I would like to thank the journal Anesthesiology for the opportunity to respond to these letters to the editor. In general, the purpose of our editorial1
was to educate the anesthesiologist community on the strengths and weaknesses of propensity analysis. It was not meant advocate or demean this type of analysis. I will respond to each letter sequentially.
Dr. Engoren is correct that observational studies should be encouraged as a complement to prospective randomized studies. He is also correct that there are limitations and biases to prospective randomized control trials, which he enumerates. Despite these limitations, they are still considered the gold standard.
Drs. Vincent and Sakr are correct that one of the strengths of their study is the very large size of the Sepsis Occurrence in Acutely Ill Patients database.2
It also is a weakness in that they are using data from another study that was designed for another purpose. They are correct that their statistical analysis is well performed. The comment in our editorial about the statistical process being opaque, simulating a “black box,” was intended as a general comment about propensity analysis, not specifically their propensity analysis.
Drs. Boylan and Kavanagh are correct that our editorial was long on methods and short on biology. This was intentional because we had a limited word count and the goal of our editorial was to educate the anesthesiologist community on the strengths and weaknesses of propensity analysis. In searching through the literature, I found very few articles describing propensity analysis in the anesthesia literature. The authors do a very nice job describing the biology.
Dr. Vergouwe et al. are correct that propensity scores do not necessarily lie, but to nonstatisticians they are mysterious. The authors are correct that the title was a play on the quote by the English Prime Minister Benjamin Disraeli (1804–1880). Though the title was provocative, we tried to write a balanced editorial on the strengths and weaknesses of propensity analysis. I trust that the anesthesiologist community is smart enough not to be biased by a title in a single editorial.
Gregory A. Nuttall, M.D.
Mayo Clinic College of Medicine, Rochester, Minnesota. firstname.lastname@example.org
1. Nuttall GA, Houle TT: Liars, damn liars, and propensity scores. Anesthesiology 2008; 108:3–4
2. Vincent JL, Sakr Y, Sprung CL, Harboe S, Damas P: Are blood transfusions associated with greater mortality rates? Anesthesiology 2008; 108:31–9
© 2008 American Society of Anesthesiologists, Inc.