admissions. Someone is fibbing here!

I posted my own (sight-unseen) analysis at Frank Davis'place. How far out was my conclusion?

Unknown said...

Even if the data were correct it's a circular argument. Since the ban can't directly affect children much they suggest it has caused people to smoke less at home, which they say is the reason for less asthma, which "proves" people must be smoking less at home. And so on.

BrianB said...

Thanks for the link, Alan.

Interesting methodolgy. It takes data torturing to a whole new level!

The best bit, though, is that they try to "prove" their agenda-driven hypothesis by comparing actual admission rates with their own massaged predictions which they label "counterfactual".

But the Websters on-line dictionary defines counterfactual as "contrary to fact". Isn't this another way of saying "lie"?


nisakiman said...

Not only does 'second-hand smoke' not have the biological element to cause asthma, but as Dave Atherton argues here, there is evidence of a protective effect from tobacco smoke. And perhaps that partly explains why when I was growing up in the 50s / 60s when we were constantly surrounded by a fug of tobacco smoke, there seemed to be very little incidence of childhood asthma, at least in my personal experience. I don't know what the actual figures are, but I would wager that they were a lot lower than today.

Mark Wadsworth said...

I thought you wouldn't be able to resist.

One theory is that there was a spike in admissions in 2006-07, the number of admissions now is pretty much the same it was as ten years ago.

stat's here

Jean Granville said...

Only anecdotal, but I seem to have noticed a lot more "smokefree" private homes since a decade or two (not since the smoking ban).
As for the idea that it's better for health if there is no smoke around, that seems to me entirely possible.
The problem is that tobacco control people have been selling a link with cancer that is dubious, when there may very well be real links with other diseases not as bad as cancer.
So keeping some places "smoke free" is not a bad idea in general, as long as things are done in a reasonable way, which is never the case with these people.

In any case, if the smoking ban has made people want "smokefree" places, then it is no longer useful. They can repeal it now, right?

sandrajean said...

Could the spike in 2006 be due to the mid-summer heatwave that swept the UK and parts of Europe? It's a fact that hot, humid weather triggers asthma attacks.

Unknown said...

Why did it take an enforced smoking ban in public enclosed places for people to ban smoking in their homes ? Why didn't they make that decision before the smoking ban came into force ?

Alan Buckingham said...

I find it ironic that so much is being made of a trend over such a short period when similar claims of '15 year trend of no warming' by global warming sceptics are dismissed as extreme statistical naivete.

In the case of the asthma data, apply confidence intervals to it and any trend will disappear.

BrianB said...

Actually Alan, they did apply confidence intervals - not that there is any more credibility in these than in the rest of their crappy analysis.

The best bit is where they inserted a "dummary variable" into their 'model' to represent the legislative effect, with a value of zero for each year prior to the smoking ban, 1 for the first year after the ban, 2 for the second and so on. So, just as with climate models, they built the outcome into the model, ie a hard presumption that the smoking ban would indeed have an effect on admissions.

No surprise then that the numbers of admissions "saved" by the smoking ban was greater after this fudge was applied than it was before.

For the avoidance of doubt, BTW, the study did not claim that the numbers of admissions went down in absolute terms, rather that they were fewer in number than their model predicted they would have been (but for the saviour smoking ban).

It is such a pile of junk that it makes Mann's Hockey-Stick appear credible!

BrianB said...

Sorry "dummy variable"

I'm the "dummary"!

Alan Buckingham said...

Thanks for the analysis Brian. I wasn't able to open the article at work and so should have saved my powder until reading it!
What I now see as highly problematic is building a model around so few years prior to the ban, especially when those few years show an uptick in admissions which runs contrary to other datasets (e.g. GP consulation data for asthma) which show longer trends and a relatively long period of plateau/slight decline prior to the law.

Junican said...

"I moved to the UK from Australia in 1994 after obtaining a social science degree from the University of Sydney."

Christopher Millett.

Perfectly qualified to be a Tobacco Control Zealot, clearly.

BrianB said...


If you look at the numbers, recorded by the NHS (and reported through HES) you will see that they (the total number of admissions for under 15s) are all over the place during recent times.

Indeed there were 5 consecutive years (2004/05 to 2008/09 incl.) where the year-on-year % changes were in double digits. namely:

2004/05 +15.8%; 2005/06 -12.3%; 2006/07 +18.3%; 2007/08 -15.3% and 2008/09 +12.7%

Up, down, up, down up!

Now I would conclude that such a time-series is far too chaotic (and limited) to attempt to derive an 'average trend'. But that's what these charlatans did. They were careful to choose to use monthly data for their model, but in my view this was merely, and deliberately obfuscatory, as it added nothing to the overall analysis, and they still reported their results as annual changes.

No doubt Chris will highlight even more data torturing in his promised post for tomorrow. I look forward to it!

proglodyte said...

From DM article 2010

'Since the 1950s childhood asthma has risen three-fold and around 1.1 million of the 5.4 million Britons suffering from the condition are children.

Allergies are also on the rise and since 2001 the numbers of patients receiving treatment for some form of allergic reaction or intolerance has increased by 25 per cent.'

About twice the smoking rate 60 years ago?

Anonymous said...

Asthma Death Rates Are Lower in States With Higher Rates of Smoking. The states of Utah and California, which have the lowest rates of smoking at 13.0 and 17.1 percent of adults respectively, are also among the states with the highest death rates from asthma. (Asthma Deaths, 2000; and: Smoking Among Adolescents, 2001, and Smoking Among Adults, 2001. Centers for Disease Control and Prevention, 2003 State Health Profiles, Atlanta, GA: US Deparment of Health and Human Services, 2003.

The EPA's Sorry Status Report on Children and Asthma
"America's Children and the Environment. Measures of Contaminants, Body Burdens, and Illnesses," Second Edition, US EPA, Feb. 2003. EPA Administrator Christine Todd Whitman boasts that "This report marks the progress we have made as a nation to reduce environmental risks faced by childen," including "Implementing the Smoke-Free Home Pledge campaign, designed to protect millions of children from the risks of tobacco smoke at home." On pdf p. 75, "Between 1980 and 1995, the percentage of children with asthma doubled, from 3.6 percent in 1980 to 7.5 percent in 1995." The graph on pdf page 67 boasts of declines in cotinine levels during this same period.