Discussion of Paper by Horisberger

Abstract

Horisberger’s paper highlights many of the health services research problems inherent in attempting to understand the effects of medical care using available epidemiological, economic, and effectiveness data. The difficulties of using data that are collected for nonresearch reasons, such as administration, and are then used to answer specific research questions or hypotheses have been commented upon extensively and are the bane of all health services researchers. It too often leads to making heroic assumptions about the data, to complex, convoluted, and incomplete analyses, and to results couched in the most constrained terms and limited by a variety of caveats.