BACKGROUND The dramatic increase in adolescent obesity is a serious public health crisis in the world. The World Health Organization has projected that by 2030, adolescent obesity will reach 254 million children worldwide. Increasing evidences show that obesity in adolescence would increase the risk of type 2 diabetes and cardiovascular disease in adulthood. A prediction model for adolescent obesity could help clinicians and adolescents monitor, control, and identify risk factors before children become overweight, enabling more personalized healthy lifestyle improvement for adolescent obesity. OBJECTIVE This study aims to develop a risk prediction model for adolescents using lifestyle factors, living environment data, and health literacy for the prediction of becoming overweight and obese in the upcoming month, and to explore living environment and lifestyle factors that may predispose youth to overweight and obesity. METHODS This prospective study was conducted at National Taiwan University Hospital. Parents and eligible adolescents were enrolled in the study. Living environment and lifestyle factors were collected by a wearable device, a smartphone app, the open environmental data API, and a case management platform. Standardized questionnaires were designed to evaluate the health literacy value of adolescents. To analyze the large amounts of heterogeneous data, we implemented six machine learning models: Random Forest, Decision Tree, SVM, KNN, LDA, and AdaBoost, and used Shapely Additive exPlanations and feature selection process to find the most cost-effective feature set to account for the problem of incomplete data in the real world. RESULTS All data from 120 adolescents were collected prospectively during a mean 1-year follow-up. For the risk prediction, the proposed model produced the best performance an accuracy of 94.3%, precision of 99,9%, and F1 score of 78.8%. Overall, the accuracy of the test set was 81.6%-94.3% for six machine learning algorithms. After the process of feature selection, the combination of daily consumption in calories, health literacy value, average heart rate and minimum heart rate was identified as the most cost-effective feature set. The purposed model with only these four features could achieve an accuracy of 93.7%, sensitivity of 71%, precision of 88.8%, and F1 score of 78.6%. CONCLUSIONS In contrast with previous existing studies, the proposed model could yield reliable prediction of the risk of becoming overweight and obesity in adolescent by adding objective lifestyle and environmental data. Our results indicated that lower values for features such as health literacy, consumption in calories, average heart rate, and rapid eye movement time would increase the risk of becoming overweight and obese. This information would help adolescents understand exactly how to improve their lifestyle and health outcomes. Furthermore, we have constructed the most cost-effective model that only needs four features to complete the prediction task, which is very helpful for deploying the risk prediction model in real life. CLINICALTRIAL The study protocol was approved by the institutional review board of the National Taiwan University Hospital (201710066RINB).

This paper presents an integrated and scalable precision health service for health promotion and chronic disease prevention. Continuous real-time monitoring of lifestyle and environmental factors is implemented by integrating wearable devices, open environmental data, indoor air quality sensing devices, a location-based smartphone app, and an AI-assisted telecare platform. The AI-assisted telecare platform provided comprehensive insight into patients' clinical, lifestyle, and environmental data, and generated reliable predictions of future acute exacerbation events. All data from 1,667 patients were collected prospectively during a 24-month follow-up period, resulting in the detection of 386 abnormal episodes. Machine learning algorithms and deep learning algorithms were used to train modular chronic disease models. The modular chronic disease prediction models that have passed external validation include obesity, panic disorder, and chronic obstructive pulmonary disease, with an average accuracy of 88.46%, a sensitivity of 75.6%, a specificity of 93.0%, and an F1 score of 79.8%. Compared with previous studies, we establish an effective way to collect lifestyle, life trajectory, and symptom records, as well as environmental factors, and improve the performance of the prediction model by adding objective comprehensive data and feature selection. Our results also demonstrate that lifestyle and environmental factors are highly correlated with patient health and have the potential to predict future abnormal events better than using only questionnaire data. Furthermore, we have constructed a cost-effective model that needs only a few features to support the prediction task, which is helpful for deploying real-world modular prediction models.

BACKGROUND:The automatic coding of clinical text documents by using the International Classification of Diseases, 10th Revision (ICD-10) can be performed for statistical analyses and reimbursements. With the development of natural language processing models, new transformer architectures with attention mechanisms have outperformed previous models. Although multicenter training may increase a model's performance and external validity, the privacy of clinical documents should be protected. We used federated learning to train a model with multicenter data, without sharing data per se.OBJECTIVE:This study aims to train a classification model via federated learning for ICD-10 multilabel classification.METHODS:Text data from discharge notes in electronic medical records were collected from the following three medical centers: Far Eastern Memorial Hospital, National Taiwan University Hospital, and Taipei Veterans General Hospital. After comparing the performance of different variants of bidirectional encoder representations from transformers (BERT), PubMedBERT was chosen for the word embeddings. With regard to preprocessing, the nonalphanumeric characters were retained because the model's performance decreased after the removal of these characters. To explain the outputs of our model, we added a label attention mechanism to the model architecture. The model was trained with data from each of the three hospitals separately and via federated learning. The models trained via federated learning and the models trained with local data were compared on a testing set that was composed of data from the three hospitals. The micro F1 score was used to evaluate model performance across all 3 centers.RESULTS:The F1 scores of PubMedBERT, RoBERTa (Robustly Optimized BERT Pretraining Approach), ClinicalBERT, and BioBERT (BERT for Biomedical Text Mining) were 0.735, 0.692, 0.711, and 0.721, respectively. The F1 score of the model that retained nonalphanumeric characters was 0.8120, whereas the F1 score after removing these characters was 0.7875-a decrease of 0.0245 (3.11%). The F1 scores on the testing set were 0.6142, 0.4472, 0.5353, and 0.2522 for the federated learning, Far Eastern Memorial Hospital, National Taiwan University Hospital, and Taipei Veterans General Hospital models, respectively. The explainable predictions were displayed with highlighted input words via the label attention architecture.CONCLUSIONS:Federated learning was used to train the ICD-10 classification model on multicenter clinical text while protecting data privacy. The model's performance was better than that of models that were trained locally.

Background: Classifying diseases into ICD codes has mainly relied on human reading a large amount of written materials, such as discharge diagnoses, chief complaints, medical history, and operation records as the basis for classification. Coding is both laborious and time consuming because a disease coder with professional abilities takes about 20 minutes per case in average. Therefore, an automatic code classification system can significantly reduce the human effort. Objectives: This paper aims at constructing a machine learning model for ICD-10 coding, where the model is to automatically determine the corresponding diagnosis codes solely based on free-text medical notes. Methods: In this paper, we apply Natural Language Processing (NLP) and Recurrent Neural Network (RNN) architecture to classify ICD-10 codes from natural language texts with supervised learning. Results: In the experiments on large hospital data, our predicting result can reach F1-score of 0.62 on ICD-10-CM code. Conclusion: The developed model can significantly reduce manpower in coding time compared with a professional coder.

Background The International Classification of Diseases (ICD) code is widely used as the reference in medical system and billing purposes. However, classifying diseases into ICD codes still mainly relies on humans reading a large amount of written material as the basis for coding. Coding is both laborious and time-consuming. Since the conversion of ICD-9 to ICD-10, the coding task became much more complicated, and deep learning– and natural language processing–related approaches have been studied to assist disease coders. Objective This paper aims at constructing a deep learning model for ICD-10 coding, where the model is meant to automatically determine the corresponding diagnosis and procedure codes based solely on free-text medical notes to improve accuracy and reduce human effort. Methods We used diagnosis records of the National Taiwan University Hospital as resources and apply natural language processing techniques, including global vectors, word to vectors, embeddings from language models, bidirectional encoder representations from transformers, and single head attention recurrent neural network, on the deep neural network architecture to implement ICD-10 auto-coding. Besides, we introduced the attention mechanism into the classification model to extract the keywords from diagnoses and visualize the coding reference for training freshmen in ICD-10. Sixty discharge notes were randomly selected to examine the change in the F1-score and the coding time by coders before and after using our model. Results In experiments on the medical data set of National Taiwan University Hospital, our prediction results revealed F1-scores of 0.715 and 0.618 for the ICD-10 Clinical Modification code and Procedure Coding System code, respectively, with a bidirectional encoder representations from transformers embedding approach in the Gated Recurrent Unit classification model. The well-trained models were applied on the ICD-10 web service for coding and training to ICD-10 users. With this service, coders can code with the F1-score significantly increased from a median of 0.832 to 0.922 (P Conclusions The proposed model significantly improved the F1-score but did not decrease the time consumed in coding by disease coders.

In this research, we aim to construct an automatic ICD-10 coding system.ICD-10 is a medical classification standard which is strongly related to scope of payment in health insurance.However, the work of ICD-10 coding is time-consuming and tedious to ICD coders.Therefore, we build an ICD-10 coding system based on NLP approach to reduce their workload.The result of f1-score in whole label prediction task is up to 0.67 and 0.58 in CM and PCS, respectively.In addition, recall@20 in whole label prediction task is up to 0.87 and 0.81 in CM and PCS, respectively.In the future, we will keep working on combining the current work with the rule-based coding system and applying the other brand new NLP techniques to improve our performance.