Untangling the complexity of multimorbidity with machine learning

Abstract

Highlights Machine learning is making significant contributions towards our understanding of the complex relationships between diseases. Advanced models take a range of modalities from big datasets with little pre-processing or information loss at study design. Developments in matrix factorisation and deep learning allow a better understanding of evolving patterns of multimorbidity.

Characterizing Multimorbidity from Type 2 Diabetes

“…The key clustering approaches used in the reviewed studies included agglomerative 24 , 31 and divisive 32 hierarchical clustering, 28 latent class analysis, 29 , 33 and graph theory, 34 all of which require substantially large health data sets. Moreover, novel machine learning approaches have been used to characterize and learn multimorbidity patterns, 35 , 36 but these methods are yet to be applied to T2DM specifically. Although the study approaches and settings were varied, predictors and predominant groups of conditions across clusters were identified and are discussed in the following sections.…”

AI and Big Data in Healthcare: Towards a More Comprehensive Research Framework for Multimorbidity

“…A general trend, nowadays, in assessing multimorbidity, is to move from disease-only to multi-modal presentation of phenotypes, including also information on medications, laboratory findings and functional health status, in addition to disease labels, in order to achieve better understanding of disease pathways. To meet this challenge, new algorithms and matrices have been developed, with improved capabilities to handle large and multi-modal datasets, and to extract hidden information from them, as presented in the recent review paper of Hassaine et al [ 31 ]. The major innovations include a shift from static to probabilistic implementations of basic ML methods, that allows a shift from qualitative descriptive to quantitative test methods, and developments in “deep phenotyping”.…”

“…The presentation of temporal dynamics of multiple interactions inherent to multimorbidity is a special challenge for data scientists as it requires more sophisticated algorithms and a multi-step analytics, that go beyond a case-control classification and the use of linear regression analysis to show its progression over time [ 105 , 106 , 107 ]. The study designs for learning about temporal dynamics of this complexity are still poorly developed and are maintained within the framework of unsupervised learning (an outcome is not known) and disease-disease relationships [ 31 , 107 ]. Although data scientists have an absolute authority in building these sophisticated data analytics, without incorporation of domain knowledge in creation of the research task and evaluation of the interpretability of the performed analytics, the real-life usability of these innovations will be questionable [ 104 ].…”

“…The purpose of this review is to summarize the limitations of current approaches to data analysis and to present the potential and shortcomings of alternative methods in multimorbidity research. The recently published paper by Hassaine et al provided an in-depth analysis of methodological advances in identifying multimorbidity-associated patterns hidden in data sequences in electronic health records (eHRs) and methods for tracking the time courses of these patterns [ 31 ]. On the contrary, in fact, because our overview shows what the problem of multimorbidity research looks like in the eyes of medical professionals.…”

Application of machine learning in understanding atherosclerosis: Emerging insights

“…Machine learning provides tools that can be applied to overcome the many challenges in multimorbidity, but only a small percentage have been used for the study of multimorbidity. 45 Hu et al demonstrated that ML can effectively predict prediabetics at risk for rapid atherosclerosis progression. 21 …”