5 Ways Data Lakes Improve Healthcare Processes, Outcomes
With availability of healthcare data rapidly growing, providers want more efficient tools to analyze a broader set of clinical and business data points. Roberta Katz, director of healthcare-life sciences at information technology applications and services vendor EMC Corp., says data lakes can be par
5 Ways Data Lakes Improve Healthcare Processes, Outcomes
With availability of healthcare data rapidly growing, providers want more efficient tools to analyze a broader set of clinical and business data points. Roberta Katz, director of healthcare-life sciences at information technology applications and services vendor EMC Corp., says data lakes can be part of the solution. She explains the concept and gives examples. (Photo: Fotolia)
What is a Data Lake?

A data lake offers a technology infrastructure that pools data from multiple sources across the healthcare ecosystem and applies analytics models to provide a new approach to information management, reporting and predictive analytics, Katz explains. A data lake acts as the technology for capturing maximum value from all data being created across the continuum of care. (Photo: Fotolia)
A Data Lakes Role in Analytics

Simply put, a data lake provides an IT environment that incorporates structured, semi-structured and unstructured data from internal and external sources, ultimately improving effectiveness and quality of critical business and clinical practices, Katz says. A data lake applies advanced analytics to produce actionable insights, enabling timely interventions to prevent adverse health events and elevate overall population wellness. (Photo: Fotolia)
Designing the Lake

A well-designed data lake architecture includes three core capabilities:
Ingest and Store: It stores structured and unstructured data for all types of analytics from numerous sources, blending capacity and performance as needed for the analytics use case.
Analyze: It supplies modern data management and analytics tools for all types of analytics, including Hadoop-based, In-Memory No-SQL and Scale-out MPP (massively parallel processing).
Surface and Act: It provides users and applications with data that will enable real-time changes in outcomes and influence critical decisions.
With a data lake, providers can consolidate legacy systems data with new data sources that include new applications, cloud services, social, mobile and third-parties such as insurers, and shorten their analytics time-to-value. Katz suggests five use cases. (Photo: Fotolia)
Use Case 1: Patient Care Delivery

Need: Identify potential factors in organizational risk.
Method: Through data lake analytics, identify how, where and when patients were contracting hospital-acquired conditions.
Result: Readjust hospital procedures where conditions were most prevalent. (Photo: Fotolia)
Use Case 2: Predictive Analytics for Operational Improvement

Need: Identify likelihood of a surge in patient volume.
Method: Use data lake analytics to review community event data, weather conditions, flu rates and other variables.
Result: Determine when a surge in patient volume is likely and ensure sufficient staffing resources are on hand. (Photo: Fotolia)
Use Case 3: Clinical Research

Need: Enable the medical community to develop new insights for disease detection and management.
Method: Data lakes enable researchers to focus high-level skills on discoveries, speeding research to the clinical setting.
Result: New insights improve disease diagnostics and treatment. (Photo: Fotolia)
Use Case 4: Population Health Management

Need: Improve community wellness management.
Method: Use data lakes to create lifestyle-based profiles to risk stratify patients.
Result: Detailed outreach and care strategies based on patient profiles contribute to more positive patient outcomes and reduced cost of care. (Photo: Fotolia)
Use Case 5: Security Analytics

Need: Quickly and efficiently spot cybersecurity breaches.
Method: Consolidate all system logs and network activity into a data lake.
Result: Providers are enabled to identify anomalies faster; leading to more targeted responses, and also can predict when and where most cybersecurity problems will occur. (Photo: Fotolia)