Multi-Tenant Health Care Analytics

Platform receives HL7 feeds from multiple hospitals and processes them in real-time. Tenant-specific micro-services processes the transforms them into agnostic structure. The platform is powered by our flagship product 'LogiCrunch' which processes these messages in real-time, predicts and calibrates code blue, sepsis pathway for the patients and notifies them in real-time. Platform is built on Hortonworks, a popular Big Data stack.

IoT Scale Platform - Hospital(s) are typical organized as facilities, high and low acuity units within facilities, fixed and non-fixed devices within units. Transaction with in hospitals from these devices, Lab observations, ventilators, workflows around patient, are all manifested as HL7 messages. Platform receives these events/messages in real-time across multiple hospitals, processes and notifies back in real-time.

Streaming data - High-Acuity units usually comprise of 20% - 25% of the hospital beds. Devices attached to the patients stream-in messages at less than 1 minute per patient. Devices from the low-acuity units stream in data from about 15 minutes to 2 hours per patient, depending on the care-site. Lab and other transaction stream-in at a much lower velocity.

Predictive Analytics - We have built predictive models, to predict code blue and sepsis pathway based on the patients' vital and lab data. Shape and form of the data characteristics of Low and High acuity vary. We have models built for each of this population.

Multi-Tenant Data Fabric -Complete Data layer is Tenant-Aware. Platform is powered by our metalayer product 'LogiThink' that is key for tenancy support. The business and technical metadata that are loaded during/post-tenant onboarding processes, drive the tenant specific and tenant agnostic processes'. DataLake, usually stores the raw feeds received. Ponds are created and managed based on the scenarios of interest. Processes data is also stored on ElasticSearch, to enable fast tenant-specific storage and retrieval scenarios.

Data Governance - Platform leverages Hortonworks Data Governance framework, especially Apache Ranger, Atlas and Knox. Konx wraps over the enterprise LDAP and cascades the authentication tokens, while Ranger is configured to authorize access. Atlas configured in conjunction with Ranger and Solr, logs all the access Tenant specific rules, policies, access are isolated and metadata driven.

Secure Access - Platform hosted on a private cloud establishes trust with the tenant LDAPs. An authentication token from the clinicians, other hospital staff is mapped to access-permission sets on the assets/resources. All the resource(s) accessed are logged and monitored over search engine.