11436 SSO

Big Data Predictive Analytics in Action: Healthcare/Medicare

Feb 25, 2014

For healthcare payers, the Affordable Care Act (ACA) can be a double-edged sword. The law, which was enacted to improve health care affordability and quality, enforces more stringent reporting rules that require payers (insurers) to monitor quality of care, performance metrics, and member satisfaction. But high performance and quality can earn healthcare payers a coveted five-star rating on their Medicare Advantage products.

Under the ACA, bonus payments, payments in future years, and member decisions to enroll in a plan are all dependent on health payers’ star ratings. Not surprisingly, there’s a big incentive to improve this rating, and maintain a high one. Predictive analytics can play a very important role in this.

The challenge: catching falling stars

Medicare Advantage plan ratings are affected by many factors: keeping people healthy via preventive services such as screenings and vaccines; effectively managing chronic conditions; offering high quality care and being responsive; the number of complaints, appeals, and voluntary disenrollments; and the quality of telephone customer service. Some of these factors are influenced by the provider; some, by the payer.

Regardless of what drags down a star rating, the Centers for Medicare and Medicaid Services (CMS) holds the payer responsible. And herein lies a challenge.

Insurers don’t have the personal relationships with members that point-of-care providers do; any issues members have with their coverage or customer service typically only surface if members decide to share their complaints with either Medicare or the payer. Unfortunately, that’s too late, for both members and payers.

Complaints are tracked by CMS in their Complaint Tracking Module (CTM). The only option for payers is to react after the fact. In some cases, they might only have a couple of days to do so. Yet the complaint has already been filed with the CMS, so it can impact a star rating regardless of how the problem is resolved. Equally unfortunate is the fact that by this point, it becomes harder for payers earn back their members’ trust.

But what if insurers had the opportunity to address member issues before they boiled over into a complaint?

Predictive analytics: time to intervene

Predictive analytics and big data offer just such an opportunity; they can help payers predict which members are likely to complain, before they complain. This gives payers sufficient time to intervene and reach out to the potentially dissatisfied member. Reaching out proactively sets a positive tone too: it shows the member that the payer cares.

In many cases, this element can make all the difference. Reaching out to members before they complain not only prevents an issue from becoming a complaint, but it also increases member satisfaction. In addition, the results of the outreach—in particular, the notes captured by the customer service representative during the outreach—serve as new data that can help further fine tune predictive models. This helps the system learn and become smarter, setting up a virtuous cycle.

Insights for CTM: higher ratings, healthier members

Apigee, with our Insights for CTM product, has worked with leading healthcare payers to predict and prevent complaints, grievances, and appeals. Our powerful predictive analytics has been proven to predict complaints 6x more accurately than traditional analytics approaches, and predict them well before they would occur.

We’ve helped healthcare payers investigate the root cause of a complaint, execute outbound intervention programs, and report on the results. With Insights for CTM, some of the largest healthcare payers have reduced complaints to Medicare by over 50% within the first year of deployment.

The net result of fewer complaints is higher overall star ratings, larger bonus payments, and a growing community of happier, healthier members.

For more information, check out the Apigee Insights for CTM data sheet

image: Medisave UK/Flickr


Scaling Microservices