Available online 28 August 2022, 104165

This study proposes a new framework for subsequent treatment item recommendations. The method adopts a sequence recommendation approach to effectively solve the daily partial temporality learning problem in EMRs data effectively. 

We propose dual-level feature learning with recurrent neural networks, which can effectively learn different diagnostic features of patients, including the current treatment results features and the historical medical records features.

In order to fully learn the patient’s diagnostic features, we propose two attention mechanisms, including the elemental attention mechanism and the sequential attention mechanism.

The algorithm learns both the partial temporality and patient diagnostic features in the data to effectively meet the personalized treatment requirements of patients. Experiments on real-world datasets demonstrate the effectiveness and state-ofthe-art of our method.

In recent years, the massive electronic medical records (EMRs) have supported the development of intelligent medical services such as treatment recommendations. However, existing treatment recommendations usually follow the traditional sequential recommendation strategies, ignoring the partial temporality of the practical process and the patient’s diagnostic features. To this end, in this paper, we propose a new Dual-level Diagnostic Feature Learning with Recurrent Neural Network for treatment sequence recommendation (DDFL-RNN), where the dual-level diagnostic features including patients’ historical medical records and current treatment results. Firstly, we divide the dataset into several sequential sets of treatment item based on the patient’s treatment days. Secondly, we propose two kinds of attention mechanisms to learn diagnostic features, including the elemental attention mechanism and the sequential attention mechanism. Finally, the dual-level learned diagnostic features are brought into the recurrent neural network for encoding and recommendation. Extensive experiments on a breast cancer dataset from a first-rate hospital have shown that our model achieves significantly better performance than several classical and state-of-the-art baseline methods.

Treatment item recommendation

Sequential sets recommendation

Deep learning

Attention mechanism

Feature learning

© 2022 Elsevier Inc. All rights reserved.