[ Proc. Natl. Acad. Sci. vol. 106 pp. 14716 – 14721 ’09 ] is an interesting paper which performed gene profiling on (easily obtainable) white blood cells in 103 healthy adults ages 25 – 40. Half of them were of low socioeconomic status (SES) in the first 5 years of life (as judged by what their parents did). All were of the same socioeconomic status at the time of testing (again judged by occupation). They found differences in gene expression between the two groups which didn’t correlate with lifestyle or perceived stress at the time of testing.
It took a lot of work (and probably money). People don’t do this sort of thing for fun. Why were they interested in the first place? Because of an even more interesting paper from Johns Hopkins [ Arch. Int. Med. vol. 166 pp. 2356 – 2361 ’06 ]. This work followed 1131 male Hopkins medical students for FORTY years. 19% came from backgrounds of low socioeconomic status (again judged by what their parents did). The striking conclusion was that there was a 240% increased risk of coronary artery disease by age 50 if you came from a background of low SES. This, in spite of the fact that for the duration of the study MDs were living a high SES existence. Impressive no?
As a practicing MD, I had to plow through this stuff year after year. Very quickly you begin reading papers in the medical literature with the attitude ‘how are they lying to me’. Well, maybe not actually lying, but drawing conclusions not warranted from the data or, worse, missing the forest for the trees. Some scientific training helps but isn’t necessary. My cousin’s boy wrote an absolutely brilliant article for his high school newspaper dissecting the methodology behind the annual college rankings in US News and World Report and he wants to be a writer.
What’s wrong with this paper? Certainly nothing is amiss with the data, painstakingly acquired year after year. Also, they were quite careful to control for lifestyle issues such as weight, smoking, exercise, alcohol consumption etc. etc. However they plotted two curves of coronary artery disease incidence vs. age (one for the low SES and the other for the 81% of the classes not of low SES) and cherrypicked the age at which the curves separated the most (e.g. 50). Also to be noted is that there wasn’t much coronary artery disease at 50 in either group — 13/218 in the low SES and 23/(1141 – 218) in the rest.
Also stated in the paper is that the mortality at age 70 was the same for both groups, even though the low SES group continued to have more coronary artery disease (and death from it). This implies that low SES in childhood actually protects against other fatal diseases (cancer perhaps?). They had to die of something after all. Which way would you want to go? That could have been the title of the paper, but wasn’t.
Even more interesting is the comparison they didn’t make –e.g. with the expected mortality and morbidity of a group of young men who remained in low SES throughout their working lives. This data is likely available. We are always reading about increased morbidity and mortality in one disadvantaged group or another (usually as a way of slamming the current system). My guess is that it would be much worse. That being the case the paper could have been titled, “A high SES in adult life negates the disadvantage of growing up poor”.
Reading the summary of the paper would have missed all this. What the authors chose to present was certainly an attention getter, and they made no attempt to hide their data. However, what your patients are paying for is your ability to evaluate data like this, think about it, and apply it to them. There’s nothing wrong with thinking, but dissecting paper after paper like this becomes tedious after a time. Chemistry, math and molecular biology are so much cleaner intellectually (but far less immediately important to your patients).
A truly awful example of missing the forest for the trees is the following: [ J. Am. Med. Assoc. vol. 259 p. 3158 ’88 ] An overview of the Physicians’ Health study in which 22,000 American physicians took either one adult aspirin or a placebo every other day. It’s pretty old but the study was widely cited and was very well worth doing because it dealt with a potentially simple (and cheap) way of preventing heart attack and stroke.
The study was double blind so identical packets of aspirin or placebo had to be prepared and delivered to all 22,000 docs in a timely fashion. Cardiovascular mortality was cut by aspirin, but overall mortality was not. Severe stroke was increased slightly, and there were 80 strokes in the aspirin group versus 70 in the placebo group. However the group experienced just 88 deaths when 733 would have been expected. The authors noted that low numbers of deaths made the data more difficult to interpret. Their discussion focused on whether the aspirin was adding anything extra. The conclusion was that this dose of aspirin probably wasn’t doing much.
Where’s the forest?
It’s the EIGHTFOLD reduction in mortality from what was expected. It could be due to a beneficial life style (money, social class) but the paper never discussed it. What we need is to reduce mortality in our patients eightfold and then worry about giving aspirin.