This Article  • References  • Full text PDF  • Send to a friend  • Save in My Folder  • Save to citation manager  • Permissions  Citing Articles  • Contact me when this article is cited  Related Content  • Similar articles in JAMA  Social Bookmarking   What's this?

Rolla Edward Park, PhD; Robert H. Brook, MD; Katherine L. Kahn, MD
The RAND Corporation Santa Monica, Calif

Jacqueline Kosecoff, MD; Mark R. Chassin, MD
Value Health Sciences Santa Monica, Calif

JAMA. 1991;265(4):459-460.

	Since this article does not have an abstract, we have provided the first 150 words of the full text PDF and any section headings.

In Reply.—

As a practical matter, randomness depends on our state of knowledge. Dr Redelmeier demonstrates that if our knowledge were good enough to sort patients into two groups that live with probability.99 or die with probability.8, random variance would be markedly reduced. Indeed, if we truly knew and understood everything about a patient and his or her treatment at a cellular or molecular level, we might banish randomness entirely and explain outcomes perfectly. We took as our state of knowledge a relation between outcome and age-sex-race that approximates the targeting method used by HCFA when we started our project 5 years ago (New York Times. March 12, 1986:1). The method currently used by HCFA makes better use of information in the claims data¹ but does not reduce random variance very much. For example, principal diagnosis from claims data lets us sort patients with acute myocardial infarction into two of . . . [Full Text PDF of this Article]

CiteULike    Connotea    Del.icio.us    Digg    Reddit    Technorati    Twitter     What's this?