Accounting for parameter uncertainty in the definition of parametric distributions used to describe individual patient variation in health economic models

Abstract

BackgroundParametric distributions based on individual patient data can be used to represent both stochastic and parameter uncertainty. Although general guidance is available on how parameter uncertainty should be accounted for in probabilistic sensitivity analysis, there is no comprehensive guidance on reflecting parameter uncertainty in the (correlated) parameters of distributions used to represent stochastic uncertainty in patient-level models. This study aims to provide this guidance by proposing appropriate methods and illustrating the impact of this uncertainty on modeling outcomes.MethodsTwo approaches, 1) using non-parametric bootstrapping and 2) using multivariate Normal distributions, were applied in a simulation and case study. The approaches were compared based on point-estimates and distributions of time-to-event and health economic outcomes. To assess sample size impact on the uncertainty in these outcomes, sample size was varied in the simulation study and subgroup analyses were performed for the case-study.ResultsAccounting for parameter uncertainty in distributions that reflect stochastic uncertainty substantially increased the uncertainty surrounding health economic outcomes, illustrated by larger confidence ellipses surrounding the cost-effectiveness point-estimates and different cost-effectiveness acceptability curves. Although both approaches performed similar for larger sample sizes (i.e. n = 500), the second approach was more sensitive to extreme values for small sample sizes (i.e. n = 25), yielding infeasible modeling outcomes.ConclusionsModelers should be aware that parameter uncertainty in distributions used to describe stochastic uncertainty needs to be reflected in probabilistic sensitivity analysis, as it could substantially impact the total amount of uncertainty surrounding health economic outcomes. If feasible, the bootstrap approach is recommended to account for this uncertainty.Electronic supplementary materialThe online version of this article (doi: 10.1186/s12874-017-0437-y) contains supplementary material, which is available to authorized users.

Comparing Strategies for Modeling Competing Risks in Discrete-Event Simulations: A Simulation Study and Illustration in Colorectal Cancer

“…A previously developed DES model 5 was implemented in R Statistical Software version 3.3.2, 19 according to the structure of the CAIRO3 study: postinduction, reintroduction, salvage, and death (Figure 1b). Model state postinduction refers to observation (control) or CAP-B maintenance treatment (intervention) after 6 cycles of CAPOX-B.…”

“…Uncertainty in parametric distributions’ parameters can be accounted for in probabilistic sensitivity analyses, so that both stochastic uncertainty (i.e., first-order uncertainty) and parameter uncertainty (i.e., second-order uncertainty) are reflected. 5 Although parametric distributions can also be used to populate STMs, this requires an additional discretization step, that is, evaluation of the cumulative density functions at fixed time points, to obtain discrete-time transition probabilities.…”

Estimating Maximum Daily Precipitation in the Upper Vistula Basin, Poland

“…As stated in [51], even when we have abundant, good-quality data to work with and a good model, our parameter estimates are still subject to a standard error. Although general guidance is available on how parameter uncertainty should be accounted for in probabilistic sensitivity analysis, there is no comprehensive guidance on the estimation of uncertainty in the parameters of the distributions used to represent stochastic uncertainty in statistical models [52]. Therefore, to assess the consistency of the theoretical distribution with the empirical distribution, the Kolmogorov-Smirnov (K-S) test was adopted [53].…”

Uncertainty Analysis of Standardized Precipitation Index Due to the Effects of Probability Distributions and Parameter Errors

“…The theory of MLE states that for large sample sizes n and a k-dimensional parameter vector, MLE estimators are approximately distributed as a multivariate normal. Thus, the uncertainties of distribution parameters are then quantified by randomly generating parameters based on the multivariate asymptotic normality assumption (Nixon et al, 2010;Degeling et al, 2017). The detailed process is as follows:…”