Application scenario: Various physiological monitoring devices in the hospital ward

Through the IoMT smart hospital platform, data from hospital facilities such as patient information, networked devices (sensors, wearables), and medical equipment in the ward are collected in real-time and transformed into FHIR format for uploading to the platform. The data can be further developed for web and app applications, data analysis and research, and even data exchange with other hospitals. The platform can also be integrated with the customer service center to analyze real-time patient data. In case of an emergency, the customer service center can dispatch personnel immediately, increasing the flexibility of problem resolution.

The IoMT smart hospital platform integrates Microsoft IoT and Azure API for FHIR technologies. Patient basic information is standardized and converted to FHIR format through the gateway and uploaded to the Azure FHIR server via the API. Data from networked devices (sensors, wearables), and medical equipment in the smart ward are received in real-time through IoT Hub or IoT Edge. The original data (raw data) from IoT devices are sent to a secure storage environment for device management. The original data is then sent to the IoMT FHIR connector, which matches the device and patient association before uploading the data to the FHIR server. These data can be used for the development of web and app applications (such as personal health record systems (PHR) and SMART on FHIR), machine learning and AI research, and data exchange with other hospitals to improve medical referral efficiency.

Application scenario: Integrating all data from patient visits

Integrating clinical data such as patient records and reports, academic research results, imaging data, and physiological signal data, etc. through a medical data platform, anonymizing and cleaning (ETL) the data and effectively integrating it. Afterwards, the data is applied according to clinical needs, such as structuring medical records and reports through natural language processing, extracting key terms, and reusing clinical data to develop clinical research databases, evidence-based medicine, provide physician diagnostic decision support, disease analysis and warning, etc. By presenting the data in a visual and interactive way, users can better see the value of the data, creating patient-centered medical and care services.

Medical data application:
◼ Structured Pathology Reports: Structuring the unstructured pathology reports of the hospital through natural language processing techniques, referencing the AJCC-8th Diagnostic Item Checklist of the Taiwan Pathology Society's Cancer Pathology Report, extracting useful information such as disease summaries, tumor location, tumor size, nature, etc. so that users (physicians) can quickly obtain patient information and provide effective research data for future clinical studies to enhance the value of data usage.

◼ Clinical research database: Integrating anonymized clinical data, clinical researchers can apply for the necessary data, reducing the tedious process of collecting data in the past.

Application Scenario:: Medical imaging such as CT, MRI, X-ray, etc. applied in the field of artificial intelligence.

Through the AI platform, the concept of MLOps is built from the bottom-up, with hardware such as server-level DGX station or general GPU cards provided. In terms of software, it includes OS operating systems, containers, and AI development software provided by NVIDIA. If collaboration is needed, the platform also provides K8S, Harbor, and a graphical user interface. If CI/CD is needed, Git and Jenkins are also provided. The development environment uses Jupyter and provides frameworks such as TensorFlow and Pytorch.

Using the AI development platform to develop models can be applied in medical image analysis, patient relationship management, disease prediction, hospital safety management, medical education, etc., providing patient-centered medical services for hospitals, reducing the workload of medical staff, and improving patient safety management.

Medical imaging application:
Breast X-ray recognition: Through AI image recognition technology, BIRADS (The Breast Imaging Reporting and Data System) interpretation is detected from breast X-ray images. Combined with the PACS system, when the physician completes the lesion marking, AI-assisted diagnosis is activated at the same time. If AI determines that it is malignant, the attending physician is notified through text messages or communication software for further diagnosis, screening malignant patients quickly, and shortening the waiting time.

Renal Care Management System
The only three-in-one renal care system in Taiwan, including a Blood Dialysis Information System, Blood Dialysis Machine Data Transmission, Peritoneal Dialysis Information System, and Chronic Kidney Disease (CKD) Pre-ESRD Diabetes Case Management System. It integrates various data and records, providing healthcare professionals with complete medical history information, achieving paperless cloud-based management for renal care.

● Blood Dialysis Information System
Provides various blood dialysis methods and presents them in a structured and integrated dialysis order. It supports continuous recording across days and integrates hospital HIS information, facilitating healthcare professionals in performing optimal care.

● Blood Dialysis Machine Data Transmission
Automated dialysis data integration into blood dialysis records enables paperless operation, allowing nursing staff more time for clinical care work, and enhancing the quality of service and care for patients. It supports multiple international dialysis machine brands.
● Blood Dialysis Check-in System
Utilizes IoT technology to integrate a weight scale so that patients can check-in at the same time when measuring their weight. Nursing staff can receive check-in information in real-time.
● Peritoneal Dialysis Information System
Provides various peritoneal dialysis methods and can summarize patients' monthly dialysis assessment records.
● Chronic Kidney Disease (CKD) Case Management System
Provides complete and continuous tracking and treatment. When patients return for follow-up visits, they can view medical records of multiple chronic diseases and provide physicians with helpful medical information.