The Medicare Advantage Program has seen rapid growth over the last few years. Nationally, almost 50% of Medicare beneficiaries are enrolled in a MA plan. The competition for new membership is intense as some MA plans and new entrants to the market have enriched their benefits. The enrichment has especially been directed towards offering supplemental benefits not covered by traditional Medicare, such as dental or vision. Some of these supplemental benefits such as nutrition, transportation, and in-home support services, are intended to address the social determinants of health. This is a welcome development and has the potential to improve health outcomes for at-risk beneficiaries.
While growth has had a great deal of focus, there are areas that need attention. Member retention in current plans is one of them. In the current environment of aggressive marketing and a substantially high cost of member acquisition, it is critical that plans develop a strategy to retain existing market share. High disenrollment rates can have adverse outcomes for both patients and plans. It is disruptive to the continuity of care which is particularly important for patients with multiple chronic conditions. For plans it reduces their return on investment (ROI) in population health programs such as annual wellness visits (AWV), vaccinations, and disease management. A robust ROI on these population health programs is essential to financial success and sustainable growth of an MA plan.
Chart 1 below shows actual annual disenrollment rate in a general enrollment plan. For this plan, the disenrollment rate for full dual members is almost 3 times as high as non-dual members.
Oliver Wyman's Actuarial Practice has developed a member-centric analytical solution to “hot spot” disenrollment trends and leverage machine learning (ML) models to identify characteristics of the population with high likelihood of disenrollment. Further analysis is conducted to stratify these members into risk cohorts to support retention outreach or other business interventions to improve member retention.
There are a variety of models to use for this purpose. However, caution is needed when selecting models, as we are ultimately trying to understand the variables, or interaction of variables that are statistically significant and are a key driver of disenrollment. The more complex the model gets, the less interpretable the results become.
For the ease of explaining factors driving disenrollment, we have used a multivariable logistic regression model. Based on our analysis of data, we have identified AWV, Dual Status, and utilization of supplemental benefits as explanatory variables. The results show that all these variables are statistically significant.
The predicted disenrollment rate for patients with AWV is much lower than patients without AWV – see Chart 2 below. Since AWV is also associated with improved patient outcomes as well as complete and accurate coding and documentation, higher AWV rate is likely to yield lower cost and higher risk adjusted revenue along with higher retention.
For dual eligible patients in a general enrollment plan, the predicted disenrollment rate is almost 2.5 times as high as non-dual patients. It appears that in addition to AWV, a strategy to effectively address the social risk factors and member engagements through the use of member touch points (HRAs, house calls) is needed to fully address high disenrollment rates of dual members.
At minimum, a benefit design that facilitates easier access to care such as non-emergency transportation would be helpful in reducing the disenrollment rate of dual beneficiaries. In the long term, a more coordinated effort that targets social determinants of health via a care team that includes local providers and social workers might be needed to improve the retention of dual eligible members.
While there has been a significant shift towards offering supplemental benefits which makes the benefit more attractive during sales, perhaps not enough attention has been given to the impact of such benefits on member retention. The model points to the utilization of the supplemental benefit as a significant factor in reducing disenrollment. This suggests that a more accessible supplemental benefit design, and creating more awareness about the availability of these supplemental benefits should be another part of strategy to improve retention rate.
Oliver Wyman’s Three-Pronged Approach
We have developed a three-pronged approach to address the issue of member disenrollment. We take the enrollment, claims, and any member touchpoint data related to patient experience with plan or their provider network, and conduct a rapid assessment by running the data through our machine learning enabled Data Extraction and Enrichment Pipeline (DEEP). This helps us to –
- “Hot Spot” disenrollment trends by various member characteristics, delivery systems, benefit utilization variables etc. and determine the opportunity to lower disenrollment rates.
- Execute rapid predictive models to identify cohorts with high propensity to voluntarily disenroll
- Monitor the financial impact of any retention outreach initiatives
Member disenrollment is a key driver of financial performance of a MA plan. There is an additional contribution margin associated with each retained member. It is even more important for plans with significant investment in operational initiatives (care management programs, accurate and complete coding and documentation initiatives), to have a clean analytical process in place to manage retention effectively. In addition, high disenrollment rates are likely to result in poor performance on patient experience measures and low ratings of health plan. A robust retention strategy will enable plans to advance innovation, achieve sustainable growth and profitability.