This paper proposes two new frequentist methods for estimating consistency and inconsistency models in network meta-analysis (NMA), which are implemented using multivariate random-effects meta-regressions. The first method is a two-stage procedure that uses data augmentation to handle multi-arm trials, while the second method directly fits the inconsistency model. Both methods are illustrated using the mvmeta package in Stata. The paper also reviews the concept of inconsistency and discusses the choice of reference categories, data augmentation parameters, and the choice of prior distributions for Bayesian analysis. The methods are applied to a dataset of thrombolytic drugs, demonstrating their ability to test for consistency and rank treatments. The results show that the inconsistency model estimates larger heterogeneity than the consistency model, and the data augmentation approach provides an excellent approximation to the true model. The paper concludes by discussing the strengths and weaknesses of the proposed methods and suggesting future research directions.This paper proposes two new frequentist methods for estimating consistency and inconsistency models in network meta-analysis (NMA), which are implemented using multivariate random-effects meta-regressions. The first method is a two-stage procedure that uses data augmentation to handle multi-arm trials, while the second method directly fits the inconsistency model. Both methods are illustrated using the mvmeta package in Stata. The paper also reviews the concept of inconsistency and discusses the choice of reference categories, data augmentation parameters, and the choice of prior distributions for Bayesian analysis. The methods are applied to a dataset of thrombolytic drugs, demonstrating their ability to test for consistency and rank treatments. The results show that the inconsistency model estimates larger heterogeneity than the consistency model, and the data augmentation approach provides an excellent approximation to the true model. The paper concludes by discussing the strengths and weaknesses of the proposed methods and suggesting future research directions.