Abstract
The costs of providing benefits to enrollees in private Medicare Advantage (MA) plans are slightly less, on average, than what traditional Medicare spends per beneficiary in the same county. However, MA plans that are able to keep their costs comparatively low are concentrated in a fairly small number of U.S. counties. In the 25 counties where the cost differences between MA plans and traditional Medicare are largest, MA plans spent a total of $5.2 billion less than what traditional Medicare would have been expected to spend on the same beneficiaries, with health maintenance organizations (HMOs) accounting for all of that difference. In the rest of the country, MA plans spent $4.8 billion above the expected costs under traditional Medicare. Broad determinations about the relative efficiency of MA plans and traditional Medicare can therefore be misleading, as they fail to take into account local conditions and individual plans’ performance.
Background
Medicare beneficiaries have a choice of traditional Medicare coverage or private Medicare Advantage (MA) plans through what is effectively a nationwide health insurance exchange. But the best way to balance the two alternatives continues to be a matter of debate. A central question is whether MA plans operate more efficiently than traditional Medicare—a question whose answer is more complicated than might first appear.
Comparing Medicare Advantage Plan Costs with Traditional Medicare Costs
Previous work has shown that MA plans’ costs of providing enrollees with Medicare benefits were, on average, slightly less in 2012 than traditional Medicare spending per beneficiary in the same county, though there was wide variation across areas and types of plans (Exhibits 1 and 2).1 Nationwide, the difference between MA plan costs and the amount traditional Medicare would have been expected to spend on the same beneficiaries was $378 million, or $43 per MA enrollee. But while MA plan costs per beneficiary were lower than traditional Medicare costs in urban areas, they were substantially higher in rural areas.
Exhibit 2. Total MA Plan Costs Relative to Local Costs in Traditional Medicare Nationwide,
in Rural and Urban Areas, and by Plan Type, 2012
| (A) MA enrollees |
(B) Annual MA plan costs per enrollee |
(C) Annual traditional Medicare costs per beneficiary |
(D) MA plan costs per enrollee as a percent of traditional Medicare costs per beneficiary (B/C) |
(E) Total MA plan costs relative to traditional Medicare costs ((B-C)*A) (in millions) |
|
| Nationwide | 8,829,576 | $9,370 | $9,413 | 100% | -$378 |
| By urban/rural location: |
|||||
| Urban | 8,422,171 | 9,344 | 9,452 | 99 | -911 |
| Rural | 407,405 | 9,915 | 8,607 | 115 | 533 |
| By plan type: | |||||
| HMO | 6,019,570 | 8,980 | 9,681 | 93 | -4,218 |
| Local PPO | 1,551,761 | 10,335 | 8,738 | 118 | 2,478 |
| Regional PPO | 764,888 | 10,307 | 9,210 | 112 | 839 |
| PFFS | 493,357 | 9,645 | 8,584 | 112 | 523 |
Notes: The cost per enrollee figures above represent the cost per enrollee with a typical risk profile—that is, these figures are, effectively, risk-adjusted.
HMO = health maintenance organization; PPO = preferred provider organization; PFFS = private fee-for-service.
The relationship between MA plan costs and traditional Medicare costs also varied across the four types of MA plans we examined. The most tightly organized MA plans are health maintenance organizations (HMOs), which usually limit coverage to care from a network of providers that are under contract with each plan. Enrollees in preferred provider organizations (PPOs) have the option to use providers out of network, but at higher cost (MA includes both local and regional PPOs). Private fee-for-service (PFFS) plans are the least structured type of plan: they are not required to establish networks of providers and cannot restrict enrollees from choosing any provider that agrees to accept the plan’s terms and conditions.
Of these four types of MA plans, only HMOs had lower costs than traditional Medicare in 2012. HMO costs per enrollee nationwide were 93 percent of traditional Medicare spending per beneficiary in the same counties—a $4.2 billion difference. Local PPO, regional PPO, and PFFS plan costs all were higher than expected costs under traditional Medicare—$3.8 billion higher in aggregate.
Geographic Concentration of Relatively Low-Cost Medicare Advantage Plans
To examine that variation more closely, we compared the total costs of MA plans in 2012 with those of traditional Medicare for each county. We then ranked the counties according to the aggregate difference between MA costs and expected costs for the same beneficiaries under traditional Medicare. The top-ranked counties are those where MA plans have the lowest costs relative to traditional Medicare, while the bottom-ranked counties are those where MA plans have the highest costs relative to traditional Medicare. Our analysis indicates that efficiencies in MA relative to traditional Medicare are concentrated in a small number of counties. Further still, these efficiencies are concentrated among HMOs in those counties (Exhibit 3).
Of the top 25 counties we ranked, MA plan costs in 2012 were a total of $5.2 billion less than traditional Medicare would have been expected to spend for the same beneficiaries. That difference is accounted for by HMO plans in those counties; aggregate costs for the other types of plans in those counties actually were slightly higher than in traditional Medicare. Although traditional Medicare costs per beneficiary in the top 25 counties are substantially higher than in the rest of the country, MA plans in these counties were not only less costly than traditional Medicare but also substantially less costly than MA plans in the rest of country. Again, HMOs drove those results, as they have much lower costs than the other types of MA plans in those 25 counties.
Outside the top 25 counties, MA plans spent $4.8 billion above the expected costs under traditional Medicare. Even HMOs in those counties spent $1.1 billion more than traditional Medicare would have been expected to spend. Traditional Medicare costs were much lower in those counties than in the top 25 counties, but MA plans had much higher costs.
Exhibit 3. Total MA Plan Costs vs. Local Costs in Traditional Medicare in 25 Counties Where Total MA Plan Costs Are Lowest Relative to Expected Traditional Medicare Costs for Same Beneficiaries, 2012
| (A) MA enrollees |
(B) Annual MA plan costs per enrollee |
(C) Annual traditional Medicare costs per beneficiary |
(D) MA plan costs per enrollee as a percent of traditional Medicare costs per beneficiary (B/C) |
(E) Total MA plan costs relative to traditional Medicare costs ((B-C)*A) (in millions) |
|
| Nationwide: | 8,829,576 | $9,370 | $9,413 | 100% | -$378 |
| 25 counties with greatest total difference |
2,055,060 | 8,378 | 10,915 | 77 | -5,214 |
| All other counties |
6,774,516 | 9,671 | 8,958 | 108 | 4,835 |
| HMO plans: | 6,019,570 | $8,980 | $9,681 | 93% | -$4,218 |
| 25 counties with greatest total difference |
1,849,784 | 8,113 | 10,967 | 74 | -5,278 |
| All other counties |
4,169,786 | 9,365 | 9,110 | 103 | 1,061 |
| Other MA plans: | 2,810,006 | $10,206 | $8,840 | 115% | $3,840 |
| 25 counties with greatest total difference |
205,276 | 10,766 | 10,450 | 103 | 65 |
| All other counties |
2,604,730 | 10,162 | 8,713 | 117 | 3,775 |
Note: The cost per enrollee figures above represent the cost per enrollee with a typical risk profile—that is, these figures are, effectively, risk-adjusted.
The “Top 25” Counties
The 25 counties with the largest total difference between Medicare Advantage and expected traditional Medicare costs for the same beneficiaries are geographically concentrated: 10 of those 25 counties are located in Florida and six are in California (Exhibit 4). Of the remaining counties on this list, three are the core counties of the largest metropolitan areas in Texas—Houston, San Antonio, and Dallas—while two more are the major counties of the Las Vegas and Phoenix metropolitan areas. The remaining counties in the top 25 include two in New York City and the largest counties of the Chicago and St Louis metropolitan areas.
In 21 of those 25 counties, traditional Medicare spending per beneficiary was greater than the national average, with Florida’s Miami-Dade County leading the list. However, MA plans in 20 of those 25 counties had lower costs per enrollee than the national average. This indicates that plans in these areas not only had an easier benchmark against which to compete but also found a way to be more efficient than plans in other areas. Using Miami-Dade County as an example, while traditional Medicare spending per beneficiary was 78 percent higher than the nationwide average, MA plans there had costs per enrollee that were 11 percent below the national average for MA plans. The data for most of the other counties on the list are similar, though less extreme.
Moreover, just as MA plan efficiency was concentrated in a few counties across the country, just three counties—Miami-Dade and Broward in Florida and Los Angeles in California—accounted for almost as much of the aggregate difference between MA and traditional Medicare as the other 22 counties combined.
Exhibit 4. The 25 Counties in Which Total Medicare Advantage Plan Costs Are Lowest
Relative to Expected Traditional Medicare Spending on Same Beneficiaries, 2012
| County (State) | (A) MA enrollees |
(B) Annual MA plan costs per enrollee |
(C) Annual traditional Medicare costs per beneficiary |
(D) MA plan costs per enrollee as a percent of traditional Medicare costs per beneficiary (B/C) |
(E) Total MA plan costs relative to traditional Medicare costs ((B-C)*A) (in millions) |
| Nationwide | 8,829,576 | $9,370 | $9,413 | 100% | -$378 |
| Total for 25 counties | 2,055,060 | 8,378 | 10,915 | 77 | -5,214 |
| Miami-Dade (FL) | 159,555 | 8,373 | 16,737 | 50 | -1,335 |
| Los Angeles (CA) | 313,292 | 8,517 | 10,871 | 78 | -738 |
| Broward (FL) | 102,023 | 7,303 | 11,984 | 61 | -478 |
| Clark (NV) | 82,296 | 7,553 | 10,890 | 69 | -275 |
| Palm Beach (FL) | 77,530 | 7,560 | 10,878 | 70 | -257 |
| Orange (CA) | 124,588 | 8,535 | 10,445 | 82 | -238 |
| Maricopa (AZ) | 150,223 | 8,562 | 9,762 | 88 | -180 |
| Pinellas (FL) | 67,303 | 7,755 | 10,420 | 74 | -179 |
| Harris (TX) | 95,938 | 9,697 | 11,541 | 84 | -177 |
| Riverside (CA) | 103,836 | 8,015 | 9,590 | 84 | -164 |
| Hillsborough (FL) | 54,175 | 7,471 | 10,265 | 73 | -151 |
| Bexar (TX) | 62,194 | 7,398 | 9,803 | 75 | -150 |
| San Bernardino (CA) | 77,259 | 8,203 | 9,842 | 83 | -127 |
| Orange (FL) | 33,745 | 7,657 | 10,412 | 74 | -93 |
| San Diego (CA) | 123,404 | 8,594 | 9,312 | 92 | -89 |
| Cook (IL) | 58,599 | 9,620 | 11,122 | 86 | -88 |
| Pasco (FL) | 40,100 | 7,316 | 9,415 | 78 | -84 |
| Volusia (FL) | 41,848 | 7,765 | 9,212 | 84 | -61 |
| Bronx (NY) | 37,656 | 10,290 | 11,898 | 86 | -61 |
| Kings (NY) | 66,615 | 10,265 | 11,060 | 93 | -53 |
| St. Louis (MO) | 43,988 | 7,568 | 8,766 | 86 | -53 |
| Dallas (TX) | 48,602 | 9,938 | 10,980 | 91 | -51 |
| Brevard (FL) | 33,579 | 8,573 | 10,043 | 85 | -49 |
| Kern (CA) | 23,078 | 7,481 | 9,402 | 80 | -44 |
| Polk (FL) | 33,634 | 8,295 | 9,532 | 87 | -42 |
Note: The cost per enrollee figures above represent the cost per enrollee with a typical risk profile—that is, these figures are, effectively, risk-adjusted.
Discussion
The role of private plans in Medicare has been the subject of much debate over the years, and that debate has intensified as the baby boom generation born after World War II has begun to become eligible for Medicare, putting pressure on the program’s fiscal viability.2 The debate has mostly focused on the assertion that private plans have the potential to be more efficient than traditional Medicare.
According to this analysis, however, MA plans that are able to keep their costs low relative to traditional Medicare are concentrated in a relatively small number of counties. In the 25 counties with the largest total difference between MA plans and traditional Medicare costs, MA plans spent a total of $5.2 billion less than traditional Medicare would have been expected to spend on the same beneficiaries, with health maintenance organizations (HMOs)—the only type of MA plan with lower nationwide per enrollee costs than traditional Medicare—accounting for all of that difference. In the rest of the country, MA plans spent $4.8 billion more than in traditional Medicare.
These data show that broad statements regarding the relative efficiency of private Medicare plans and traditional Medicare can be misleading, given the wide variation that exists in local markets and in the performance of individual MA plans.
|
About This Study This analysis is based on data on MA plan costs in 2012 posted by the Centers for Medicare and Medicaid Services (CMS) on its website.3 Additional data on MA plan enrollment and other aspects of the MA program, as well as county data on costs in traditional Medicare, also were acquired from the CMS website.4 In June of every year, each MA plan is required to submit data to CMS on the costs to the plan in the previous calendar year of providing its enrollees with the same Medicare benefits (risk-adjusted to control for beneficiaries’ health status) provided by traditional Medicare. This amount is then trended forward to the following year by the projected inflation in Medicare costs as determined by CMS. CMS uses these data, termed the plan’s “bid,” to calculate the amount of Medicare payments to the plan in the following calendar year. Each MA plan’s bid is then compared with a MA county-level benchmark payment amount set by CMS as the projected average cost of benefits in traditional Medicare in the county in the following year, as provided by the Medicare statute. Each MA plan receives a payment rate equal to: 1) the plan’s bid plus a “rebate” that is equal to a proportion of the difference between the county benchmark amount and the plan’s bid, if the plan’s bid is less than the county benchmark amount; or 2) the benchmark payment amount for the county, if the plan’s bid is not less than the benchmark amount. Although the amount of the bid submitted by each MA plan is not publicly released by CMS, the agency does post a “payment-net-of-rebate” amount for each plan in the year after the payment year (e.g., in December 2013 for calendar year 2012). The payment-net-of-rebate amount is equal to the MA plan’s costs to provide Medicare benefits for the vast majority of plans that had bids below the benchmark rate for the county in 2012.5 This payment-net-of-rebate amount for calendar year 2012 includes data for plans in the 2,933 counties where there were at least 11 MA enrollees. Because the payment-net-of-rebate amount for each MA plan is based on its retrospectively reported costs—based on actual experience—we use that information here to represent the plan’s costs. For this analysis, MA plan costs are compared with per-beneficiary spending in traditional Medicare in the same county (also risk-adjusted to represent the average Medicare beneficiary) and the difference between the two is multiplied by the number of each MA plan’s enrollees in the county to indicate the difference between MA plan costs and what traditional Medicare would be expected to spend for the same beneficiaries. The relative costs for all MA plans of each type—HMOs, local PPOs, regional PPOs, and PFFS plans—are then summed to get a total value for each county. The analysis here does not include special-needs plans and employer-sponsored plans. These county-level amounts are used to calculate the amount by which MA plan costs are greater or less than traditional Medicare spending per beneficiary for each of the four MA plan types nationwide, both as a percentage and in terms of the absolute difference in total costs. |
Notes
1 B. Biles, G. Casillas, and S. Guterman, “Variations in County-Level Costs Between Traditional Medicare and Medicare Advantage Have Implications for Premium Support,” Health Affairs, Jan. 2015 34(1):56–63.
2 See the discussion in D. Blumenthal, K. Davis, and S. Guterman, “Medicare at 50—Origins and Evolution,” New England Journal of Medicine, Jan. 29, 2015 372(5):479–86.
3 Centers for Medicare and Medicaid Services, “Plan Payment Data for 2012,” http://www.cms.gov/Medicare/Medicare-Advantage/Plan-Payment/Plan-Payment-Data-Items/2012data.html?DLPage=1&DLEntries=10&DLSort=0&DLSortDir=ascending.
4 For more information on these data, see B. Biles, G. Casillas, and S. Guterman, “Variations in County-Level Costs Between Traditional Medicare and Medicare Advantage Have Implications for Premium Support,” Health Affairs, Jan. 2015 34(1):56–63 (online appendix).
5 For more on the bidding process and how MA plan rates are determined, see B. Biles, G. Casillas, G. Arnold, and S. Guterman, The Impact of Health Reform on the Medicare Advantage Program: Realigning Payment with Performance (New York: The Commonwealth Fund, Oct. 2012).