A hierarchical Bayesian latent class mixture model with censorship for detection of linear temporal changes in antibiotic resistance

Abstract

Identifying and controlling the emergence of antimicrobial resistance (AMR) is a high priority for researchers and public health officials. One critical component of this control effort is timely detection of emerging or increasing resistance using surveillance programs. Currently, detection of temporal changes in AMR relies mainly on analysis of the proportion of resistant isolates based on the dichotomization of minimum inhibitory concentration (MIC) values. In our work, we developed a hierarchical Bayesian latent class mixture model that incorporates a linear trend for the mean log 2 MIC of the non-resistant population. By introducing latent variables, our model addressed the challenges associated with the AMR MIC values, compensating for the censored nature of the MIC observations as well as the mixed components indicated by the censored MIC distributions. Inclusion of linear regression with time as a covariate in the hierarchical structure allowed modelling of the linear creep of the mean log 2 MIC in the non-resistant population. The hierarchical Bayesian model was accurate and robust as assessed in simulation studies. The proposed approach was illustrated using Salmonella enterica I,4,[5],12:i:-treated with chloramphenicol and ceftiofur in human and veterinary samples, revealing some significant linearly increasing patterns from the applications. Implementation of our approach to the analysis of an AMR MIC dataset would provide surveillance programs with a more complete picture of the changes in AMR over years by exploring the patterns of the mean resistance level in the non-resistant population. Our model could therefore serve as a timely indicator of a need for antibiotic intervention before an outbreak of resistance, highlighting the relevance of this work for public health. Currently, however, due to extreme right censoring on the MIC data, this approach has limited utility for tracking changes in the resistant population.

Longitudinal Trajectories of Hair Cortisol: Hypothalamic-Pituitary-Adrenal Axis Dysfunction in Early Childhood

“…The same process of fitting LCMM models within subgroups was repeated until consistent ln-HCC patterns remained within each identified group. This approach is commonly recommended in various statistical texts (25)(26)(27) and in a reference text: "Applied Latent Class Analysis" edited by Hagenaars and McCutcheon (28). These authors state that LCMM analyses may be performed sequentially or simultaneously within the subclasses identified from an initial model.…”

Overview of Quantitative Methodologies to Understand Antimicrobial Resistance via Minimum Inhibitory Concentration

“…These subpopulations are distinguished by the presence of resistance mechanisms that are mixed into a bimodal distribution consisting of WT and non-WT components [38]. Unlike logistic regression models, these models need not rely on a cutoff to classify observations, and some implementations use latent variables to account for the interval-censored data and mixing weights to estimate the prevalence of WT and non-WT isolates [60,61].…”

“…Another study employed a type of hierarchical Bayesian mixture model that also presumes the MIC data to be composed of overlapping WT and non-WT distributions [60]. This model assumes both components follow a log-normal distribution with fixed standard deviation.…”

“…In contrast to the mixture models previously described in this review, this approach introduces regression by allowing the WT population mean to vary with the covariates (e.g., time) while keeping the non-WT population mean fixed, largely due to heavy amounts of right-censoring in the data sets employed in the study. Additionally, the mixture model estimates prevalence of non-WT isolates along with a variance for this value, with this calculation relying on the model's ability to distinguish between the WT and non-WT distributions instead of using a single cutoff point such as the ECOFF [60]. It should be noted that this approach requires several assumptions including normality of the non-WT distribution, whereas the previous model assumed the non-WT to be a mixture of multiple normal distributions, and a fixed variance for the WT and non-WT normal distributions [60].…”

A Bayesian approach to modeling antimicrobial multidrug resistance

“…These resistant and non-resistant components of the bacterial populations are indicated by the bimodal distributed frequency plot of the observed MIC. Following this idea, subsequent studies on AMR temporal trends [ 25 ] and cross-population correlation in AMR [ 26 ] have improved the previous ones that did not address the censored nature or the underlying mixture distribution. Jaspers et al (2018) [ 27 ] developed a Bayesian method to model the joint MIC density of two or more antimicrobials, from which the correlation of type i can be inferred.…”