Library Header Image
LSE Research Online LSE Library Services

Adaptation of acceptability curves: probabilistic sensitivity analysis according to fixed parameters

Shillcutt, Sam, Coleman, Paul, Morel, Chantal M., Goodman, Catherine and Mills, Anne (2004) Adaptation of acceptability curves: probabilistic sensitivity analysis according to fixed parameters. In: 5th European Conference on Health Economics, 2004-09-08 - 2004-09-11, London, United Kingdom, GBR. (Submitted)

Full text not available from this repository.


Objectives: Acceptability curves present the probability that an intervention is cost-effective according to incrementally increasing levels of the ceiling ratio. Non-stochastic parameters other than the ceiling ratio may be uncertain within the scope of the model and be important drivers of cost-effectiveness. Probabilistic sensitivity analyses (PSA) can be extended to represent the probability that an intervention is cost-effective given multiple sources of non-stochastic uncertainty. This paper aims to present a methodological framework for undertaking PSA according to fixed parameters, demonstrate two- and three-dimensional PSA using curves and surface plots in an application to the diagnosis and treatment of malaria in sub-Saharan Africa and demonstrate further application of this framework to Expected Value of Perfect Information (EVPI) analysis. Methods: Key non-stochastic uncertain parameters were identified in two cost-effectiveness models evaluating rapid diagnostic tests versus presumptive treatment and artemisinin-based combination therapies versus sulphadoxine-pyrimethamine. Fixed variables used in the PSA were (1) the prevalence of malaria and the ceiling ratio; and (2) the starting level of parasite resistance, relative rate at which resistance develops between two drugs, timeframe and ceiling ratio. Results: This approach clearly shows that the cost-effectiveness of diagnostic tests is highly dependent on disease prevalence but not on the ceiling ratio, while EVPI is affected by both parameters. In the treatment model, this approach shows that the relative growth rate of resistance, levels of initial resistance and timeframe impact cost-effectiveness, with the ceiling ratio having little effect. EVPI is affected by all of these parameters. Conclusion: PSA according to fixed parameters provides a clear, visual representation of the effect of multiple sources of non-stochastic uncertainty, and interactions between them, in assessments of cost-effectiveness and EVPI.

Item Type: Conference or Workshop Item (Paper)
Official URL:
Additional Information: © 2004 The authors
Divisions: Social Policy
LSE Health
Subjects: H Social Sciences > HB Economic Theory
JEL classification: I - Health, Education, and Welfare > I1 - Health > I10 - General
Date Deposited: 14 Jul 2010 15:32
Last Modified: 16 May 2024 10:58

Actions (login required)

View Item View Item