The Evolution of Health Data Analytics

Healthcare systems have progressed in recent years through three kinds of health data analytics: collection, sharing, and analysis. However, the collection and sharing phases using electronic records and exchanges for healthcare information haven’t appreciably affected the price and quality of healthcare. As for data, despite the fanfare and application of these approaches in other sectors and industries, healthcare is a different story; the industry has actually just started to have the capacity to facilitate system-wide improvement and spending reduction. The most exciting thing about analytics is related to its potential to allow for a data-driven culture in healthcare that will ultimately enhance patient care.

The Healthcare Analytics Adoption Model of health data analytics is an example of such a framework. It was created by experienced healthcare industry vets and was intended to be a helper for providing a systematic way of implementing analytics and classifying capabilities groups. For sustainable analytics, foundational elements are key for supporting the upper tiers as the organization matures. This model includes:

  • An evaluation framework for the industry related to its analytics adoption
  • A roadmap so that the organization can chart its progress
  • A structure that assists in the adoption of products from vendors

Lowering costs while improving patient care is the goal, and there are eight levels in the optimal adoption of this model. Before implementation, there are often inconsistent records and analytics in the organization that are addressed on a sporadic, as-needed basis; there is no integration, consistency or data governance. However, the first level shows immediate improvement:

1. An Enterprise Data Warehouse

This foundational step for technology and data involves a warehouse with a minimum of HIMSS EMR Stage 3 data, Patient Experience, Financial, Costing, Revenue Cycle and Supply Chain. A searchable metadata repository and insurance claim records are available. Updates take place within a month or less. Data governance starts to take shape.

2. Standardized Patient Registries and Vocabulary

Core data is organized and standardized, with patient registries based upon ICD billing data. Registries and data management capacity evolves gradually.

3. Internal Automated Reporting

Consistent, efficient production is in evidence as analytics are focused upon the production of reports in an efficient, consistent manner that supports the operation and management of the healthcare organization at a basic level.
Key performance indicators can be accessed from both front-line managers and executives.

4. External Automated Reporting

This stage shows consistent, efficient agility and production focused upon consistent report production for key regulatory requirements. Basic keyword searches and centralized data governance are available.

5. The Variability Reduction of Waste and Care

This stage relates to the management and measurement of evidence-based care and is focused on minimizing waste, reducing variability and measuring adherence to best practices. Management teams focused upon improving patient health are assisted by population-based analytics. Multi-discipline teams monitor how to improve quality and lower cost and risk in a much more precise manner. Data is standardized and evidence-based, a combination of patient registry cost and clinical data that includes all insurance claims. Data is updated within about a week.

6. Suggestive Analytics and Population Health Management

This step involves a financial commitment and preparing the workplace for more integrated analytics and best practices. Half or more of cases use bundled payments, and analytics from point of care to patient care quality, the economics of care, and population management are covered. Bedside devices, external pharmacy data, home monitoring data and more detailed information are also covered. Updates take place within about a day.

7. Predictive Analytics and Clinical Risk Intervention

Higher financial risk can be managed proactively, and analytics expands to cover fixed-fee per capita, diagnosis-based reimbursement models. Beyond just case management, there is also collaboration with payer and clinician partners to manage care with forecasting, predictive modeling and risk stratification. Home monitoring data, protocol-specific outcomes and long-term care facility data are also reported. Updates take place in an hour or less.

8. Prescriptive Analytics and Personalized Medicine

This level involves managing and contracting for health, prescriptive analytics/personalized medicine, wellness management, physical and behavioral functional health, interventional decision support and mass customization of care. Prescriptive analytics at the point of care improve patient outcomes due to analyzing overall population outcomes. There is rapid updating of all data including familial data, biometrics and genomic data.

While health data analytics can be confounding without an adequate structure, if a systematic framework is in place, it can be very helpful in guiding both priorities and approach for success.