JMIR Publications

ABSTRACT

Background:

Traditional Chinese medicine practitioners usually follow a four-step evaluation process during a patient diagnosis: Observation, Auscultation, Olfaction, Inquiry, Pulse feeling and Palpation. The information gathered in this procedure along with lab test results and other measurements such as vital signs is recorded in the patient’s electronic health record. All the information needed for making a treatment plan is contained in the electronic health record. But only a seasoned Traditional Chinese medicine physician could utilize well this information to make a good treatment plan since the reasoning process is very complicated and it takes years of practice for a medical graduate to master the reasoning skill. In this digital medicine era, with a deluge of medical data, and ever increasing computing power and more advanced artificial neural network models, it becomes not only desirable but also readily possible for a computerized system to mimic the decision-making process of a Traditional Chinese medicine physician

Objective:

To develop an assistive tool that can predict prescriptions for hospital inpatients based on patients’ clinical electronic health records.

Methods:

Clinical health records containing medical histories as well as current symptoms and diagnosis information are used to train a Transformer-based neural network model using the corresponding doctor’s prescriptions as the aimed target. This is accomplished by extracting relevant information, such as the patient’s current sickness, medicines taken, nursing care given, vital signs, examinations and lab results from the patient’s electronic health records. The obtained information is then sorted chronologically to produce a sequence of data for each patient. This time-sequence data is then used as input to a modified Transformer network, which is chosen as a prescription prediction model. The output of the model is a prescription for the patient. The ultimate goal is for this tool to generate a prescription that matches what an expert Traditional Chinese medicine physician would prescribe. To alleviate the issue of overfitting, a Generative Adversarial Network (GAN) is used to augment the training sample dataset by generating noise-added samples from the original training samples.

Results:

A total of 21,295 copies of inpatient electronic medical records from Guang’anmen traditional Chinese medicine hospital was used in this research. These records were created between January 2017 and December 2018, covering a total of 6352 kinds of medicines. These medicines were sorted into 819 types of first category medicines based on the class relationships among medicines. As shown by the test results, the performance of a fully trained Transformer model can have an average precision rate of 80.58%，and an average recall rate of 68.49%.

Conclusions:

As shown by the preliminary test results, the Transformer-based TCM prescription recommendation model outperforms the existing conventional methods. The extra training samples generated by the GAN network helps to overcome the overfitting issue, leading a further improved recall rate and precision rate.