Analysis of Doctors Study.

Doctors Study has been selected for this analysis because it is a well known survey. In addition, the causes of death in this survey have been collected very meticulously.

The survey was started in the 1950's by Austin Bradford Hill and Richard Doll. After 1964, Austin Bradford Hill no longer participated in the surveys, and was replaced by Richard Peto.

The first report was published in 1954, and was followed by reports in 1956, 1964, 1976, 1980, 1994 and 2004.

This analysis will focus on the mortality of cohorts from the surveys published in 1964, 1976, 1980 and 1994.

Alcohol and lung cancer.

First the analysis will concentrate on alcohol and lung cancer.

The 1964 report.

This report is very extensive. It was published in two parts, in seperate issues of The British Medical Journal. The first part of the report was published in BMJ 1964, 1, 1399 – 1410. The second part of the report was published in BMJ, 1964, 1, 1460 – 67. The report published the results after 10 years of surveys.

The report contains a plethora of tables that attempt to connect smoking, especially of cigarettes, with lung cancer.

Table 1. Excerpt of table 23, page 1406. Mortality rate per 1.000. There is some variation in the relation between mortality from lung cancer and cirhosis of liver, but for all smokers the relation is 6.45. Cirhosis of liver is a good predictor of occurance of lung cancer.

The report notes that there is a connection between smoking and alcoholism. Excerpt from page 1465:

“These two conditions [cirrhosis of liver and alcoholism] show a close association with smoking, most marked with cigarettes and particulary with the smoking of more than 25 a day. ... Several studies have shown that heavy drinkers also tend to be heavy cigarette smokers (Wynder et. al., 1956, 1957; Heath, 1958), ....”

The 1976 report.

This report was published in BMJ, 1976, Volume 273, page 1525 – 36. The report published the results after 20 years of surveys.

Table 2. All cohorts from table III and IV sorted according to mortality. Mortality rate per 100.000. Note that the relation between mortality from lung cancer and cirhosis of liver caused by alcoholism is approximately 6.5 regardless of type of smoking. For former smokers the relation is 4.3. For all participants it is about 6. Cirhosis of liver caused by alcoholism is a strong predictor of lung cancer.

Legend for smoking categories: smokers (5) – pipe/cigar smokers, smokers (6) – mixed smokers, ex (2) – former smokers, all (1) – current and former smokers, smokers (3) – current smokers, smokers (4) – cigarette smokers.

The 1994 report.

This report was published in i BMJ, 1994; Volume 309, page 901 – 911. The report published the results after 40 years of surveys.

Table 3. Excerpt from table V. Mortality rate per 100.000. The relation between mortality from lung cancer and cirhosis of liver caused by alcoholism is approximately the same regardless of smoking type. Former smokers have a relation of about 4. Cirhosis of liver caused by alcoholism is a strong predictor of lung cancer.

Conclusion for lung cancer.

Alcohol appears to be the dominant factor for development of lung cancer. There is a small effect from smoking. It can be determined with the formula:

(Relation smokers) / (Relation former smokers) x 100.0 %

From table 3 the effect on lung cancer from smoking can be calculated to be (6.5 / 3.8) x 100.0 % = 171 % or (7 / 4.2) x 100.0 % = 167 %.

The increase in risk for lung cancer from smoking is thus in the order of 70 %. This is eqiuvalent to an RR of 1.7. That is a very modest risk.

Alcohol as indicator of overall mortality.

In the following mortality from cirhosis of liver and overall mortality is examined to establish if alcohol is the dominant factor for overall mortality in the study.

The 1976 report.

Table 4. Excerpt from table III, page 1527. Mortality rate per 100.000. Liver cirhosis as indicator of alcohol consumption is a strong predictor of increased mortality.

Table 5. Excerpt from table IV, page 1529. Mortality rate per 100.000. Liver cirhosis as indicator of alcohol consumption is a strong predictor of increased mortality.

Table 6. All cohorts from table III and IV, sorted according to occurance of cirhosis of liver. Liver cirhosis as indicator of alcohol consumption is a strong predictor of increased mortality.

The 1994 report.

Table 7. Excerpt from table V. Mortality per 100.000. Liver cirhosis as indicator of alcohol consumption is a strong predictor of increased mortality.

Grafic presentation.

Figure 1. The numbers from table 6 and 7 shown graphically. It is quite obvious that there is a correlation between occurrance of liver cirhosis and overall mortality. Liver cirhosis an an indicator of alcohol consumption is a better predictor of overall mortality than smoking status. The information is from the 1976 and 1994 reports.

Verification with the 1980 survey.

In 1980 a report about a survey of the female British doctors was published. The report was published in BMJ, 1980, Volume 280, page 967 – 71.

Table 8. Excerpt from table III, page 969. Mortality per 100.000. Liver cirhosis is a considerably better predictor of overall mortality and occurance of lung cancer than smoking.

The cohort of cigarette smokers with a daily consumption of between 1 and 14 cigarettes per day is remarkable. The reason that this cohort of smokers fares so well is that there is a very modest level of alcohol consumption among those smokers. There is no mortality from liver cirhosis. This cohort of smokers is the cohort in the entires study, with the lowest mortality rate. This demonstrates how strong a killer alcohol is.

The numbers from the 1980 report confirm that alcohol is the decisive factor for overall mortality and occurrance of lung cancer.

Emphysema.

Visual inspection of the mortality rate from emphysema in the 1980 report demonstrates that emphysema is the most significant disease that smoking can cause. Close inspection reveals that alcohol must also be a factor in the development of emphysema, which is evident from the smokers with daily consumption of 15 or more cigarettes per day. It is unclear how strong the effect from smoking is, but it appears to be of the same order of magnitude as for smoking.

Emphysema is a disease of genetic origin. Smoking does not cause this disease, but accelerates the decay of the lungs that the disease itself causes.

Phantom effect from alcohol.

The apparant dose response from smoking is demonstratable with results from the 1994 report.

Figur 2. Excerpt from table V. This is an apparant dose response from smoking. In reality it is the phantom image of the smoking habits of heavy drinkers and alcoholics.

Estimate of the risk of lung cancer from alcohol consumption.

From the 1994 report it is possible to estimate the risk increase for lung cancer from alcohol consumption. In table III it is presented as if the relative risk (RR) for smoking and lung cancer is 19.5, i.e. a risk increase of 1950 %. That is of course the combined effect of smoking and drinking. The risk from alcohol consumption alone cane be derived from the formula:

RR Alcohol = RR smoking and alcohol / RR smoking = (19,5 ) / (1,7) = 11,5 i.e. 1150 %

Summary.

The smoking survey Doctors Study is a piece of magic science that inquires on smoking and registers the damages caused by alcohol consumption.. The effects of drinking alcohol are erroneously ascribed to smoking. There is a small effect from smoking which can be seen on lung cancer. There is also an effect from smoking on emphysema. It is unclear if smoking in itself can cause cancer.