JMIR Publications

ABSTRACT

Background:

Multimorbidity has become a major global public health challenge. However, existing research primarily emphasizes the identification of disease patterns at the population level, while lacking the capacity to provide predictive insights into individual future pattern membership.

Objective:

This study aims to propose an innovative framework that integrates population-level multimorbidity pattern recognition with individual-level predictive modeling, thereby advancing multimorbidity research from descriptive analysis toward prospective risk prediction.

Methods:

Using longitudinal health follow-up data, we first applied Latent Transition Analysis (LTA) to identify temporally stable multimorbidity patterns, and subsequently transformed these patterns into predictive labels for constructing a novel deep learning model, CLA-Net (Cross-Lag Attention Network), designed for predicting individual future multimorbidity patterns. CLA-Net leverages the complementary strengths of GRU and Transformer architectures and introduces a bi-temporal directed cross-attention mechanism to simultaneously capture temporal dependencies and complex feature interactions.

Results:

Experimental results demonstrate that CLA-Net significantly outperforms several advanced baseline models in multimorbidity pattern prediction, achieving a precision of 0.8326±0.0053, recall of 0.8312±0.0056, F1-score of 0.8319±0.0051, and accuracy of 0.8352±0.0048, confirming the effectiveness and robustness of the proposed framework.

Conclusions:

This study extends the scope of LTA beyond descriptive statistical modeling and, for the first time, establishes the scientific value of multimorbidity pattern prediction as an independent research task. By bridging population-level insights with individual-level prediction, the proposed framework provides an intelligent tool for clinical risk stratification and personalized intervention, and offers new methodological and practical value for precision medicine and public health policymaking.