Abstract: During treatment and rehabilitation, patients may experience multiple nursing-related complications concurrently. Conventional single-outcome prediction models are limited in their ability to characterize the co-occurrence, temporal relationships, and interactions among these outcomes. Interpretable multi-label models (IMLMs) integrate multi-label learning with explainable artificial intelligence, enabling simultaneous prediction of multiple complications at the individual patient level while elucidating key predictive features, inter-label dependencies, and prediction uncertainty. This review summarizes recent advances in the application of IMLMs for nursing-related complication risk identification, focusing on their technical foundations, clinical applications, potential advantages, and existing challenges, to provide insights for nursing risk prediction, early warning, and clinical decision support.