IS IT TIME TO ELIMINATE THE ICER? USING NET BENEFITS TO REPORT THE RESULTS OF DETERMINISTIC COST-EFFECTIVENESS ANALYSES
Author(s)
O'Day K, McLaughlin T, Bramley TXcenda, LLC., Palm Harbor, FL, USA
Presentation Documents
BACKGROUND: Incremental cost-effectiveness ratios (ICERs) are used to report the results of cost-effectiveness (CE) analyses and represent the cost per unit of effectiveness of a more costly and more effective option. However, numerous conceptual and practical problems limit the usefulness of ICERs for decision making. These problems include, but are not limited to, negative ICERs, one-way sensitivity analyses, complexity of multiple comparator analyses, and statistical limitations with ratios. The net benefits approach was developed to address the statistical limitations of ICERs and is now an accepted methodology used in probabilistic sensitivity analysis to estimate CE confidence intervals and plot acceptability curves. However, despite the remaining challenges and limitations the use of ICERs persists, raising the question: Is it time to eliminate the use of ICERs in the reporting of CE analyses? METHOD: We propose expanding the net benefit method to present deterministic CE analysis results using a net monetary benefit (NMB) chart or table. A NMB chart is plotted with the x-axis representing the WTP threshold and the y-axis representing the NMB. The NMB of each option is a line with the intercept representing the cost and the slope representing the effectiveness across a specified range of WTP values. The line with the greatest NMB at a given WTP represents the most cost-effective option at that WTP. The vertical distance between two lines represents the incremental NMB. Dominance, extended dominance, and the frontier are captured graphically and intuitively. Multiple comparator analyses are simplified and one-way sensitivity analyses are enhanced due to the elimination of negative ICERs. CONCLUSION: A net benefits approach provides a more intuitive, informative, and useful method to present CE results than the use of ICERs. Moreover, it has the benefit of facilitating a uniform and consistent approach to presenting the results of deterministic and probabilistic CE analyses.
Conference/Value in Health Info
2011-05, ISPOR 2011, Baltimore, MD, USA
Value in Health, Vol. 14, No. 3 (May 2011)
Code
PRM41
Topic
Methodological & Statistical Research
Topic Subcategory
Confounding, Selection Bias Correction, Causal Inference
Disease
Multiple Diseases