Health information and technology solutions specialist Healthline launched its HealthData Engine big data analytics platform, which leverages the company’s HealthTaxonomy clinical natural language processing (NLP) technologies and semantic analysis.
The HealthData Engine is aimed at health care organizations that bear the risk of their patient populations and the cost of care delivery, and will be available to select health care enterprises during an Early Adopter Partner Program and to the general public in early 2015.
Those risk-bearing entities–integrated delivery networks (IDNs), accountable care organizations (ACOs), providers and payers–are the organizations that are investing in analytics.
“Specific applications powered by the HealthData Engine are focused on specific use cases. For example, our Risk Insight application is focused on reducing readmissions, initially for heart failure, and is primarily aimed at hospitals and large provider practices,” Murray Brozinsky, chief strategy officer of Healthline, told eWEEK. “Longer term, we hope to have the platform embedded as part of the EMR and other health IT systems.”
Brozinsky said analytics require the use of structured and unstructured patient data, such as physician transcriptions, hospital admission notes and discharge summaries, to accurately identify, group and risk-score patients.
The HealthData Engine aggregates and normalizes unstructured information from disparate sources and combines it with structured data, such as problem lists and medications, for a more comprehensive view of the patient’s health.
Using predictive factors from peer-reviewed studies, it leverages predictive modeling to risk-score and group patients into cohorts for targeted interventions.
“The scale and complexity of health care requires a data-driven approach to deliver personalized care. In 2009, HITECH injected tens of billions of dollars to spur the adoption of electronic health records. Wearable sensors tracking all manner of metrics are proliferating and, with Apple’s entry, will accelerate,” Brozinsky said. “The cost of collecting personal genomic data is dropping precipitously. Enormous amounts of data are being generated, but without powerful analytics to detect the signal from the noise and to make that signal actionable, we are blind.”
Readmissions Risk InSight for Congestive Heart Failure (CHF) will be Healthline’s first application in a suite of clinical applications built on the HealthData Engine.
The application leverages predictive clinical factors, combined with non-clinical predictive indicators such as socioeconomic and environmental factors, to score patients for risk for 30-day hospital readmissions due to CHF and help providers direct appropriate care to the patients who need it most.
Brozinsky also noted security and privacy are major issues when it comes to personal health information.
“There is heightened awareness with the recent consumer data security breaches at Target, Home Depot and Apple. Although these breaches are troubling, health data is arguably more sensitive to most people than credit cards and private pictures,” Brozinksy said. “Health care organizations and health IT providers will move slowly and deliberately on this dimension. Many health care organizations have been rightly conservative and have chosen to require patient data be kept on premises. However, we are seeing many more getting comfortable with cloud-based services.”