Enhancing Clinical Evidence Recommendation with Multi-Channel Heterogeneous Learning on Evidence Graphs

Clinical evidence encompasses the associations and impacts among patients, interventions (such as drugs or physiotherapy), problems and outcomes. The objective of recommending clinical evidence is to provide medical practitioners with relevant information to support their decision-making processes and to generate new evidence. Our specific task focuses on recommending evidence based on clinical problems. However, the direct connections between certain clinical problems and related evidence are often sparse, creating a challenge of link sparsity. Furthermore, recommending appropriate evidence requires jointly exploiting both the topological relationships among evidence and the textual information describing them. To address these challenges, we define two knowledge graphs: the Evidence Co-reference Graph and the Evidence Text Graph, to represent the topological and linguistic relations among evidential elements respectively. We also introduce a multi-channel heterogeneous learning model and a fusional attention mechanism to manage the co-reference-text heterogeneity in evidence recommendation. Our experiments demonstrate that our model outperforms state-of-the-art methods on open data.