Quantitative prediction error analysis to investigate predictive performance under predictor measurement heterogeneity at model implementation

Table 1 Simulation parameters

Parameter	Value
Baseline hazard of an event	0.1
Conditional hazard ratio for association predictor X and survival times	2
Time point of administrative censoring	15
Baseline hazard of censoring	0.01
Conditional hazard ratio for association between random variable for censoring and censoring times	3
Mean of predictor X and random variable for censoring	0
Variance of predictor X and random variable for censoring	1
Predictor W at implementation*
ψ	− 0.3 to 0.3
θ	0.5 to 2
σ_ε	0 to √2

*At implementation, a different measurement of predictor X was available, denoted measurement W. The connection between X and W was defined using the following measurement heterogeneity model: \( \mathbbm{E}(W)=\psi +\theta \mathbbm{E}(X)+\epsilon, \) where \( \epsilon \sim \mathcal{N}\left(0,{\sigma}_{\epsilon}^2\right) \), and where ψ denotes an additive shift in W with respect to X, θ denotes a multiplicative linear association between W and X, and σ_ε denotes random deviations from X

ISSN: 2397-7523