The U.S. healthcare system has been moving toward a system that rewards positive outcomes while attempting to reduce or eliminate unnecessary services, and the Medicare 5-Star Program is part of the efforts of the Centers for Medicare and Medicaid Services (CMS) to define, measure, and reward quality healthcare.
The system measures how well Medicare Advantage and prescription drug (Part D) plans perform in several categories spanning quality of care, satisfaction, member retention, and customer service. Ratings that range from one to five stars (five being the highest) are assigned to individual categories, and one overall Star rating is assigned to summarize the plan’s performance as a whole.
Because overall Star ratings can change year to year and consumers are free to choose a new plan each year, consumers are encouraged to review coverage, plan costs, overall premiums, as well as ratings as part of their annual enrollment decisions. Clearly, a plan’s Star rating can influence consumer decisions regarding enrollment.
But the Star rating system also plays a role in reimbursements to plans. In 2010, the Affordable Care Act (ACA) introduced a quality-based payment structure for Medicare Advantage plans. CMS uses the Star ratings to reward highly rated plans with higher payments, and Medicare Advantage and MA-PD plans with a rating of four or more stars receive quality bonus payments (QBPs). Plans are required to use to use QBPs to provide extra benefits, such as eyeglasses, transportation to/from medical appointments, etc. Plans with higher ratings that receive QBPs should be able to offer more attractive benefits than their competitors, which in turn leads to greater enrollment and even higher quality ratings.
High Star ratings are also important for reasons unrelated to payments: plans with five stars may market to and enroll Medicare beneficiaries year-round and are not limited to a specified open-enrollment period.
Data Analysis in Patient Engagement and Population Management
One of the most complex challenges facing Medicare Advantage plans in the context of Star is delivering positive outcomes across disparate health and clinical domains. The most effective way to deploy a population-level, Star-focused engagement strategy is though big data engagement analytics. Big data engagement analytics enables plans to identify behavioral risk for each member proactively and prioritize members and measures based on their potential impact, and deliver personalized communications that address a member’s specific barriers to engagement.
For example, a member may be at high risk for non-compliance with preventative screening and chronic care measures because of health literacy issues, but at very low risk for disenrollment, dissatisfaction, and re-admissions. In this simple example, a directed communication focused on educating the member on self-care and prevention is the best and most efficient communication because it addresses the area of greatest risk in a manner that has the highest probability of generating sustained behavior change.
Most members, however, do not fall neatly into a single segment. A key component of any big data approach to improving Star ratings, therefore, is the ability to predict accurately member behavior across a variety of different measures. These types of big data models are designed to use a range of different data inputs, including member enrollment history, clinical profile, health-seeking behavior, utilization patterns, physician profiles and relationships, member engagement, social/economic profiles, and many other factors. Plans should strive to include as many data inputs as possible. Additional data boosts accuracy and provides valuable insights into potential behavior.
In addition, designing these models as dynamic systems enables predicting outcomes of interest at various times of the year and as members’ experiences and behaviors change. Predicted preventive screening behavior at the start of the year is different from at six and nine months after the start of the year, and it is critical to re-evaluate risk continually to maximize adherence to evidence-based guidelines.
Finally, big data helps determine the most effective communication channel by predicting member-specific channel engagement and effectiveness. This provides an even greater opportunity for behavior change because it further personalizes the outreach.
Creating a Continuous Learning Model
Part of implementing a big data approach to Star population health management is being prepared for some sort of failure. Failures are opportunities to understand what did and did not work and to analyze the impact of programs. This understanding enables program modifications and re-crafting that creates sustained success. An initiative to improve diabetic medication adherence, for example, may have failed to improve adherence in the aggregate, but it is critical to analyze each member segment and sub-segment and continually retain or discard certain portions of the program to improve performance.
A big data approach to Medicare Star offers a range of advantages, from greater efficiency in resource deployment to improved patient engagement and outcomes to improved Medicare Star ratings. Some results that Decision Point clients that have realized by deploying this type of approach include the following:
- Improved retention. Member engagement efforts led by big data engagement models resulted in a 21 percent improvement in disenrollment rates in less than four months.
- Lower re-admission rates. The best models can proactively identify more than 70 percent of readmissions by targeting less than 10 percent of members, which has consistently resulted in a 1 percent annual reduction in re-admissions rates.
- Improved compliance. Targeting members at high risk of compliance issues with engagement initiatives has resulted in breast cancer screening rates improving by more than 5 percent in one year compared with a control group, and also improved comprehensive diabetic care measures by 5.25 percent, to 10.3 percent (depending on the measure), in one year compared with a control group.
Saeed Aminzadeh, founder and CEO of Decision Point Healthcare Solutions, has more than 25 years of health information technology experience. He has held key senior management positions at Eliza Corporation, Ingenix (currently OptumInsight), IHCIS, and ProVentive, where he built and managed successful sales and marketing organizations focused on revenue growth and expansion of client relationships. Saeed is a graduate of The Johns Hopkins University with a BA in Economics.
Subscribe to Data Informed for the latest information and news on big data and analytics for the enterprise.