[Skip to Content]
[Skip to Content Landing]
July 14, 1989

To Best Was Always to Worst, but Worse Is Better, if Good and Bad Are Relevant-Reply

JAMA. 1989;262(2):210. doi:10.1001/jama.1989.03430020052024

This article is only available in the PDF format. Download the PDF to view the article, as well as its associated figures and tables.


In Reply. —  The decision to respond to a letter such as that from Dr Banzhaf from "SPELL" represents a complicated problem. Decision theory provides a framework to analyze this problem, as it has for other complicated decisions.

Methods. —  We have constructed a decision tree with five branches: (1) ignore the letter completely (IGNORE); (2) admit our error, grovel, and beg for forgiveness (GROVEL); (3) withdraw the article in shame (WITHDRAW); (4) resign our faculty positions at the University of Rochester in disgrace (RESIGN); and (5) compose a tongue-incheek reply indicating how minor we considered SPELL's grammatical concerns (RAILLERY).Each branch of the tree had several possible outcomes, each with differing probability of occurrence. Probability estimates were obtained by DELPHI techniques using colleagues familiar with medical decision theory. Each outcome included the effects on the number of infants' eyes treated correctly by ophthalmologists (weight = 99), the longrun effect on