Until very recently, the objects of common use such as wearable devices, cars, watches, refrigerators and health-monitoring devices, did not produce or handle data at a large scale and lacked internet connectivity. However, furnishing such objects with computer chips and sensors that enable data collection and transmission over the internet has created a network of billions of physical devices around the world, commonly known as Internet of Things (IoT).
In the field of health care specifically, IoT has become nothing short of a revolutionary movement. Devices like fitness or health-tracking wearable devices, biosensors, clinical devices for monitoring vital signs etc. create a continuous stream of data. This makes them a major contributor towards revealing critical information that is potentially beneficial in improving the health and lifestyle of the population at large.
However, as with any technological disruption, IoT has led to an emergence of new datasets with extremely painful data management demands. Everything connected means that even the simplest of IoT applications will demand extremely open, flexible, and fundamentally connected data models. With the rich ecosystem of products that IoT presents, the management of such an information systemically.
Why graph databases are perfect for IoT
The traditional relational databases fail miserably when dealing with high volume, sensitive, and interconnected data ingested into organizational system sat a very high velocity from disparate sources. They cannot deliver real-time, a capability that is critical for virtual healthcare, due to technical limitations such as complex joins. Typical examples include master data management, ensuring compliance with GDPR, HIPAA/other regulations, failing to uncover or discover patterns in real-time in fraud detection, implementing symbolic AI/reasoning etc.The list of similar cases is endless, where traditional databases fail.
Graph databases are schema-less and built of nodes to store data entities and with edges to store relationships between them. They are a perfect choice for understanding complex, connected, and dynamic systems. As each smart device in a virtual cloud of devices is likely to have multi-faceted interrelationships with other devices, graph technology allows these relationships to be manifested more realistically, without the need to force fit into arbitrary relational models.
Graphs are especially useful for discovering previously unknown or little understood relationships. These relationships can include those arising from behavioral patterns or coincident patterns of change. This significantly advances the ability to unveil insights on everything in IoT, including data control and security, and facilitate real-time analytics on the complex relationships between connected devices.
A good example would be that of discovering fraud rings in real-time, which is prevalent in banking, and relevant in healthcare. Conducting entity link analysis to detect organized or collusive activities, kickbacks, fake referrals, and other healthcare frauds is almost impossible to achieve with traditional databases. This is because queries to find relations are incredibly complex to build, expensive to run and scaling them in a way that supports real-time access poses significant technical challenges.
In order to leverage the vital relationships and growing swarm of real-time devices that make up IoT, graph databases are optimized to not only query thedata quickly, but also to persist relationship data for perpetual real-time performance.
Graph enables us to ask questions we haven’t even considered asking before we had technologies optimized for providing these answers. However, graph databases still largely remain an untapped asset that can truly help the healthcare industry in dealing with real-time, saturated, and complex data.
Graph and IoT – making the health system work better for everyone
Graphs can easily manage data inflow from IoT devices and analyze it in real time. By integrating this data stream with historic IoT data and other sources like EMRs and PHRs in graph, an early intervention and treatment can be provided to patients within the comfort of their homes/offices.
More importantly, in the current COVID-19 pandemic situation, such a strategic graph database-IoT based solution cannot only help in providing virtual personalized healthcare butcan also help prevent spread of this contagious disease and the life-threatening risks associated with it.
Here is an example of how this could work – A mechanism has been built in the medical system for real-time streaming of all the patient vitals from wearable and home care devices into Graph DB, which also has an input of other connected patient data like EHR/EMR, gaps in care, biometrics, lab data, X-rays, MRI scans etc.Additionally,important features such sending an SOS to the concerned clinician for immediate action the moment a patient’s vitals show an abnormality, are built in as feeds to the Graph DB.
So, if an otherwise healthy individual were to suddenly develop a fast pulse, high blood pressure or undergo a change in weight, an SOS event would be sent to a cardiologist/nurse the moment this real-time event is received by the graph DB. The clinician would then immediately take a holistic view of the patient’s condition,taking into consideration all types of patient’s details along-with the newly received vitals, and decide the appropriate course of action.
All of this can happen in a matter of minutes.Not only this, the event can trigger a real-time stratification of the patient from low risk to medium/high risk through a ML model on/via Graph and alert an engagement specialist to do the appropriate outreach for enrolling the member into a related care management program without any delay.
The road ahead
With the advent of more and more processing power through technologies like quantum computing, high speed connectivity enabled through 5G networks and added level of digital intelligence enabled by AI, IoT networks are expected to make the very fabric of the world around us much more responsive and smarter. The devices will go beyond just transmitting data from patient to doctor but would help assist in areas such as medical adherence, remote monitoring,early warning,and telehealth.
With the adoption of FHIR standards, healthcare interoperability will improve, leading to growth of rich clinical data. As patients’ datasets evolve and AI algorithms mature, the promise of assisted diagnosis, disease prediction and precision medicine will become real. This is precisely where advances in graph technology, coupled with IoT will help build the foundation for evolving the next generation of connected healthcare.
(The article has been co-authored by Lalit Singla, Senior Director, Software Engineering, Optum Global Solutions and Harmeet S. Gambhir, Principal Engineer, Optum Global Solutions (India) Pvt. Ltd. and the views expressed in this article are their own)
