Like most industries, healthcare is suffering from data overload, presenting both a challenge and an opportunity. This phenomenon, known as Big Data—massive volumes, variety and velocity of data—has permeated nearly every industry (e.g., retail, media, advertising, banking). Market leaders that have adopted a Big Data strategy have experienced impressive results, such as new revenue streams, lower costs, and improved customer engagement and employee satisfaction. However, based on the technical complexity and market dynamics of the healthcare industry, it has been slower to respond and adapt. All of that is changing now out of urgency and necessity. As with all emerging technologies, there have been several leading, innovative healthcare organizations that were early adopters of Big Data analytics—the practice of integrating and analyzing large, disparate datasets, often in real time, to gain new insights, uncover hidden patterns, influence outcomes, and predict events. Big Data analytics is now at the forefront, showing great promise for healthcare organizations in addressing clinical, operational, and financial aspects of care delivery.
The U.S. healthcare industry is experiencing a seismic shift under the Patient Protection and Affordable Care Act (ACA) and other market dynamics. The end result is increased attention to managing risk, gaining efficiency, improving quality, and reducing costs. As payers are no longer allowed to actuarially cherry-pick members, they are seeking to mitigate increased risk by seeking various forms of value-based payments with providers. However, providers are seeking contractual arrangements that limit risk exposure, especially in the short term, as they build capacity and expertise in deploying and using analytics-based solutions for population health management, clinical decision support, and care coordination.
The Centers for Medicare & Medicaid Services (CMS) had been piloting certain forms of value-based payments long before the passage of the ACA, but in a recent statement, CMS has indicated that 50% of reimbursements will be based on quality and outcomes by 2018. Private payers are not far behind: Aetna has established a goal of 75% of reimbursements via value-based contracts by 2020. A group of the top U.S. health systems, payers, and stakeholders announced in January the formation of the Healthcare Transformation Task Force, a private-sector alliance aimed at accelerating the healthcare industry’s transformation to value-based care. As of January 1, 2015, according to Managed Care, there were 424 Medicare Shared Savings Program Accountable Care Organizations (ACOs) and 159 private ACOs.
These circumstances necessitate that both payers and providers leverage their vital information assets to improve clinical, operational, and financial performance. The recent trend toward payer-provider mergers, partnerships, and joint ventures seeks to address fragmented healthcare delivery, emphasize patient-centered care, and promote care coordination. With the expansion of high-deductible health plans and emphasis on healthy lifestyles and wellness, there is also more focus on patient-generated data and the consumerization of healthcare. Aligning the goals and incentives among payers, providers, and patients sets the stage for data sharing, integration, and harmonization and coordinated Big Data analytics strategy. To be successful and competitive, stakeholders must share data to leverage analytics for (1) risk stratification and identification of high-cost or high-risk patients by using predictive models; (2) performance monitoring and evaluation of key quality measures; (3) development of longitudinal patient records; and (4) establishment of market-based models for bringing greater price transparency, measuring variation in cost and quality, and delivering predictable outcomes.
Healthcare organizations across the United States are becoming more interested in Big Data analytics and a related suite of applications to improve care delivery, but barriers to adoption include lack of skilled resources, limited access to data, inability to integrate data, and privacy concerns. Moreover, with so many vendor solutions, there is a sense of confusion and frustration on how to establish an analytics capability. To address these concerns, EHR vendors are moving toward open, cloud-based platforms that allow easier access to data via Application Programming Interfaces (APIs) that connect data sources to facilitate integration of new features into existing applications and to develop innovative point-of-care applications that bring context to each patient encounter in real time. This capability can be extended to detect, correlate, and monitor events across the enterprise as they occur for pattern matching, anomaly detection, and aggregation of data in motion with data at rest to make it actionable and apply advanced analytics and predictive modeling.
A related development is the emergence of Fast Health Interoperability Resources (FHIR, pronounced “fire”), a proposed Health Level 7 standard describing data formats and APIs for exchanging health data. FHIR solutions are built from a set of modular components called “resources,” which can be assembled into working systems. FHIR offers improved interoperability via Web services and common protocols, allowing users to interact with clinical data in new and different ways. By capturing data in a standard format, FHIR also offers a streamlined approach to implement complex clinical workflows and execute transactions and rules within FHIR resource bundles, thereby improving evidence-based medicine to make accurate diagnoses and select the most effective care pathways.
A common practice among healthcare experts is to collect as many health data as possible to explore and discover the “unknown unknowns,” especially for research in genomics and clinical informatics. Big Data requires a new generation of scalable technologies designed to extract insights from very large volumes of disparate, multistructured data by enabling high-velocity capture, discovery, and analysis. This is now possible with scale-out, clustered compute platforms such as Hadoop and Spark, which utilize low-cost, commodity hardware for collecting massive amounts of data for low-latency batch processing and real-time analysis. In conjunction with clinical data warehouses, NoSQL and in-memory databases, and data discovery and visualization tools, it is feasible to establish a unified analytics platform over a multitenant environment with domain-specific enclaves to evolve from systems of record to systems of engagement. As these systems and the standards governing them become more mature, it will be possible to analyze data more efficiently—to identify adverse drug interactions, misdiagnoses, missed routine exams, healthy lifestyle recommendations, disease prevention, and others. The scope of analysis can range from the individual patient level to the entire healthcare system to several healthcare systems connected via a single insurer.
The promise of Big Data analytics is becoming a realization, and the pace of innovation is accelerating; but this landscape requires a methodical, deliberative approach to achieve desired results. To be successful, the following components are essential:
The U.S. healthcare industry is experiencing a seismic shift under the Patient Protection and Affordable Care Act (ACA) and other market dynamics. The end result is increased attention to managing risk, gaining efficiency, improving quality, and reducing costs. As payers are no longer allowed to actuarially cherry-pick members, they are seeking to mitigate increased risk by seeking various forms of value-based payments with providers. However, providers are seeking contractual arrangements that limit risk exposure, especially in the short term, as they build capacity and expertise in deploying and using analytics-based solutions for population health management, clinical decision support, and care coordination.
The Centers for Medicare & Medicaid Services (CMS) had been piloting certain forms of value-based payments long before the passage of the ACA, but in a recent statement, CMS has indicated that 50% of reimbursements will be based on quality and outcomes by 2018. Private payers are not far behind: Aetna has established a goal of 75% of reimbursements via value-based contracts by 2020. A group of the top U.S. health systems, payers, and stakeholders announced in January the formation of the Healthcare Transformation Task Force, a private-sector alliance aimed at accelerating the healthcare industry’s transformation to value-based care. As of January 1, 2015, according to Managed Care, there were 424 Medicare Shared Savings Program Accountable Care Organizations (ACOs) and 159 private ACOs.
These circumstances necessitate that both payers and providers leverage their vital information assets to improve clinical, operational, and financial performance. The recent trend toward payer-provider mergers, partnerships, and joint ventures seeks to address fragmented healthcare delivery, emphasize patient-centered care, and promote care coordination. With the expansion of high-deductible health plans and emphasis on healthy lifestyles and wellness, there is also more focus on patient-generated data and the consumerization of healthcare. Aligning the goals and incentives among payers, providers, and patients sets the stage for data sharing, integration, and harmonization and coordinated Big Data analytics strategy. To be successful and competitive, stakeholders must share data to leverage analytics for (1) risk stratification and identification of high-cost or high-risk patients by using predictive models; (2) performance monitoring and evaluation of key quality measures; (3) development of longitudinal patient records; and (4) establishment of market-based models for bringing greater price transparency, measuring variation in cost and quality, and delivering predictable outcomes.
Healthcare organizations across the United States are becoming more interested in Big Data analytics and a related suite of applications to improve care delivery, but barriers to adoption include lack of skilled resources, limited access to data, inability to integrate data, and privacy concerns. Moreover, with so many vendor solutions, there is a sense of confusion and frustration on how to establish an analytics capability. To address these concerns, EHR vendors are moving toward open, cloud-based platforms that allow easier access to data via Application Programming Interfaces (APIs) that connect data sources to facilitate integration of new features into existing applications and to develop innovative point-of-care applications that bring context to each patient encounter in real time. This capability can be extended to detect, correlate, and monitor events across the enterprise as they occur for pattern matching, anomaly detection, and aggregation of data in motion with data at rest to make it actionable and apply advanced analytics and predictive modeling.
A related development is the emergence of Fast Health Interoperability Resources (FHIR, pronounced “fire”), a proposed Health Level 7 standard describing data formats and APIs for exchanging health data. FHIR solutions are built from a set of modular components called “resources,” which can be assembled into working systems. FHIR offers improved interoperability via Web services and common protocols, allowing users to interact with clinical data in new and different ways. By capturing data in a standard format, FHIR also offers a streamlined approach to implement complex clinical workflows and execute transactions and rules within FHIR resource bundles, thereby improving evidence-based medicine to make accurate diagnoses and select the most effective care pathways.
A common practice among healthcare experts is to collect as many health data as possible to explore and discover the “unknown unknowns,” especially for research in genomics and clinical informatics. Big Data requires a new generation of scalable technologies designed to extract insights from very large volumes of disparate, multistructured data by enabling high-velocity capture, discovery, and analysis. This is now possible with scale-out, clustered compute platforms such as Hadoop and Spark, which utilize low-cost, commodity hardware for collecting massive amounts of data for low-latency batch processing and real-time analysis. In conjunction with clinical data warehouses, NoSQL and in-memory databases, and data discovery and visualization tools, it is feasible to establish a unified analytics platform over a multitenant environment with domain-specific enclaves to evolve from systems of record to systems of engagement. As these systems and the standards governing them become more mature, it will be possible to analyze data more efficiently—to identify adverse drug interactions, misdiagnoses, missed routine exams, healthy lifestyle recommendations, disease prevention, and others. The scope of analysis can range from the individual patient level to the entire healthcare system to several healthcare systems connected via a single insurer.
The promise of Big Data analytics is becoming a realization, and the pace of innovation is accelerating; but this landscape requires a methodical, deliberative approach to achieve desired results. To be successful, the following components are essential:
- Realization that the status quo will not be sufficient to exceed future standards of care and stakeholder expectations;
- Visionary leadership that empowers, inspires, and is not afraid to fail;
- A corporate culture that encourages collaboration and values ingenuity and creativity at all levels;
- A methodology that increases likelihood for successful implementations, based on due diligence, business case, value proposition, and evaluation criteria;
- Multidisciplinary teams of specialists who guide, build, deploy, and scale solutions;
- An imperative to share data with safeguards for privacy and security; and
- An embrace of next-generation solutions and technologies for data collection, integration, and consumption.
RSS Feed
