Evaluating individual heterogeneity in mental health research: an overview of clustering methods and guidelines for applications

Clustering models or cluster analyses have been widely used to explore individual heterogeneity in mental health research and psychology. Despite advances in new algorithms and increasing popularity, there is little guidance on model choice, analytical framework and reporting requirements. In this paper, we aimed to address this gap by first providing an introduction to the philosophy, design, advantages/disadvantages and implementation of major algorithms that are particularly relevant in mental health research with examples in the R software language. Extensions of basic models, such as kernel methods, deep learning, semi-supervised clustering, and clustering ensembles are subsequently introduced. How to choose algorithms to address common issues as well as methods for pre-clustering data processing, clustering evaluation and validation are then discussed. Importantly, we also provide general guidance on clustering workflow and reporting requirements. A rapid review of publications (December 2020-December 2021) focusing on the top six psychology and psychiatry journals publishing most of the clustering papers is also presented. The review highlights a few issues, including a lack of diversity in the algorithm of choice, robust internal and external validation processes (e.g., via resampling), and available data and analysis code to improve reproducibility. This comprehensive paper offers researchers (in both mental health and other broader health application areas) advanced tools and guidelines to use these methods efficiently, robustly and transparently.