Exploring heterogeneity in clinical trials with latent class analysis

Abstract

Case-mix is common in clinical trials and treatment effect can vary across different subgroups. Conventionally, a subgroup analysis is performed by dividing the overall study population by one or two grouping variables. It is usually impossible to explore complex high-order intersections among confounding variables. Latent class analysis (LCA) provides a framework to identify latent classes by observed manifest variables. Distal clinical outcomes and treatment effect can be different across these classes. This paper provides a step-by-step tutorial on how to perform LCA with R. A simulated dataset is generated to illustrate the process. In the example, the classify-analyze approach is employed to explore the differential treatment effects on distal outcomes across latent classes.

Subtypes in patients with opioid misuse: A prognostic enrichment strategy using electronic health record data in hospitalized patients

“…LCA models were considered for one to eight classes for these class-defining variables. The optimal number of latent classes was selected using fit statistics including the Bayesian information criterion (BIC), adjusted Bayesian information criterion (aBIC), consistent Akaike information criterion (cAIC), class prevalence, class separation, and model interpretability [22]. Each patient encounter was assigned a class according to the highest latent class posterior probability.…”

“…Analysis was performed using Python Version 3.6.5 (Python Software Foundation) and RStudio Version 1.1.463 (RStudio Team, Boston, MA). Latent class analysis was performed with the poLCA package in R (http://dlinzer.github.com/poLCA) and followed the analysis plan by Zhang et al [22]. Our open-source code to perform LCA may be viewed in S3 Appendix as well as at: https://bitbucket.org/afsharjoycelab/opioid-misuse-lca/.…”

Application of Latent Class Modelling in Students' Life Skills: The Case of Iran University of Medical Sciences

“…This property can let the researcher compare the classes according to the participants' responses (9,10). In the latent class models, the effect of predictors on each class of participants can be modeled at the same time (11)(12)(13). In this study, the unconditional latent class model (ignoring the effects of predictors) and the conditional latent class model were fitted to the data.…”

“…The low value of BIC indicates a better fit and the significant tests of BLRT, LMRLR, and ALMRLR indicate that the model has a better fit than other models with different classes. The entropy index with the minimum value of 0.8 indicates that the quality of classes in the model is fine (11,13). Also, we used logistic regression to review the effect of predictors on the classes.…”

Machine Learning Method for the Management of Acute Kidney Injury: More Than Just Treating Biomarkers Individually

“…Intuitively, unsupervised ML aims to identify latent subclasses by features of the subjects, and there is no supervisor to guide the training iteration. Unsupervised ML encompasses principal component analysis, clustering analysis, latent profile (class) analysis and latent Markov models [12][13][14]. The unsupervised ML plays an important role in the exploration of the heterogeneity in clinical researches.…”