Cost - Effectiveness Concepts and the CE Plane

T. Ganiats

DOI: https://doi.org/10.1177/0272989X9101100213

1991-06-01

Medical Decision Making

Abstract:plane. In table 2 of his article, Black presents strategy 1 as a program costing $250,000 (i.e., saving $250,000) and providing five QALYs. He then calculates the program’s costeffectiveness ratio (CER) as $50,000/QALY (i.e., a savings of $50,000/QALY). Calculating such a ratio in quadrant IV of the CE plane is misleading. This is illustrated in table 1 of this letter. Strategy 1 is the same as Black’s example. This strategy saves as much money as, but provides fewer QALYs than, strategy 2. Obviously, strategy 2 is more cost-effective than strategy 1. Looking at the CER for each strategy, however, reveals that the CER for strategy 1 is $50,000/QALY vs. strategy 2’s CER of $25,000/QALY (i.e., strategy 1 saves more dollars per QALY than does strategy 2). In other words, strategy 1 appears the preferred strategy after calculating the CERs. Calculating d for strategies 1 and 2 shows the expected outcome with d, < d~, meaning strategy 2 is the preferred strategy. The explanation of this apparent paradox is straightforward. Both strategies 1 and 2 lie in quadrant IV of the CE plane. Net costs in quadrant N are beneficial (c < 0; dollars saved), as are the net health effects (e > 0). The process of obtaining the ratio divides the costs by the health effects. When both costs and health effects are beneficial they are additive; calculating their ratio diminishes their value and is inappropriate. A better presentation of strategy 1 is, &dquo;$25,000 saved and 5 QALYs produced.&dquo; In quadrant II, both cost and health effects are detrimental (c > 0; e < 0), and calculating the ratio is also inappropriate.