Performance of Methods for Handling Missing Categorical Covariate Data in Population Pharmacokinetic Analyses

Abstract

In population pharmacokinetic analyses, missing categorical data are often encountered. We evaluated several methods of performing covariate analyses with partially missing categorical covariate data. Missing data methods consisted of discarding data (DROP), additional effect parameter for the group with missing data (EXTRA), and mixture methods in which the mixing probability was fixed to the observed fraction of categories (MIX(obs)), based on the likelihood of the concentration data (MIX(conc)), or combined likelihood of observed covariate data and concentration data (MIX(joint)). Simulations were implemented to study bias and imprecision of the methods in datasets with equal-sized and unbalanced category ratios for a binary covariate as well as datasets with non-random missingness (MNAR). Additionally, the performance and feasibility of implementation was assessed in two real datasets. At either low (10%) or high (50%) levels of missingness, all methods performed similarly well. Performance was similar for situations with unbalanced datasets (3:1 covariate distribution) and balanced datasets. In the MNAR scenario, the MIX methods showed a higher bias in the estimation of CL and covariate effect than EXTRA. All methods could be applied to real datasets, except DROP. All methods perform similarly at the studied levels of missingness, but the DROP and EXTRA methods provided less bias than the mixture methods in the case of MNAR. However, EXTRA was associated with inflated type I error rates of covariate selection, while DROP handled data inefficiently.

Population Pharmacokinetics of Methylphenidate in Healthy Adults Emphasizing Novel and Known Effects of Several Carboxylesterase 1 (CES1) Variants

“…Genotype information regarding CES1 diplotypes, rs71647871, and rs115629050 was missing in 0.8%, 5%, and 42%, respectively, of the total studied population. Missing data were considered as being missing not at random and was handled using the EXTRA (or EST) method . Gender was included on MTT due to the improvement of model fit assessed through decrease of OFV and improvement of visual diagnostics.…”

“…Taking the aforementioned limitation of the genetic analysis into account, as well as the proportion of different missing genotypes, in this study, missing data were considered as being missing not at random. The EXTRA (or EST) method, which quantifies the effect of missing covariate data through full maximum likelihood modeling and adds a parameter for each covariate containing missing data, was therefore chosen due to its precise and unbiased parameter estimates when dealing with missing not at random data, as well as its ease of implementation in stepwise covariate procedures . Genotype information regarding rs115629050, located in CES1A1 exon 7, was found to be missing in 42% of the study population, which was high compared with the amount of missing data observed in the other final model‐included CES1 variants.…”

Pharmacokinetics of Capecitabine and Four Metabolites in a Heterogeneous Population of Cancer Patients: A Comprehensive Analysis

“…In case of missing categorical covariates a separate parameter for the missing group was estimated, 34 as follows (Eq. 5):…”

Effect of diurnal variation,CYP2B6genotype and age on the pharmacokinetics of nevirapine in African children

“…44 Proportionality and correction factors were applied on RUV to test for differences between the assays and laboratories used.…”