Motivation

Data transmitted by internet of things (IoT) medical devices into the cloud needs to be efficiently aggregrated, stored, processed and democratized for use in real-time monitoring, data analytics, data science and machine learning. A data platform can organize the data into a central repository and processing center where data is ingested, transformed and operationalized. This platform serves as a foundation for patient, treatment and medical device data to be presented and shared to the main stakeholders, namely

• Medtech businesses to remotely track the performances and usage of their installed medical devices and patients outcomes.
• Physicians to monitor their many patients and each of their treatment progress individually and in aggregate.
• Patients to follow their specific journey from diagnosis to treatment to prognosis.

Solution

The ingestion of structured or unstructed data from the many IoT medical devices deployed anywhere in the world may occur at anytime, simultaneously or not. The potential for high volume of concurrent, real-time ingestion and ETL (extraction, transformation and loading) of IoT data can be consider as OLTP (online transaction processing), often used in e-commerce or banking. In this application, each medical device procedure applied to each patient is treated as a transaction. Once the data is in their proper stores, OLAP (online analytical processing) occurs through querying medical device procedures in the databases, processing and transforming the data to prepare it for use in analytics to find hindsights, insights and foresights. These data-driven findings on the device performance, usage and patient outcomes help medtech businesses make informed decisions on the reliability and treatment succeses of their current devices as well as offer guidance into future potential advancements to improve their medical devices.

With this is architecture and design, it is not limited to just one medical device, this can be extended to multiple other connected devices. The same cloud resources can be replicated for the other devices. Since patients, data format and data types may be different for each device, it is best to keep the data storages separate. As the data collected for each device may need to be processed in different ways, separate cloud resources and custom microservices will be needed. Additionally, if for some reason services for one device get disrupted, the services for other devices will remain unaffected and will still be up and running, limiting the interruptions.