I have been staring at that graph and it gives remarkable information but there are two things I don't quite understand.
1. Why has someone joined up all the dots?
2. Would it be clearer as a table re-sortable so that either of the axes could be in the left hand column?
Probably, the linked points make no sense except to create dramatic peaks.
Have a look at the original data for the cholesterol levels and the rest of the paper. I actually can't understand why Kendrick didn't use the Monica data for mortality which are available. (not withstanding all the problems with Mortality data in that an other studies) It needs to be noted that for the total cholesterol measurement, four populations (Perth, Australia, Kaunas,
Lithuania, and
Ticino and Vaud/Fribourg, Switzerland) had major problems in the external quality control
http://ije.oxfordjournals.org/content/34/1/181.full#F1
I wonder what a graph comparing mortality and treated/untreated high cholesterol would show?
I'm not drawing it but have look at the treated cholesterol figures for Lithuania and Russia and Toulouse France.
My thoughts could be wrong but basically in Toulouse' a much higher percentage of those with high levels of cholesterol were treated, compared with nil in Lithuania and some parts of Russia.)
Note the countries where fewer people were regularly tested and where individuals knew their cholesterol levels.
Some geographical patterns in the frequency of cholesterol measurements emerged. The highest frequency was observed in France, Belgium, Germany, Italy, and Spain in both men and women, while the lowest were seen in Russia and Lithuania
And of those that knew they had high cholesterol
The highest awareness among people with hypercholesterolaemia was in populations in Spain, France, Belgium, Germany, Italy, and North America and, in men, also in Australia. The awareness was lowest in Lithuania and Russia
So might the high mortality rate for CVD be something to do with treatment of those with high cholesterol, rather to do with the actual prevalence of high cholesterol in the country ?
(Just an hypothesis!)
Areas where there people with high levels of cholesterol are neither detected nor treated have higher CVD mortality
But what about the Swiss figures, paradox perhaps??? (caveat on data)lowish treatment levels but also , according to Kendrick, lower mortality .
A 25 year follow up of the same individuals found that for individuals each CVD risk (obesity, smoking, high BP and high cholesterol was associated with a higher CVD death rate, all 4 together was associated with a 9.6 years shorter expected survival for a man aged 50.
http://www.ncbi.nlm.nih.gov/pubmed/22080525
But then again within country studies of both Lithuania and Russia show a rather different picture with areas of high incidencce of hypercholesterolaemia.
Diet and serum lipids: changes over socio-economic transition period in Lithuanian rural population
http://www.biomedcentral.com/1471-2458/11/447
Cardiovascular epidemiology in Lithuania
http://www.seminarsincardiology.com/pdf ... -68-75.pdf
And in Russia
Total Cholesterol and Mortality in China, Poland, Russia, and the US
http://www.annalsofepidemiology.org/art ... 47-2797(03)00388-0/abstract
Simple graph but not the whole story.
edit to fix transposed figure in link. ( thank you, to the person who p'm'd me about it... some people obviously read!)