Do People Who Sign Up for Medicare Advantage Plans Have Lower Medicare Spending?

Overview

Medicare payments to Medicare HMOs and PPOs, known as Medicare Advantage plans, have always been based on Medicare spending by similar people in traditional Medicare, partly because Medicare has never had accurate, complete data on the use of services or health care spending for beneficiaries in Medicare Advantage plans.¹ The assumption has been that adjusting payments to plans for health status and other factors accounts for differences between beneficiaries in traditional Medicare and those in Medicare Advantage plans. Profits are assumed to be due to plans reducing spending by either managing fees (e.g., by having lower-cost hospitals in their network) or changing patterns of care (e.g., reducing hospital readmissions), rather than to favorable selection. Nonetheless, selection bias has been an ongoing concern and the subject of many studies over the years, with mixed evidence of favorable selection.^2,3,4,5,6 This question is important because it affects the accuracy of Medicare payments to plans on behalf of 20 million Medicare beneficiaries, and rising.

This is the first known analysis to examine whether beneficiaries who choose to enroll in Medicare Advantage plans have lower spending and use fewer services – before enrolling in Medicare Advantage – than similar people in traditional Medicare. If Medicare Advantage enrollees use fewer services and have lower Medicare spending before they enroll in Medicare Advantage plans, compared to similar beneficiaries in traditional Medicare, then basing payments to Medicare Advantage plans on the Medicare spending for similar beneficiaries in traditional Medicare would overestimate the expected costs of Medicare Advantage enrollees and overpay plans by billions of dollars over the next decade. Studies that have looked at differences in the use of services and Medicare spending for Medicare Advantage enrollees compared to beneficiaries in traditional Medicare that did not account for actual prior differences may have overestimated the extent to which plans are reducing enrollees’ spending or use of services.

To address this question, we examine Medicare Part A and B spending and service use for traditional Medicare beneficiaries in 2015. We compare average Medicare spending and use of services for traditional Medicare beneficiaries who enrolled in Medicare Advantage plans versus those who remained in traditional Medicare in 2016, after adjusting spending values for health conditions and other relevant factors (Figure 1). We examine how the results differ across demographics, chronic conditions, and counties, and also examine how the results change when Part D spending is included. The analysis is based on a five percent sample of Medicare claims data and excludes beneficiaries who may not have been active choosers in 2016; more details about the analysis are included in the Methods.

Differences in Medicare Spending

Among beneficiaries in traditional Medicare in 2015, those who enrolled in Medicare Advantage in 2016 had spending (for Part A and Part B) that was $1,253 lower (13% difference), on average, than beneficiaries who remained in traditional Medicare in 2016, after adjusting for health risk factors (Figure 2; Tables 1 and 2).⁷

When Part D spending is included, the results changed only slightly. Traditional Medicare beneficiaries in 2015 who switched to Medicare Advantage in 2016 had total Medicare spending (including Part D) that was 15 percent lower than spending for beneficiaries who remained in traditional Medicare in 2016.

Comparison to Other Payments Received by Medicare Advantage Plans. To put the difference in Medicare spending in context, the $1,253 average difference in spending is nearly four-times larger than the average per capita quality-based bonus payment ($336) paid to Medicare Advantage prescription drugs plans that qualified for bonuses in 2015 (Figure 3). The average difference in spending is also more than twice as large as the average annual premium paid by Medicare Advantage enrollees in 2015, including enrollees in plans with no premium.

Differences in Medicare Spending, by Demographics

Traditional Medicare spending in 2015 was lower for beneficiaries who enrolled in Medicare Advantage plans in 2016 than for similar beneficiaries who remained in traditional Medicare that year, by age and gender, and among beneficiaries dually eligible for Medicare and Medicaid, after adjusting for health risk and other factors (Table 1).

• Age: The difference in average traditional Medicare spending in 2015 among beneficiaries who switched to Medicare Advantage in 2016, compared to those who remained in traditional Medicare, was evident for beneficiaries of all ages, and increased with age for beneficiaries over the age of 65, after risk adjustment. For example, among beneficiaries ages 65-69, average traditional Medicare spending in 2015 was $1,119 lower among beneficiaries who switched to a Medicare Advantage plan in 2016 than for similar beneficiaries who remained in traditional Medicare; among beneficiaries ages 85-89, the difference in spending was $1,314. This finding suggests that selection bias, and the associated potential overpayments, may increase with age.
• Gender: The average difference in spending between the two groups was similar among men and women ($1,271 and $1,247, respectively).
• Dual eligibility for Medicaid: Traditional Medicare spending in 2015 for Medicare beneficiaries with full Medicaid benefits (full dual eligible) who enrolled in Medicare Advantage in 2016 was $1,142 lower, on average, than spending for similar full dual eligibles who stayed in traditional Medicare in 2016, after adjusting for health and demographic factors. Similarly, partial dual eligibles who enrolled in Medicare Advantage in 2016 had traditional Medicare spending in 2015 that was $1,162 lower than spending for those who remained in traditional Medicare in 2016, after adjusting for risk factors. In other words, among dually eligible beneficiaries – a group of beneficiaries with relatively high Medicare spending – those who used more services and incurred higher Medicare spending in 2015 were more likely to remain in traditional Medicare in 2016 while dual eligibles with lower service use and spending were more likely to enroll in a Medicare Advantage plan in 2016.
• Institutional status: Among Medicare beneficiaries living in institutions, such as nursing homes, traditional Medicare spending in 2015 was $1,825 lower among those who enrolled in Medicare Advantage plans in 2016 than among similar institutional residents who stayed in traditional Medicare that year. If higher-cost nursing home residents are remaining in traditional Medicare while lower-cost residents are moving to Medicare Advantage plans, it could make it easier for Medicare Advantage plans serving the nursing home population to be profitable, which may explain the relatively recent increase in firms offering Special Needs Plans for this population (I-SNPs).⁸ 

Differences in Medicare Spending, by Chronic Conditions

Even among beneficiaries with the same chronic conditions, those who enrolled in Medicare Advantage plans in 2016 consistently had lower Medicare spending in 2015 than similar beneficiaries who remained in traditional Medicare in 2016 (Table 1).

For example, among traditional Medicare beneficiaries with diabetes in 2015, those who enrolled in Medicare Advantage plans in 2016 had Medicare spending that was $1,072 lower in 2015, on average, than similar beneficiaries with diabetes who stayed in traditional Medicare in 2016, after adjusting for differences in health status (Figure 4). In other words, it would appear that lower-cost beneficiaries with diabetes are more inclined to enroll in Medicare Advantage than higher-cost diabetics. Likewise, traditional Medicare beneficiaries with asthma who enrolled in Medicare Advantage plans in 2016 had Medicare spending that was $1,410 lower in 2015, on average, than similar beneficiaries with asthma who remained in traditional Medicare in 2016, even after adjusting for health risk factors.

The difference in average, adjusted 2015 traditional Medicare spending between beneficiaries who subsequently enrolled in Medicare Advantage versus those who remained in traditional Medicare increases with the number of chronic conditions, rising from $226 among those with no chronic conditions to $1,629 or more among beneficiaries with 5 or more chronic conditions (Figure 5). This finding suggests that potential overpayments may be largest for the Medicare Advantage plans that are serving the sickest beneficiaries.

Differences in Medicare Spending, by County

In this section, we looked at whether the observed differences in spending and service use persist across markets, and the extent to which differences may vary from one market to another. We compared average spending in 2015 among beneficiaries who switched to Medicare Advantage in 2016 versus those who remained in traditional Medicare, without adjusting for other factors. We were not able to replicate the analysis by county with the adjustment for risk factors, such as health conditions and demographics, due to sample size constraints. For this analysis, we looked at 20 relatively large markets that vary geographically, and vary by Medicare Advantage penetration and payment quartiles.

Among large, urban counties, the differences in spending between Medicare Advantage enrollees and beneficiaries in traditional Medicare varied greatly across the country (Figure 6; Table 3). In some counties, such as Los Angeles, CA, San Bernardino, CA, Wayne, MI (Detroit), and Cuyahoga, OH (Cleveland), beneficiaries who enrolled in Medicare Advantage plans in 2016 had significantly lower traditional Medicare spending in 2015 ( ≥$3,000 lower) than beneficiaries in the county who remained in traditional Medicare in 2016.

In other counties, such as Allegheny, PA (Pittsburgh), Baltimore City, MD, Mecklenburg, NC (Charlotte), Erie, NY (Buffalo), and Multnomah, OR (Portland) beneficiaries who enrolled in Medicare Advantage plans in 2016 had higher prior year traditional Medicare spending ( ≤-$1,000) than beneficiaries in the county who remained in traditional Medicare in 2016. These differences across counties suggest that the selection bias into Medicare Advantage may vary across markets.

Discussion

This analysis examines whether beneficiaries who choose to enroll in Medicare Advantage plans have lower spending and use fewer services – before enrolling in Medicare Advantage – than similar people in traditional Medicare. The study found that beneficiaries who chose to enroll in a Medicare Advantage plan in 2016 had average expenditures in traditional Medicare (in 2015) that were $1,253 less, on average, than similar beneficiaries who remained in traditional Medicare. Similar differences in spending were found across all demographics and chronic conditions, even after adjusting for health risk factors. The results suggest that favorable self-selection into Medicare Advantage plans is occurring, even among traditional Medicare beneficiaries with similar health conditions. The findings raise questions as to why beneficiaries who are higher utilizers are less likely to go into Medicare Advantage and instead remain in traditional Medicare.

Other studies have examined services used by people while they were enrolled in Medicare Advantage plans, based on limited data, and have generally found that beneficiaries in Medicare Advantage plans use fewer services than those in traditional Medicare.^9,10,11 Notably, the authors of these studies almost universally attribute differences in service utilization to care management by the plans – rather than to pre-existing differences in care seeking behavior and use of health services. This study suggests that differences in health care use, and spending, are evident before beneficiaries decided to enroll in Medicare Advantage plans or remain in traditional Medicare, raising questions about the extent to which plans are actually lowering spending or managing care.

It is not clear whether the differences in spending observed in this study increase, decrease, or persist over time as beneficiaries age, which has implications for whether a similar difference in spending could be assumed for all Medicare Advantage enrollees.¹² Likewise, it is not clear how this difference in spending will change as the share of counties with the majority of beneficiaries in Medicare Advantage plans grows. This missing information could have important implications for Medicare spending. Potential overpayments could amount to billions in excess Medicare spending over a ten-year period if the observed differences in spending hold up as beneficiaries age and Medicare Advantage enrollment continues to rise. To illustrate, if the difference in average Medicare spending ($1,253) applied to just 10 percent of all Medicare Advantage enrollees in 2016, or 1.8 million enrollees, it would amount to more than $2 billion in excess spending in one year alone.

Policymakers could consider adjusting payments to reflect Medicare Advantage enrollees’ prior use of health care services, which could lower total Medicare spending and in turn reduce Medicare Part B premiums and deductibles for all beneficiaries. With more than 20 million enrollees in Medicare Advantage plans and Medicare payments to plans projected to reach $250 billion in 2019, the stakes are high for making payments to plans as accurate as possible.^13,14