Characterization and valuation of uncertainty of calibrated parameters in stochastic decision models

by   Fernando Alarid-Escudero, et al.

We evaluated the implications of different approaches to characterize uncertainty of calibrated parameters of stochastic decision models (DMs) in the quantified value of such uncertainty in decision making. We used a microsimulation DM of colorectal cancer (CRC) screening to conduct a cost-effectiveness analysis (CEA) of a 10-year colonoscopy screening. We calibrated the natural history model of CRC to epidemiological data with different degrees of uncertainty and obtained the joint posterior distribution of the parameters using a Bayesian approach. We conducted a probabilistic sensitivity analysis (PSA) on all the model parameters with different characterizations of uncertainty of the calibrated parameters and estimated the value of uncertainty of the different characterizations with a value of information analysis. All analyses were conducted using high performance computing resources running the Extreme-scale Model Exploration with Swift (EMEWS) framework. The posterior distribution had high correlation among some parameters. The parameters of the Weibull hazard function for the age of onset of adenomas had the highest posterior correlation of -0.958. Considering full posterior distributions and the maximum-a-posteriori estimate of the calibrated parameters, there is little difference on the spread of the distribution of the CEA outcomes with a similar expected value of perfect information (EVPI) of $653 and $685, respectively, at a WTP of $66,000/QALY. Ignoring correlation on the posterior distribution of the calibrated parameters, produced the widest distribution of CEA outcomes and the highest EVPI of $809 at the same WTP. Different characterizations of uncertainty of calibrated parameters have implications on the expect value of reducing uncertainty on the CEA. Ignoring inherent correlation among calibrated parameters on a PSA overestimates the value of uncertainty.


Uncertainty Quantification in Friction Model for Earthquakes using Bayesian inference

This work presents a framework to inversely quantify uncertainty in the ...

On Calibrated Model Uncertainty in Deep Learning

Estimated uncertainty by approximate posteriors in Bayesian neural netwo...

Bayesian Calibration of Force-fields from Experimental Data: TIP4P Water

Molecular dynamics (MD) simulations give access to equilibrium and/or dy...

Policy Implications of Statistical Estimates: A General Bayesian Decision-Theoretic Model for Binary Outcomes

How should scholars evaluate the statistically estimated causal effect o...

Detecting Parameter Symmetries in Probabilistic Models

Probabilistic models often have parameters that can be translated, scale...

Percentile-Based Residuals for Model Assessment

Residuals are a key component of diagnosing model fit. The usual practic...

Microsimulation Model Calibration using Incremental Mixture Approximate Bayesian Computation

Microsimulation models (MSMs) are used to predict population-level effec...

Please sign up or login with your details

Forgot password? Click here to reset