Interaction-aware Drug Package Recommendation via Policy Gradient

Abstract

Recent years have witnessed the rapid accumulation of massive electronic medical records (EMRs), which highly support intelligent medical services such as drug recommendation. However, although there are multiple interaction types between drugs, e.g., synergism and antagonism, which can influence the effect of a drug package significantly, prior arts generally neglect the interaction between drugs or consider only a single type of interaction. Moreover, most existing studies generally formulate the problem of package recommendation as getting a personalized scoring function for users, despite the limits of discriminative models to achieve satisfactory performance in practical applications. To this end, in this paper, we propose a novel end-to-end D rug P ackage G eneration (DPG) framework, which develops a new generative model for drug package recommendation with considering the interaction effect within drugs that affected by patient conditions. Specifically, we propose to formulate the drug package generation as a sequence generation process. Along this line, we first initialize the drug interaction graph based on medical records and domain knowledge. Then, we design a novel message passing neural network to capture the drug interaction, as well as a drug package generator based on a recurrent neural network. In detail, a mask layer is utilized to capture the impact of patient condition, and the deep reinforcement learning technique is leveraged to reduce the dependence on the drug order. Finally, extensive experiments on a real-world dataset from a first-rate hospital demonstrate the effectiveness of our DPG framework compared with several competitive baseline methods.

Research directions in recommender systems for health and well-being

“…As the Web has matured and developed, so have recommender systems, and specifically recommender system applications in health (Alhijawi et al 2022). Recent applications include drug recommendation (Zheng et al 2022), diabetes treatment recommendation (Oh et al 2022), cancer drug response prediction (Wang et al 2022), and recommendations for cancer patients and caregivers (Rahdari et al 2022), to name a few.…”

Dynamic Sparse Learning: A Novel Paradigm for Efficient Recommendation

“…Recommendation systems [44,45,57] provide personalized services for today's web and have achieved significant success in various fields, such as e-commerce platforms, medical care [56,59,60], education [58] and job search [55]. At its core is learning high-quality representations for users and items based on historical interaction data.…”