[Skip to Content]
[Skip to Content Landing]
January 4, 1995

Measuring Quality of Hospital Care-Reply

Author Affiliations

The Permanente Medical Group Oakland, Calif
University of California Berkeley

JAMA. 1995;273(1):22. doi:10.1001/jama.1995.03520250035016

In Reply.  —Drs Silber and Rosenbaum agree that selection bias is present, but they argue it is difficult to address as we proposed, and they argue that unbiased studies can be done on inpatient samples by adjustment for "all variables simultaneously relevant to outcomes and to selection into the groups being compared." As we stated, in a perfect world with complete information, we agree. However, we do not have perfect information. They then suggest that controversy over methods to address selectivity supports ignoring it. Selectivity is a ubiquitous problem in the study of hospital quality. While there is continuing discussion in the literature on methods to address selectivity, there is no controversy on the need to address it. To be skeptical that any one proposed solution is a panacea seems prudent. To propose that researchers simply ignore a serious problem seems foolhardy.Dr Huff incorrectly posits as our key finding

First Page Preview View Large
First page PDF preview
First page PDF preview