Model-based reasoning methods for diagnosis in integrative medicine based on electronic medical records and natural language processing

Background This study aimed to investigate model-based reasoning (MBR) algorithms for the diagnosis of integrative medicine based on electronic medical records (EMRs) and natural language processing. Methods A total of 14,075 medical records of clinical cases were extracted from the EMRs as the development dataset, and an external test dataset consisting of 1,000 medical records of clinical cases was extracted from independent EMRs. MBR methods based on word embedding, machine learning, and deep learning algorithms were developed for the automatic diagnosis of syndrome pattern in integrative medicine. MBR algorithms combining rule-based reasoning (RBR) were also developed. A standard evaluation metrics consisting of accuracy, precision, recall, and F1 score were used for the performance estimation of the methods. The association analyses were conducted on the sample size, number of syndrome pattern type, and diagnosis of lung diseases with the best algorithms. Results The Word2Vec CNN MBR algorithms showed high performance (accuracy of 0.9586 in the test dataset) in the syndrome pattern diagnosis. The Word2Vec CNN MBR combined with RBR also showed high performance (accuracy of 0.9229 in the test dataset). The diagnosis of lung diseases could enhance the performance of the Word2Vec CNN MBR algorithms. Each group sample size and syndrome pattern type affected the performance of these algorithms. Conclusion The MBR methods based on Word2Vec and CNN showed high performance in the syndrome pattern diagnosis in integrative medicine in lung diseases. The parameters of each group sample size, syndrome pattern type, and diagnosis of lung diseases were associated with the performance of the methods. Strengths and limitations of this study A novel application of artificial intelligence – natural language processing approaches on diagnosis of integrative medicine A study of medical artificial intelligence based on real-world data of electronic medical records Multiple approaches on artificial intelligence to include traditional machine learning algorithms, neural network, and deep learning algorithms Rule-based combining model-based reasoning to be explored in this dataset