[Skip to Content]
[Skip to Content Landing]
March 8, 1995

Case-Mix Adjustment: Making Bad Apples Look Good-Reply

Author Affiliations

Harvard Community Health Plan Brookline, Mass

JAMA. 1995;273(10):773. doi:10.1001/jama.1995.03520340026018

In Reply.  —Drs Welch, Black, and Fisher raise a valid point, that in some instances differential discovery and/or use of diagnoses can influence diagnostic adjusting models. They correctly identify the diagnosis of disk disease as one condition where clinical practice can be associated with the presence of a diagnosis. Unfortunately, no objective criterion standard for presence of disease exists free from the "bias" of physician and/or patient perspective. It would be wrong, however, not to adjust for differential burden of disease simply because we cannot do it perfectly. The existence of some degree of measurement error should not lead us to abandon the case-mix adjustment effort.Our data demonstrated significant variation in identifying "outlier" physicians when analyses were performed with and without case-mix adjustment. Given the choice, we prefer the risk of overadjusting to the risk of underadjusting. Since we are not able to control the administrative purposes for which

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