[Skip to Content]
[Skip to Content Landing]
June 25, 1997

The Negative Side of Cost-effectiveness Analysis-Reply

Author Affiliations

Lilly Research Laboratories Indianapolis, Ind
Lilly SA Madrid, Spain

JAMA. 1997;277(24):1932-1933. doi:10.1001/jama.1997.03540480032026

In Reply.  —While we sympathize with many of the points raised by Drs Stinnett and Mullahy, we cannot agree that the magnitude of a negative C/E ratio "conveys no useful information." This is an important part of what you "see" when you plot the outcome of a health economics study as a point (△C,△E on the cost-effectiveness plane.1Polar coordinates (radius and angle) are helpful to understand the technical issues here (Figure). Stinnett and Mullahy point out that the C/E ratio is an imperfect measure of angle; any given ratio corresponds to 2 angles that differ by exactly 180°. Outcomes in quadrants II and IV generate negative C/E ratios. An outcome in quadrant IV means that the experimental treatment dominates the standard treatment on long-range average, whereas an outcome in quadrant II indicates the opposite (ie, the standard treatment dominates the experimental treatment).Quadrant II (0,0) Quadrant III AC Quadrant I