Decision-making under uncertainty: using MLMC for efficient estimation of EVPPI

Giles, Michael B.; Goda, Takashi

doi:10.1007/s11222-018-9835-1

Cited by 33 publications

(53 citation statements)

References 20 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…for the same maximum level L with the MLMC estimator, so that the mean square error (7) of the NMC estimator is bounded above by ε 2 . As the theoretical result predicted, we see that the total cost of the MLMC estimator is of O(ε −2 ), whereas that of the NMC estimator is of O(ε −3 ).…”

Section: Test For the Bias Convergencementioning

confidence: 99%

Multilevel Monte Carlo estimation of expected information gains

Goda

Hironaka

Iwamoto

2019

Stochastic Analysis and Applications

Self Cite

View full text Add to dashboard Cite

The expected information gain is an important quality criterion of Bayesian experimental designs, which measures how much the information entropy about uncertain quantity of interest θ is reduced on average by collecting relevant data Y . However, estimating the expected information gain has been considered computationally challenging since it is defined as a nested expectation with an outer expectation with respect to Y and an inner expectation with respect to θ. In fact, the standard, nested Monte Carlo method requires a total computational cost of O(ε −3 ) to achieve a root-mean-square accuracy of ε. In this paper we develop an efficient algorithm to estimate the expected information gain by applying a multilevel Monte Carlo (MLMC) method. To be precise, we introduce an antithetic MLMC estimator for the expected information gain and provide a sufficient condition on the data model under which the antithetic property of the MLMC estimator is well exploited such that optimal complexity of O(ε −2 ) is achieved. Furthermore, we discuss how to incorporate importance sampling techniques within the MLMC estimator to avoid arithmetic underflow. Numerical experiments show the considerable computational cost savings compared to the nested Monte Carlo method for a simple test case and a more realistic pharmacokinetic model.

show abstract

Section: Test For the Bias Convergencementioning

confidence: 99%

Multilevel Monte Carlo estimation of expected information gains

Goda

Hironaka

Iwamoto

2019

Stochastic Analysis and Applications

Self Cite

View full text Add to dashboard Cite

show abstract

“…We estimated the per-person expected value of perfect information (EVPI), which is the expected improvement in decision making, or value, from removing all uncertainty in all parameters 11. We used multilevel Monte Carlo to estimate per-person expected value of partial perfect information (EVPPI), the expected value of removing all uncertainty in only a subset of parameters 28. We focused on sets of parameters likely to be informed by studies comparing cAVR with the Ross procedure.…”

Section: Methodsmentioning

confidence: 99%

Clinical and cost-effectiveness of the Ross procedure versus conventional aortic valve replacement in young adults

et al. 2019

View full text Add to dashboard Cite

ObjectivesIn young and middle-aged adults, there are three current options for aortic valve replacement (AVR), namely mechanical AVR (mechAVR), tissue AVR (biological AVR) and the Ross operation, with no clear guidance on the best option. We aim to compare the clinical effectiveness and cost-effectiveness of the Ross procedure with conventional AVR in young and middle-aged adults.MethodsThis is a systematic literature review and meta-analysis of AVR options. Markov multistate model was adopted to compare cost-effectiveness. Lifetime costs, quality-adjusted life years (QALYs), net monetary benefit (NMB), population expected value of perfect information (EVPI) and expected value of partial perfect information were estimated.ResultsWe identified 48 cohorts with a total number of 12 975 patients (mean age 44.5 years, mean follow-up 7.1 years). Mortality, bleeding and thromboembolic events over the follow-up period were lowest after the Ross operation, compared with mechAVR and biological AVR (p<0.001). Aortic reoperation rates were lower after Ross compared with biological AVR, but slightly higher when compared with mechAVR (p<0.001). At a willingness-to-pay threshold of £20effective. At a willingness-to-pay threshold of £20, 000 per QALY000 per QALY, the Ross procedure is more cost-effective compared the Ross procedure is more cost-effective compared withwith conventional AVR, with a lifetime incremental NMB of £60 conventional AVR, with a lifetime incremental NMB of £60 952 (952 (££3030 236236 to to ££7979 464). Incremental costs were £12464). Incremental costs were £12 323 (323 (££61086108 to to ££1515 972) and incremental QALYs 3.66 (1.81972) and incremental QALYs 3.66 (1.81 to to 4.76). The population EVPI indicates that a trial costing up to £2.03 million could be cost 4.76). The population EVPI indicates that a trial costing up to £2.03 million could be cost--effective.At a willingness-to-pay threshold of £20 000 per QALY, the Ross procedure is more cost-effective compared with conventional AVR, with a lifetime incremental NMB of £60 952 (£30 236 to £79 464). Incremental costs were £12 323 (£6108 to £15 972) and incremental QALYs 3.66 (1.81 to 4.76). The population EVPI indicates that a trial costing up to £2.03 million could be cost-effective.ConclusionsIn young and middle-aged adults with aortic valve disease, the Ross procedure may confer greater quality of life and be more cost-effective than conventional AVR. A high-quality randomised trial could be warranted and cost-effective.

show abstract

“…Probabilistic sensitivity analysis is an attempt to provide a framework for evaluating how the uncertainty of input parameters propagates to the uncertainty of model outputs, and has been studied intensively in the context of health economic evaluations [2][3][4][5] . For instance, the expected value of partial perfect information (EVPPI) evaluates how the perfect knowledge on an individual variable or a group of variables yields an increment of the expected cost-effectiveness [6][7][8][9] . Let θ θ θ = (θ θ θ 1 , θ θ θ 2 ) be a partition of the components in the vector θ θ θ .…”

Section: Introductionmentioning

confidence: 99%

Probabilistic threshold analysis by pairwise stochastic approximation for decision-making under uncertainty

Goda

Yamada

2021

Preprint

Self Cite

View full text Add to dashboard Cite

The concept of probabilistic parameter threshold analysis has recently been introduced as a way of probabilistic sensitivity analysis for decision-making under uncertainty, in particular, for health economic evaluations which compare two or more alternative treatments with consideration of uncertainty on outcomes and costs. In this paper we formulate the probabilistic threshold analysis as a root-finding problem involving the conditional expectations, and propose a pairwise stochastic approximation algorithm to search for the threshold value below and above which the choice of conditionally optimal treatments changes. Numerical experiments for both a simple synthetic testcase and a chemotherapy Markov model illustrate the effectiveness of our proposed algorithm, without any need for accurate estimation or approximation of conditional expectations which the existing approaches rely upon. Moreover we introduce a new measure called decision switching probability for probabilistic sensitivity analysis in this paper.

show abstract

Decision-making under uncertainty: using MLMC for efficient estimation of EVPPI

Cited by 33 publications

References 20 publications

Multilevel Monte Carlo estimation of expected information gains

Multilevel Monte Carlo estimation of expected information gains

Clinical and cost-effectiveness of the Ross procedure versus conventional aortic valve replacement in young adults

Probabilistic threshold analysis by pairwise stochastic approximation for decision-making under uncertainty

Contact Info

Product

Resources

About