TABLE 2.

Summary of parameters in the simulation study.

Data distribution	Lognormal (1, 2)
	Gamma (1.07, 1)
	Weibull (0.7, 1)
Sample size	20–100
	100–500
	500–1000
% censored	0–30
	30–50
	50–80
Statistical methods	Substitution (zero, 1/2 LOD, LOD)
	Kaplan–Meier
	Regression on order statistics
	MLE
	Lognormal‐MLE
	Gamma‐MLE
	Weibull‐MLE
Summary statistics	Mean, 95th percentile
Parameters used to identify the best fitting model when using the parametric MLE method	Akaike information criterion
	Bayesian information criterion
	Root mean square error

Abbreviations: LOD, limit of detection; MLE, maximum likelihood estimation.