I would like to make an investigation on smoking in numbers.
At this site i found your estimations: "Death rate is 2-3 times higher than non-smokers". I processed data given in http://www.pubmedcentral.gov/articlerender.fcgi?tool=pubmed&pubmedid=15213107 (R.Doll, R.Peto, Table5), and found the same relation between deaths of smokers and nonsmokers (~3 times).
But I see that these numbers (R.Doll, R.Peto) are not statistically enough - only 55 men deaths in group "nonsmokers, age 35-44".
Another data source, R.Peto et al ( http://www.ctsu.ox.ac.uk/~tobacco/SMK_All_PAGES.pdf ), is statistically enough, but R.Peto made an integration of deaths for ages 35-69. In some cases (for some countries), higher gap (69 yrs) is too big. I would like to find data such as R.Peto had, but with detailed age distribution, to make calculations as there in (R.Doll, R.Peto) performed.
Can you help me in my investigation?
I found also, that in death certificates ( http://www.cdc.gov/nchs/data/dvs/blue_form.pdf ) the field "35. Did tobacco use contribute to death?" exist and statistics on filled death certificates can give the data I want.
I think that half-year statistics on ~1 million deaths around any developed country would be very enough to gain exact(!) (not extrapolated or calculated by epidemilogical methods) data on tobacco-assisted deaths.