A reader pointed out that there is a logical fallacy in the post below about the abortion-premie link. This article was not written by me, but came from a newsletter. While I recognize that there may be some methodological flaws in the reporting, I none the less think that it raises important questions that need to be asked.
Epidemiology is the study of large scale human disease and illness. Originally, this branch of science was limited to studying infectious disease epidemics. It has a long history with the usual proportion of heroism and scandal.
The logical fallacy for which I was reprimanded was the confusion of correlation and cause. Simply because events occur in conjunction does not mean that one causes the other. With this I do agree. However, what my reader either ignored or refused to see is that the basic tool of epidemiology is to find correlations and then investigate to see if there may indeed be a causal event. This is something that may be ignored or denied due to political or personal reasons. The history of medicine abounds with unfortunate examples of this kind of hubris. An excellent book (alas, out of print) on this is Retrolental Fibroplasia: A Modern Parable.
Statistical inference is the tool used to help determine if the correlation is strong enough to justify further investigation. One examines a large population who had experienced a particular event or engaged in a certain behavior such as abortion, smoking, an earthquake, or whatever. One also looks at certain possible outcomes such as depression, cancer, preterm birth, or whatever the investigator chooses to examine. The group is compared to another group as similar as possible. The rules of statistics are used to see if the differences between the groups can be explained by random occurance. If not, there is a strong suspicion of cause.
The best studies of this kind collect the data up front, either from the moment of the event forward, or in context of collecting voluminous data about many events. There is bias in collecting data after the fact, but in reality, much epidemiology is done retrospectively.
One classic study in epidemiology was of a cholera epidemic in London, which was stopped by one physician who noted retrospectively that the majority of cases came from families using water from a certain pump. He removed the handle, and the epidemic was arrested. Another classic study was by Dr. Ignacz Semmelweiss. He noted that the laboring women attended by the physicians had significantly more death and illness from childbed fever than those attended by the midwives in the same hospital. He noticed that the midwives did not do surgery or attend autopsies. He mandated that the physicians wash their hands after surgery or autopsy and before attending women in labor. The death rate plummeted. He was laughed out of town despite his improved outcomes, and when his hand washing ritual was abolished, women started dying at increased rates again.
All I ask those who do not think abortion is associated with or potentially causative of such health risks as future preterm birth, future breast cancer, or depression, is that they do the research. If only 1% of the women who walked through abortion clinics would agree to be followed prospectively, there would be a statistical universe to answer some of the questions that I think many abortion supporters are afraid to even ask.
