Some years ago, I thought to help my daughter understand statistics by reanalyzing the data from a 2004 study on coffee and Parkinson’s disease mortality, “Coffee consumption, gender, and Parkinson’s disease mortality in the cancer prevention study II cohort: the modifying effects of estrogen” , Am J Epidemiol. 2004 Nov 15;160(10):977-84, see it here

For the study, a cohort of over 1 million people was enrolled in 1982 and assessed for diet, smoking, alcohol, etc. Causes of deaths were ascertained through death certificates from January 1, 1989, through 1998. Death certificate data suggested that coffee decreased Parkinson’s mortality in men but not in women after adjustment for age, smoking, and alcohol intake. They used a technique I didn’t like though, ANOVA, analysis of variance. That is they compare the outcome of those who drank a lot of coffee (4 cups or more) to those who drank nothing. Though women in the coffee cohort had about 49% the death rate, it was not statistically significant by the ANOVA measure (p = 0.6). The authors of the study understood estrogen to be the reason for the difference.

I thought we could do a better by graphical analysis, see plot at right, especially using R² to analyze the trend. According to this plot it appears that coffee significantly reduces the likelihood of death in both men and women, confidence better than 90%. Women don’t tend to drink as much coffee as men, but the relative effect per cup is stronger than in men, it appears, and the trend line is clearer too. In the ANOVA, it appears that the effect in women is small because women are less prone Parkinson’s.

The benefit of coffee has been seen as well, in this study, looking at extreme drinkers. Benefits appear for other brain problems too, like Alzheimer’s. It seems that 2-4 cups of coffee per day also reduces the tendency for suicide, and decreases the rate of gout. It seems to be a preventative against kidney stones, too.

There is a confounding behavior that I should note, it’s possible that people who begin to feel signs of Parkinson’s, etc. stop drinking coffee. I doubt it, give the study’s design, but it’s worth a mention. The same confounding is also present in a previous analysis I did that suggested that being overweight protected from dementia, and from Alzheimer’s. Maybe pre-dementia people start loosing weight long before other symptoms appear.

Dr. Robert E. Buxbaum, and C.M. Buxbaum, December 15, 2022