Re: A letter to a statistician. Emailed and airmailed and no reply

On Mar 30, 9:11 am, Ray Johnstone <r...@xxxxxxxxxxxx> wrote:
On Thu, 29 Mar 2007 09:26:47 -0500, David Winsemius





<doe_s...@xxxxxxxxxxx> wrote:
Ray Johnstone <r...@xxxxxxxxxxxx> wrote innews:emom03di9v20dgu9t9f313biisldniotj2@xxxxxxx:

http://members.iinet.com.au/~ray/DeVeaux.htm
snip
Chapter 13 is about controlled
intervention trials. Seven such trials have been conducted
to determine the association between smoking and health, with
a hundred thousand test and control subjects followed for seven
years - Whitehall, MRFIT, Goteburg, Finnish Businessmen, Oslo,
WHO collaborative and North Karelia. (References are available
on my website.) The results of all are uniform, forthright and
unequivocal: they show no association between smoking and life
expectancy, deaths from cancer or deaths from any other cause.

It is very difficult to cause smokers to quit. The intervention trials
generally have a very low success in modifying smoking rates.

No, they are generally successful, according to their authors. For
example:

"In conclusion we have shown that it is possible to apply an intensive
long-term intervention program against three coronary risk factors
with considerable success in terms of risk factor changes. The overall
results do not show a beneficial effect on CHD or total mortality from
this multifactor intervention." (Multiple Risk Factor Intervention
Trial Research Group, 1982)





Further,
since the incidence and mortality from lung cancer does not decrease
after cigarette smoking, the short-term (5 or 6 years typically) studies
all had very low power to detect an impact on that outcome. The relative
lung cancer incidence of continuing smokers to that of former smokers does
increase, but only because of the strong association of total dose with
incidence (and mortality).

These studies were NOT looking at the 'association of smoking and life
expectancy' and they were NOT "determining the association of smoking and
health"; they were looking at the association of smoking cessation
programs on differences in short-term event rates. s "unequivocal"

If you want to see the _association_ between smoking and life expectancy in
otherwise healthy individuals, you should look at current life insurance
estimates. The Society of Actuaries collects experience from multiple
companies and prepares a report on aggregate mortality outcomes. This is
observational data, of course, but it is an unmistakable _association_.

But whose data is it? Certainly not that collected by the US Bureau of
the Census. If it were then those paying the lowest rates would be
light smokers:http://members.iinet.com.au/~ray/SG8.htm





<http://www.actuary.org/life/cso/appendix_a_jun02.xls>

If Johnstone needs help on constructing a life expectancy estimate from
a mortality table, he need only ask for help here. You can see from
inspection of the select or ultimate one year q's that the mortality of
smokers is roughly twice that of the non-smokers.

The references from Johnstone's website:

12 Rose, G., P.J.S. Hamilton, L. Colwell and M.J. Shipley (1982), 'A
randomised controlled trial of anti-smoking advice: 10-year results',
Journal of Epidemiology and Community Health, 36, pages 102-108

13 Multiple Risk Factor Intervention Trial Research Group (1982),
'Multiple risk factor intervention trial: risk factor changes and
mortality results', Journal of the American Medical Association, 248,
pages 1465-1477.
MRFIT had multiple interventions. Subgroup analyses found that only the
successful smoking quiters had lower cardiac event rates.

One of the fundamentals of experimentation is that having set the
statistical design you don't change it after the data is collected.
Cherry-picking leads to spurious results.

<http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?itool=abstractplus&db=p...>

14 WHO European Collaborative Group (1986), 'European collaborative
trial of multifactorial prevention of coronary heart disease: final
report on the 6-year results', Lancet, 1, pages 869-872.

15 Wilhelmsen, L., G. Berglund, E. Elmfeldt, G. Tibblin, H. Wedel, K.
Pennert, A. Vedin, C. Wilhelmsson and L. Werks (1986), 'The multifactor
primary prevention trial in Goteborg', European Heart Journal, 7, pages 279-288.

If persons are looking for citations, they should use "Helsinki
Businessmen Study" in preference to "Finnish Businessmen" which was not
its name.

Neither name appears to have been used by the authors.>16 Miettinen, T.A., J.K. Huttunen, V. Naukkarinen, T. Strandberg,
S. Mattila, T. Kumlin and S. Sarna (1985), 'Multifactorial primary
prevention of cardiovascular diseases in middle-aged men: risk-factor
changes, incidence and mortality', Journal of the American Medical
Association, 254, pages 2097-2102.
See also:
<http://heart.bmj.com/cgi/content/abstract/74/4/449>

British Heart Journal, Vol 74, 449-454,
"CONCLUSION--The traditional risk factors (smoking, blood pressure, and
cholesterol) are significantly associated with 28-year mortality in this
high social class population with previous health education."

Read it again: their numbers belie their words.
Death rates
Intervention group 155.2/1000
Control group 106.6/1000
The death rate in the quit group was 45% higher than in the control
group. They give more details at:http://members.iinet.com.au/~ray/2consequences.html



17 Puska, P., J. Tuomilehto, J. Salonen, L. NeittaanmSki, J. Maki, J.
Virtamo, A. Nissinen, K. Koskela and T. Takalo (1979), 'Changes in
coronary risk factors during comprehensive five-year community programme
to control cardiovascular diseases (North Karelia project), British
Medical Journal, 2, pages 1173-1178.

You might want also to look at more recent data from that effort:
<http://www.ktl.fi/portal/english/osiot/research,_people___programs/he....>
Mortality changes in North Karelia in 1970-1995 (per 100 000, 35-64
years, men, age adjusted.

Rate in 1970 Change in 1970-1995
All causes 1,509 -49.00%
All cardiovascular 855 -68.00%
Coronary heart disease 672 -73.00%
All cancers 271 -44.00%
Lung cancers 147 -71.00%

It may be more recent data but it does not concern that trial: where
are the figures for control and intervention groups?

18 Leren, P., E.M. Askenvold, O.P. Foss, A. Fr¨ili, D. Grymyr, A.
Helgeland, I. Hjermann, I. Holme, P.G. Lund-Larsen and K.R. Norum (1975),
'The Oslo study. Cardiovascular disease in middle-aged and young Oslo men',
Acta Medica Scandinavica [Suppl.], 588, pages 1-38.

Johnstone has apparently not kept up with the literature in this area.
The Lung Health Study was published in 2005:
"The Effects of a Smoking Cessation Intervention on 14.5-Year Mortality:
A Randomized Clinical Trial" Nicholas R. Anthonisen, et al
<http://www.annals.org/cgi/content/full/142/4/233>

There may well be other trials like 7 I have described but this is
not one of them:
"Limitations: Results apply only to individuals with airway
obstruction."- Hide quoted text -

- Show quoted text -- Hide quoted text -

- Show quoted text -- Hide quoted text -

- Show quoted text -- Hide quoted text -

- Show quoted text -

What is your job Ray?

illywhacker;

.

Relevant Pages

  • Re: A letter to a statistician. Emailed and airmailed and no reply
    ... they show no association between smoking and life ... The intervention trials ... results do not show a beneficial effect on CHD or total mortality from ... health"; they were looking at the association of smoking cessation ...
    (sci.stat.math)
  • Re: A letter to a statistician. Emailed and airmailed and no reply
    ... they show no association between smoking and life ... The intervention trials ... since the incidence and mortality from lung cancer does not decrease ... health"; they were looking at the association of smoking cessation ...
    (sci.stat.math)
  • Re: A letter to a statistician. Emailed and airmailed and no reply
    ... mortality benefit of smoking cessation. ... To assess the long-term effect on mortality of a randomly ... intervention program and were compared with usual care participants. ... disease, lung cancer, and other respiratory disease. ...
    (sci.stat.math)
  • The Scientific Scandal of Antismoking
    ... Emeritus Professor of Mathematical Statistics ... Smoking and health have been the subject of argument since tobacco was ... Hill showed that smokers seemed more likely to develop lung cancer. ...
    (alt.support.stop-smoking)
  • Re: DOF Sept 4, 2005
    ... I'm sorry you have had so many health issues to deal with, ... > not smoking. ... > all the years, who had suffered hourly, trying to quit smoking at AS3. ... > for 3 weeks for a med adjustment in April as one of the new meds I was ...
    (alt.support.stop-smoking)