There’s an unaddressed problem in the industry for Medicare Advantage Part D (MAPD) plans. That problem is the ambiguous role of “predictive analytics.” Well today, we’re throwing down the gauntlet and assessing the purpose and value of predictive analytics.
What is Predictive Analytics Supposed to Do?
The deceptively simple answer is: “predict.” Unfortunately, the term “predictive analytics” has become so commonplace in the healthcare industry that it almost holds no meaning. Every vendor, marketer, and data team out there – especially in the health tech space – has some version of it. Sometimes it’s just one or two people making use of Google Analytics dashboards. Sometimes it’s full-blown departments of people calculating extensive statistical analyses based on publicly available decades of EHR or CMS data. What’s right, then? What is predictive analytics really supposed to do for health plans?
The fault in this question lies in thinking of “predictive analytics” as the sole answer to a given problem. In reality, it’s a step in a deeply involved processing model. After many mechanisms, the output – the prediction – comes to life. When done right, it’s rooted in large, relevant datasets that use historical information paired with past and current trends to offer up a realistic view of the future, inform a plan of action, and drive day-to-day operations through targeting and prioritization. And that is what it’s really supposed to do.
So Can Predictive Analytics Actually Improve Star Ratings?
Yes, but that doesn’t mean it should be the only tool you use. MAPD plans should use predictive analytics (or find partners who do – and do it well) to get insight into what’s ahead. However, it is important to remember that using this practice to change Star Ratings also depends on what plans do next. Let’s take a look at the ideal sequence of events that can enable plans to beat cut points via the use of predictive analytics and follow-on solutions.
(For this example, we’ll focus on the triple-weighted and entirely influenceable CMS Star Ratings medication use measures.)
- MAPD plans employ a predictive analytics model to generate an advanced understanding of potential pitfalls and opportunities across Star Ratings medication use measures. The output is a report showing which members are nonadherent and which are at risk for becoming nonadherent but can modify their behavior to improve adherence and plan Star Ratings.
- Using the insights found through analytics, plans or their partners should select tactics that minimize administrative efforts, maximize resources, and align with the plans’ performance and business goals. Intelligent interventions, which include provider activation and interactive voice response (IVR) calls transferring to retail pharmacies, have shown promise in both increasing member engagement and adherence—one by reaching members through a trusted care provider, the other by expediting the process for medication refill. Blending the most appropriate tactics allows plans to use cost-effective approaches on the members identified as most critical to improving population-level adherence.
- After solutions are applied, ongoing monitoring and reassessment is used to course-correct throughout the year. This includes reidentifying high-value target members and which outreach or intervention method will be most beneficial to those members.
- Repeat steps one through three.
But Can Health Plans Get ROI from Predictive Analytics?
Absolutely. Which brings us back to our original claim. Predictive analytics must be better integrated into a quality improvement strategy for health plans. Although there is certainly value to using data retroactively, predictive analytics adds more value by helping to anticipate, plan for, and/or avoid potential pitfalls in adherence strategies. In tandem with the right tactics, it drives better planning and program execution, and as a result—more desirable outcomes and ROI. Predictive analytics—and in particular, predictive analytics with the right simulations, regular reporting, and intelligent interventions—enables plans unlock star bonuses and ROI. This makes predictive analytics what it was designed to be: a step in an integrated and ongoing cycle.
You can be one of those plans. By applying forecasting to Star Ratings measures early in the year and complementing predictive analytics with the right tactics and tools, you really can make a difference to your members and your bottom line.