In 2013 there was some good progress made in the implementation of health information technology. Electronic health record adoption, interoperability, and health data exchange are making great gains. We have laid a good foundation where we can begin to realize the value of health IT by beginning to make use of population health management and health data analytics tools that will help us to transform our healthcare system. However, with government mandates around meaningful use, HIPAA security and patient data access rules, and ICD-10 facing already strained healthcare organization IT departments, there are significant disparities in the ability to implement these critical business intelligence tools. 2014 will have some great opportunities and some difficult challenges, particularly in the area of analytics.
Growth of Accountable Care Organizations and Patient-centered Medical Homes
The next year will see a marked shift in both healthcare delivery and payment away from volume and towards value. This shift in healthcare spending will lead to a focus towards diagnostics and preventative care as new payment models like Accountable Care Organizations (ACOs) and delivery mechanisms such as Patient-centered Medical Homes flourish. This will require the ability to leverage data and analytics in order to evaluate risk factors, identify patients within a population requiring specific intervention, and provide improvements in coordinating care. New care coordination models will involve increased data collection from remote patient monitoring technologies as well as data aggregated from clinical systems and data sharing among providers across the care continuum. A focus on leveraging data from multiple clinical information sources is key to moving from a system spending large sums in acute care settings to one more focused on early intervention to preserve health, targeting at-risk patients who generate high costs.
Prepare for Value-based Purchasing
This next year will be a perfect storm of regulatory change, business drivers, and technology solutions that will make 2014 the year of health data analytics. There will continue to be a deep need for analytics capabilities to maximize payments under Value-based Purchasing (VBP) arrangements. The Patient Protection and Affordable Care Act (ACA) established the VBP program, and it will certainly be a factor in the coming year. Aggregating both clinical and financial data is necessary to succeed in VBP, which will pay for the quality of care and not the quantity. However, using these data is not as straightforward as it may seem. Significant barriers exist to leveraging data effectively to drive value-based decision-making. Four good steps to prepare for VBP are:
- Assess your current performance
- Implement education programs
- Develop a healthcare analytics strategy
- Identify areas for clinical quality and cost improvement
Taking these steps, with a sharp focus on the technology tools that will enable this strategy, are key to success in the changing healthcare reimbursement landscape. Adopting new technologies that use advanced data analytics will help to identify and process information in order to make more successful decisions and improve outcomes and value. There are, however, still some significant challenges ahead.
Focus on Population Health Management
Since a major focus of these new care and reimbursement models is population health management, success will depend on the application of data analytics to continuously monitor the population, identify system risk factors, stratify patients by risk, and target those patients requiring intensive care management. Healthcare providers will need to determine which of their patients both individually and as a cohort are at greatest risk in order to better manage costs. This will require a dedicated focus on data analytics which combines clinical data with financial and administrative data. Simply focusing on clinical data will not be enough. Organizations that maintain profitability in the new healthcare marketplace will need tools to marry clinical and financial data. A data set which includes clinical, financial, operational, and patient experience data will be needed to determine the quality/cost equation.
Dealing with data in clinical information systems which is stored in an unstructured manner will be another barrier. Unstructured data has no pre-defined and identifiable structure. Successful solutions will need the ability to analyze both structured and unstructured information in patient records. This can create a challenge as the bulk of the data in current electronic health record systems is unstructured including registration forms, physician notes, discharge summaries and assessments, and a massive volume of images. Extracting these data is difficult, and because unstructured data is difficult to locate and extract, it is often not included in business and clinical analysis for effective decision-making. Natural language processing (NLP) can help, but is not sufficient.
Success Depends on Healthcare Data Analytics
Analytics tools can detect trends and patterns, identify deviations, and help determining relationships. Predictive analytics solutions can generate and evaluate hypotheses, and determine a confidence level for the hypotheses. But comparative analytics, predictive analytics, and NLP will not solve all of health care’s problems. We’ve had these solutions for years in health care and still have seen increased costs, inefficient care delivery, and declining quality compared to other developed nations. You need outcomes data, as well as all of the data types mentioned above, in order to succeed. A successful organization must have tools with the ability to score predicted outcomes to better guide the care team on the need to intervene, when and how to intervene, and a feedback loop to create a learning healthcare system. 2014 will be the year of health data analytics.
Find out where your organization sits on the Healthcare Analytics Adoption Model.
What do you see happening in health care during 2014? Does your organization have analytics capabilities in place?
How would you rate this article?