MPH Master pieces

Funnel plots, outliers and smoking

I don't often blog the Journal of Public Health but a paper in the latest issue caught my eye - almost literally: there is an interesting graph of death rates from cancer in the 100 or so PCTs of England. The data are displayed as a funnel plot with age-standardised death rate on the 'y' axis and number of deaths on the 'x' axis. Doing it this way allows you to put a statistical outlier funnel on the map which narrows as the numbers get bigger.

In the first graph there is a group of PCTs which are outliers. Now comes the clever bit - if you take out lung cancer none of them are outliers. So these high death rate PCTs owe their unenviable position to smoking. The picture is somewhat different in females.

You could have done all of this with tables of data but it's the graphic which makes the story compelling.

While on the subject of images, here is a sorry tale. A coffee creamer, for reasons not entirely clear, uses on its packaging a graphic of a bear nursing a baby bear. All very cuddly but the effect of the picture is that 18% of women in Laos have used it as infant food. Analysis of images is known as semiotics, an important discipline because as this and other examples show, images often override words. (Which means you should go back to those cancer graphs and make sure the underlying analysis is fully rigorous.) 

Urban greenspace

Here's a nice piece about parks in cities (Ontario to be precise). You might have thought that proximity was the most important thing - the nearer the better. Not so! and the results, which I leave you to discover, carry a clear message to urban planners.

I like the use of a formal assessment measure (the 'Environmental Assessment for Public Recreation Spaces' instrument).

One disappointing feature of the study was the response rate - 61%. Part of the problem was that householders had to complete activity diaries, to be collected after 10 days by the researchers. In a single-visit survey, with excellent interviewers, I'd be looking for a response rate of over 80%.

Public policy

Policy analysis is an important part of public heath - thinking through the likely effects of government (and other) policies.

This thoughtful piece by Clare Bamra looks at the likely effects of a change in rules for incapacity benefit in Britain. She points out that the policy signifies a shift towards viewing people as either 'deserving' or 'undeserving' of state support. If unable to work, a higher allowance will be paid. If judged 'sick but able to work' the full allowance is only payable to people who are participating in various back to work schemes.

As she says, this type of scheme will shine an interesting light on our attitudes to various sorts of mental illness (is it illness or laziness?), and also on the role of doctors in state certification systems.

Obese children and fat adults

The title of this paper from Japan pretty much says it all. The study allows us to say that a generation of obese children may grow up to be a generation of slim adults, which is good news.

Note that the method is repeat surveys (and hence different sets of respondents) over 60 years. So we cannot use this study to give a prognosis to individual children who are obese. That would require information from a cohort study, following the same individuals for 40 years or more: the British Birth coghort studies have done this.

Anyway all those gloomy forecasts for our current generation of children of a tide of obesity, type II diabetes, fatty liver and so on are not inevitable - but equally the growing up into slim adults will not just happen by itself: we have to make it happen with effective public health action.

Dangerous sandwiches

Here is an interesting account of an outbreak of listeriosis in a hospital. This sort of outbreak is probably only recognised in the laboratory. Listeriosis isn't common but the risk is more than just theoretical.

Bullous pemphigoid - on the increase?

Here's an interesting exchange.

Langan published a piece based on a GP database showing a large increase in the number of people diagnosed with bullous pemphigoid.

Faced with a striking and unexpected finding such as this, the correct scientific response is to test the question in other ways. A team from the St John's Institute of Dermatology looked at their lab data over the past decade. Crucially, they make the point that almost all laboratory diagnoses of pemphigoid in the south east of England would pass through their lab. Result - no change.

Langan et al assert (reasonably) that no GP would diagnose pemphigoid without confirmation by a dermatologist but even so I can't help feeling that the answer lies in the fact that Langan et al used a GP database.

Attributable deaths - problem drug use

Here is a nice demonstration of how to calculate attributable risk fraction, also known as aetiological fraction. That is, how much of this disease is caused by that factor? All comments such as '90% of lung cancer is caused by smoking' or 'deprivation kills 5000 every day' are founded in this type of calculation. 

This paper impressed me with its bold conjecture* : a single factor accounts for much of the (striking) difference in mortality between Scotland and England. That factor is drug abuse, and it accounts for 30% of the difference. Note that differences in deprivation were greater in 1981 than 1991.

Now we need a bold conjecture* to explain the French - English difference. Not garlic or red wine, by the way.

* The stuffy name that we should properly use for a bold conjecture is 'interesting hypothesis'.

Cervarix or Gardasil?

The choice of papilloma vaccine for the national programme was bound to cause argument between the suppliers. The UK opted for Cervarix, based on some mathematical modelling of the likely benefits. The criteria for choosing the vaccine were made explicit (an example of option appraisal) and were made public after a parliamentary question.  

I suggest that you ease your way into this by reading the editorial before the article.

Melanoma and sunlight

Although a key principle of teaching is to make things simpler, sometimes they just ain’t. So here is an interesting exchange on the question of whether sunlight causes melanoma.

 

Menzies says yes, Shuster says no. Unfortunately the BMJ debate format is light on references so it’s not clear to me whether Menzies and Shuster are arguing about the facts or about how they should be interpreted: I think it’s about interpretation, which means you can read the arguments and make your own mind up. Shuster has been outspoken on this for many years - I guess that's why the BMJ chose him for this exchange.

 

One key issue is whether melanoma is commoner on areas not exposed to sunlight. I guess we could all agree that sunlight doesn’t give you melanoma on the sole of your foot: as Burd points out from Hong Kong: "Although this accounts for only 2% of melanomas in white people, it accounts for more than 50% in our local Chinese population, and I have yet to see anyone sunbathing in Hong Kong with their feet in the air."

Another question is whether melanoma has genuinely become more common.  If not, the ecological correlation with increased sun bathing is less persuasive. I'm inclined to the view that what we call ‘melanoma’ may have changed over the past 20 years. This for two reasons.

 

Firstly, we know that the more you look for a particular cancer, the more you detect hitherto unnoticed forms of the disease. We know this famously from neuroblastoma but it seems to apply even in breast cancer (Gotzsche PC, Nielsen M. Screening for breast cancer with mammography. Cochrane Database Syst Rev 2006(4):CD001877 ).

  

Secondly, I went to a lecture about 10 years ago by a pathologist who ran a WHO melanoma reference centre. It was obvious that deciding on the basis of a microscope slide of dead tissue whether or not a cell was likely to invade and metastasise – hence a melanoma not some more benign cell abnormality - was very much a matter of judgement, not a black-and-white matter of fact.

 

The exchange was followed by some correspondence. Notice some subtext: Shuster receives support from his former workplace Newcastle (the English one). It is going too far to call this ‘groupthink’ but clearly schools of thought develop in particular localities. How is your thinking affected by your immediate colleagues?

Hand washing

It's amazingly difficult to reference some of the most obvious principles of public health, perhaps because in some cases the research was done so long ago. So it's nice to see a systematic review of handwashing confirming what your mother taught you. 

Soap seems to be important, which allows me to mention a 5-star account of a social marketing campaign in Ghana. Research had shown that mothers were doing well in washing their hands, but needed to be persuaded to use soap each and every time. The impressive thing about the campaign was the focus and consistency they achieved, perhaps because of the involvement of professional marketing people.