Detecting treatment-subgroup interactions in clustered data with generalized linear mixed-effects model trees.

Identification of subgroups of patients for whom treatment A is more effective than treatment B, and vice versa, is of key importance to the development of personalized medicine. Tree-based algorithms are helpful tools for the detection of such interactions, but none of the available algorithms allow for taking into account clustered or nested dataset structures, which are particularly common in psychological research. Therefore, we propose the generalized linear mixed-effects model tree (GLMM tree) algorithm, which allows for the detection of treatment-subgroup interactions, while accounting for the clustered structure of a dataset. The algorithm uses model-based recursive partitioning to detect treatment-subgroup interactions, and a GLMM to estimate the random-effects parameters. In a simulation study, GLMM trees show higher accuracy in recovering treatment-subgroup interactions, higher predictive accuracy, and lower type II error rates than linear-model-based recursive partitioning and mixed-effects regression trees. Also, GLMM trees show somewhat higher predictive accuracy than linear mixed-effects models with pre-specified interaction effects, on average. We illustrate the application of GLMM trees on an individual patient-level data meta-analysis on treatments for depression. We conclude that GLMM trees are a promising exploratory tool for the detection of treatment-subgroup interactions in clustered datasets.

Quantitation of intrathecal immunoglobulin synthesis - a new empirical formula.

BACKGROUND AND PURPOSE: Intrathecal immunoglobulin (Ig) synthesis occurs in various chronic inflammatory neurological diseases. Different formulae have been developed for quantitative determination of Ig synthesis within the cerebrospinal fluid (CSF) compartment. The hyperbolic formula of Reiber is frequently used which, however, returns a considerable number of false positive results in empirical observations. METHODS: A computerized database of more than 19 000 paired CSF and serum samples was screened for patients presumed negative for local Ig synthesis and a new formula characterizing this collective was calculated. The validity of this formula was confirmed by several validation steps. RESULTS: A cohort of 1173 patients with normal CSF findings was used for quantile regression. The 97.5th quantile of the formula Qlim(IgX)=a×Qalbb was considered as the cut-off curve for intrathecal Ig synthesis using different constants a and b for IgG, IgA and IgM. Compared to the Reiber formula, a lower level of false positive results was produced especially for IgM and IgA which was confirmed in a separate clinically well defined validation cohort. In 77 patients with discrepant findings between Reiber and our formula no specific diagnoses were found confirming the low diagnostic value of borderline Ig synthesis. CONCLUSIONS: A new approximation formula was developed for determination of intrathecal Ig synthesis which produces fewer false positive results without reducing diagnostic sensitivity.