American Journal of Epidemiology Vol. 152, No. 2 : 196
Copyright © 2000 by The Johns Hopkins University School of Hygiene and Public Health
ERRATUM |
RE: "DR. SHAPIRO RESPONDS TO DR. HERTZ-PICCIOTTO"
Because of a technical error, the invited commentary by Dr. Hertz-Picciotto (1
) provided to Dr. Samuel Shapiro for response was not the final version. An unfortunate, but unintended, consequence was that item 2 of her critique (1, pp. 946–7) was materially different from the version provided to Dr. Shapiro. He also appeared to be misquoting her. The words "...the 'null' situation might be far lower than a relative risk of 1.0" (2, p. 949), originally a direct quote, no longer appeared in her final text. Dr. Shapiro's amended response is as follows:
2. By definition, the null is a relative risk estimate of 1.0. In contravention of that definition, Hertz-Picciotto suggests that the null need not always be a relative risk of 1.0. She uses the example of a positive family history of breast cancer to argue that since high-risk women may avoid oral contraceptives, "...the 'null' situation (no association between oral contraceptives and breast cancer) would give rise to a relative risk lower than 1.0..." (1, p. 947). This argument is erroneous, and among other things, it implies that family history is uncontrolled. Yet, it is straightforward to record and control that variable; indeed, it would be remiss not to do so. Moreover, it is circular reasoning to argue that the relative risk of less than 1.0 can represent the "null": What it represents, again by definition, is either a reduced risk or a bias.
The Journal regrets the error.
REFERENCES
-
Hertz-Picciotto I. Invited commentary: Shifting the burden of proof regarding biases and low-magnitude associations. Am J Epidemiol 2000;151:946–8.
[Free Full Text] -
Shapiro S. Dr. Shapiro responds to Dr. Hertz-Picciotto. Am J Epidemiol 2000;151:949–50.
[Free Full Text]
CiteULike 
Connotea 
Del.icio.us What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||