[Skip to Content]
Access to paid content on this site is currently suspended due to excessive activity being detected from your IP address Please contact the publisher to request reinstatement.
[Skip to Content Landing]
January 5, 1990

Surgical Therapy for Coronary Disease: Randomized and Observational Studies

Author Affiliations

University of California, San Francisco

University of California, San Francisco

JAMA. 1990;263(1):33. doi:10.1001/jama.1990.03440010031011

To the Editor.—  As Califf et al1 mention in their discussion, the results of all observational studies must be interpreted cautiously because of potential selection bias. Statistical adjustment is a well-established method of attempting to control for selection bias. However, the choice of which factors to include in this adjustment are critical. As the authors state, although there are differences between patients treated medically and those treated surgically, there is a large overlap between the two groups in their baseline characteristics. This suggests two possibilities: either there are additional criteria, known by physicians, used to select either medical or surgical treatment, or, within this overlap group, selection is somehow haphazard or random. Since the latter seems unlikely, it is almost certain that additional undefined factors are used to select treatment. What the authors have done is carefully to adjust for factors associated with the prognosis of ischemic heart