The healthcare industry has lagged behind retail, banking and other sectors in the adoption and use of data to improve efficiencies and services. Recently though, healthcare has seen increased interest in this approach as new payment models and reform initiatives require greater internal visibility and external transparency.
In addition, only recently have technology and data become available allowing people to really analyze what is going on, and make improvements in how we approach health and care management.
The demand for greater efficiency, effectiveness and quality in care creates a need for data-driven strategies that empower physicians, engage individuals and enhance organizational ability to deliver on the triple aim of improved population health, patient experience and affordability.
At the same time, data mining and analytics are becoming more powerful. The volume and variety of information now available along with the exponential growth of computing capacity now require organizations to change their approach in order to transform big data into actionable insight. For example, within the healthcare industry, both the volume and variety of data are increasing as providers are required to adopt electronic health records (EHRs). Unlocking the power of these massive amounts and different kinds of data will require new data science capabilities that can potentially bring incredible value to care delivery.
New technologies predict risk better, enhance care coordination, and improve outcomes
Financial analytics have been used by healthcare for decades, mainly to manage a practice or hospital. These practice management systems now focus on business matters, such as billing and payment. Advanced analytics can be used to process volumes of different kinds of data, including claims, medical, pharmacy and self-reported information, to predict not only future risk but support the clinician with specific actions to take.
For example, advanced analytics can identify individuals most likely to be hospitalized or determine whose current lifestyle will likely turn into a chronic condition, and provide specific actions to avoid a bad outcome. This information can be used to pinpoint at-risk persons who would benefit from specific interventions. These might include disease management, complex care management, wellness programs, or transitional care, among others.
Advanced analytics are now being used by payers to improve care coordination and management. This is a significant opportunity given nearly one third of Americans have two or more chronic conditions, and individuals with chronic diseases drive more than 75 percent of healthcare costs. Many of these patients constantly move from one care setting to another. They are supported by multiple providers as well as individuals within their own social support networks, which means there are many points of data entry about a person's health and opportunities to coordinate and improve care.
Until recently, the data derived from these interactions were siloed and locked in different systems, making it impossible to truly coordinate care based on an individual's changing health status. This resulted in gaps in care, unnecessary hospital admissions, duplication of efforts and adverse events. Predictive analytics develops new algorithms to meet these challenges, and new systems can query a wide variety of data and generate dynamic care plans that update automatically as a person's health or treatment changes. These real-time updates can be shared among a patient's entire care team to improve communication and collaboration.
Taking analytics to the next level--best practices from other industries
If we look at the new capabilities from other industries, highly-advanced algorithms are now being used in "recommendation engines" built by major retailers, especially those with a large online presence. These engines filter through vast amounts of data to identify and correlate the most important factors influencing consumer preferences and behavior. As this same approach is applied to healthcare, engines can filter out irrelevant information and pinpoint the most important data that should be used to influence care recommendations. For example, advanced algorithms can scour through a patient's medical history, health conditions, phenotype, genotype, lifestyle and other variables. Based on only the most impactful information, data engines can then identify an optimal treatment plan.
Given the immense variation in patient characteristics, the growing body of medical literature and number of treatments available, it is impossible for the human brain to process all of this healthcare information and turn it into useful insight. Fortunately, today's advanced analytics can be leveraged to accomplish this goal. When applied to the growing volume and variety of data, these tools better predict and manage risk, while helping providers determine the best course of treatment based on every individual's unique needs, and enabling patients to be equal partners in their own care.
Ken Yale, DDS, JD, is vice president of clinical solutions, ActiveHealth Management and author of predictive analytics tutorials in the upcoming book "Practical Predictive Analytics and Decisioning Systems for Medicine."