The lay abstract featured today (for Evaluation of a flexible piecewise linear mixed-effects model in the analysis of randomized cross-over trials by Moses Mwangi, Geert Verbeke, Edmund Njeru Njagi, Alvaro Jose Florez, Samuel Mwalili, Anna Ivanova, Zipporah N. Bukania and Geert Molenberghs) is from Pharmaceutical Statistics with the full article now available to read here.
Evaluation of a flexible piecewise linear mixed-effects model in the analysis of randomized cross-over trials. Pharmaceutical Statistics. 2023; 1–15. doi:10.1002/pst.2357, , , et al.
Cross-over designs are commonly used in randomised clinical trials to estimate efficacy of a new treatment. They have received a lot of attention, particularly in connection with regulatory requirements for new drugs. The main advantage of using cross-over designs over conventional parallel designs is increased precision, thanks to within-subject comparisons. In the statistical literature, more recent developments are discussed in the analysis of cross-over trials, in particular regarding repeated measures. A piecewise linear model within the framework of mixed effects has been proposed in the analysis of cross-over trials. A simulation study that compares the performance of a piecewise linear mixed-effects (PLME) model with two commonly used models, Grizzle’s mixed-effects (GME) and Jones & Kenward’s mixed-effects (JKME) models, was designed. The study aimed at mimicking real-life situations by using empirical data to derive true underlying parameters. The findings from real-life data confirmed the original hypothesis that high-dose iodine salt significantly lowers diastolic blood pressure. To further evaluate the performance of PLME model, a simulation study was conducted using a 2×2 cross-over design. The fixed-effects, random-effects, and residual error parameters used in the simulation were estimated from diastolic blood pressure data using the PLME model. The initial results show that the univariate PLME model outperforms the GME and JKME models in estimating the variance-covariance matrix governing the random effects, ensuring satisfactory model convergence during estimation. When considering a hierarchical viewpoint, where outcomes are conditioned upon random effects, the variance-covariance matrix of the random effects must be positive-definite. The PLME model is preferred, especially when modeling an increased number of random effects, as it can explain more variability in the data and improve precision in estimating the effect size parameters.