Population health management and the power of predictive analytics

Each year in the United States, seven out of 10 deaths are related to chronic disease. Healthcare is shifting from a sole emphasis on the individual patient to focusing on both the individual patient and groups of patients. The convergence of advances in health information technologies and a focus on groups of patients has set the stage for population health management. 

Defining population health management

There is not an agreed upon definition of population health management. The Patient Protection and Accountable Care Act, along with the creation of accountable care organizations (ACOs), encouraged group practice leaders to identify and create a process to manage, track and monitor patient populations throughout the healthcare and palliative care continuum. Population health management represents a proactive way to improve the health outcomes of a group across the continuum of care by monitoring and identifying individual patients within that group at the lowest necessary cost.

Medical groups must make the shift from individual patient management to implementing strategies to manage their patient populations more effectively. Medical groups are at the tip of the spear when it comes to improving patient care and aiding in the move to value-based care from fee-for-service. Increasingly, reimbursement is being tied to the effective management of population health. The Department of Health and Human Services recently set a goal to have value-based care account for 50% of all Medicare provider payments by 2018.

The backbone: Health information and communication technology 

As the country moves to a value-based care model, health systems and medical groups are focusing their attention on high-risk patients outside the four walls of the practice to prevent readmission, as well as identifying those patients who may be at risk for an adverse event upon admission. The challenge for medical groups is how to collect and leverage different data elements from various EHR vendors. To date, there is no clear standardized way to extract that data into one single source suitable for stratifying risk and developing strategies to manage both clinical and financial risk. Figure 1 illustrates the number and variety of data feeds available to a provider for all patients who enter a medical group. Aggregating large amounts of data for the creation of actionable intelligence is resource-intensive. 

Technology will be the backbone of population health management. By leveraging health information technology, group practices can provide more effective and efficient care while also predicting utilization and other metrics tied to quality of care and reimbursement. These predictive tools fall under the emerging science and practice of predictive analytics. Armed with such technology and tools, clinicians can move away from simply managing diseases to preventing diseases and educating patients. When this data is presented in a way patients understand it, the patients increase their ability to manage their own health.

Predictive analytics: Leveraging data to drive the Triple Aim

Predictive analytics gives providers the insights to identify healthcare risks and proactively drive patients to the right treatment options. Imagine a patient seeing their primary care physician (PCP) and, based on six months of data, the provider can stratify that patient to the appropriate specialty for care, eliminating the need for a hospital visit. Another illustration of predictive analytics at work in medical group settings is to identify heart failure patients who have stopped filling their diuretic medications and then follow up to determine and address the barrier.

Medical groups play an important role in improving the quality care. Providers have only scratched the surface through real-time data in identifying patients upon admission who are at risk for a heart attack, patients at risk for an unplanned hospital stay or even patients who have been non-adherent to their treatment regimen. The power of predictive analytics will allow medical groups to address the populations directly attributed to them. Better care, improved health and lower costs are the pillars providers strive for. Relying on retrospective results and claims data is too late. With the large amount of unstructured data coming from the provider’s EHR, algorithms and solution sets will need to be developed either by the provider or a third-party vendor so the front-line caregivers have the insights necessary to manage populations more effectively and efficiently in the hospital, ambulatory care and home settings. 

The ACO’s role: Changing organizational structures and predictive analytics

The use of predictive data allows for better patient engagement and encourages medical groups to point the appropriate resources to the areas of top importance. By having this data at their fingertips, physicians and advanced practice clinicians will be able to pinpoint gaps in care quickly and efficiently, and move to more proactive patient care.

ACOs and integrated delivery networks are being asked to manage a large, diverse patient population. ACOs are garnering more interest among hospital, medical and health plan executives, as well as specialists and, in some cases, even retail clinics such as CVS or Walgreens. The number of lives that will be directly associated with coordinated care is growing at a staggering rate. Consulting company Leavitt Partners, Salt Lake City, projects ACOs will serve 107 million covered lives by 2020. This represents more than a fourfold increase over four years. Bottom line: There will be more patients in the healthcare system and more ACOs to treat them. Medical groups and providers will need to prepare for this paradigm shift.

Building a predictive analytics strategic roadmap

There is no one-size-fits-all approach to implementing predictive analytics. However, to be successful, you will need a well-defined strategy and roadmap to address the ever-changing regulatory requirements. This strategy must be robust enough to ensure quality data mapping and normalization has occurred. A well-defined strategy should begin with these five steps:

1. Align patient care with clinical data integration and quality metrics.
2. Engage in the proper infrastructure to prepare for changing payment models.
3. Identify additional care pathways for your patient population and empower patients to be involved in their own care.
4. Understand the distribution of cost across your patient population.
5. Eliminate silos of information to build a strong culture of collaboration among caregivers.

Conclusion

The lens through which medical groups and providers are seeing is ever-changing, and there is a need to be flexible, nimble and strategic in how predictive analytics will be deployed. You are only as strong as your least-innovative medical group or clinician in the system. Predictive analytics is going to be the force that changes healthcare for generations to come, providing meaningful context for diagnosis, treatment and quality care for patients.