The Affordable Care Act’s reductions to Medicare Advantage plan payments were not significantly associated with healthcare access or affordability for enrollees.
Objectives: To explore whether the Affordable Care Act (ACA)’s Medicare Advantage (MA) payment cuts were associated with changes in enrollees’ access to and affordability of healthcare relative to traditional Medicare (TM).
Study Design: Descriptive analyses of changes in access and affordability in MA relative to TM between 2009 and 2017 and between 2011 and 2017.
Methods: Respondents who reported Medicare coverage on the National Health Interview Survey were divided into MA and TM enrollees. Using multivariate regression to adjust for demographic, economic, and health status changes over time, we compared changes in healthcare access and affordability for the 2 groups between 2009 and 2017, as the ACA payment cuts were implemented. For some measures, the analysis covers 2011 to 2017.
Results: Between 2009 and 2017, MA respondents did not report statistically significant changes in healthcare access or affordability after adjusting for demographic, socioeconomic, and health status changes in the MA population. There were no statistically significant differences between changes in access and affordability for beneficiaries in MA relative to those in TM over this period.
Conclusions: Although MA payment cuts were expected to reduce the attractiveness of the MA program to both plans and enrollees, the program’s enrollment grew steadily from 2009 to 2017. Over this period, plans reduced their costs for providing Part A and Part B benefits to their enrollees, thereby preserving room for rebates. Our findings show that plans made such cost reductions without significantly affecting enrollees’ access to or affordability of care compared with TM beneficiaries.
Am J Manag Care. 2019;25(9):e261-e266Takeaway Points
Prior to the Affordable Care Act (ACA), Medicare Advantage (MA) plans were overpaid,1-3 costing the Medicare program an average of 114% of traditional Medicare (TM) spending per beneficiary nationally in 2009.4 The ACA reduced payments to MA plans through changes to the MA benchmark-and-bidding system, resulting in MA plans, on average, costing 100% of TM spending per beneficiary nationally in 2017.5 Research suggests that MA plans were able to reduce their costs and maintain extra benefits under this payment pressure.1,6 However, less is known about how plan responses to changes to MA payments under the ACA may have affected beneficiaries’ access to and affordability of care and whether MA plan cost reductions may have come at the expense of beneficiaries through narrower networks, stronger utilization management, or higher cost sharing. Using data from the 2009, 2011, and 2017 rounds of the National Health Interview Survey (NHIS), we assessed whether reductions in MA plan payments affected enrollees’ healthcare access and affordability relative to TM beneficiaries over this period.
Background on MA Program
The MA program allows Medicare beneficiaries to receive their Part A and Part B benefits through private plans. To participate in the program, private plans submit a bid to CMS equal to the expected cost, including administrative costs and profits, of providing Medicare Part A and Part B benefits to an average-risk Medicare enrollee. This bid is compared with a predetermined benchmark set by CMS. Plans bidding below the benchmark receive a portion of the difference between the benchmark and their bid as a rebate, which must be used to provide extra benefits, such as lower cost sharing, to enrollees. The availability of extra benefits at no additional cost to enrollees has been thought to be an important driver of MA enrollment.7-9
Impact of the ACA
The ACA made significant changes to the process for calculating benchmarks and rebates, generally lowering them to bring average payments to MA plans more in line with TM spending per beneficiary. Under the ACA, benchmarks are based on average TM spending per beneficiary in a county, with some variation across counties to preserve choice and reward plans with high quality ratings. Additionally, the ACA lowered the share of the difference between benchmarks and bids that plans would receive as a rebate from 75% to between 50% and 70%, depending on the plan’s quality rating. Nationally, the ACA lowered benchmarks from an average of 118% of TM spending per beneficiary in 2009 to an average of 106% in 2017.4,5 In turn, MA plans lowered their bids from 102% of TM spending per beneficiary in 2009 to 90% in 2017.4,5 This preserved space for rebates,6 which fell only from 12% of TM spending per beneficiary in 2009 to 10% in 2017.4,5
The reduction in MA bids relative to TM from 2009 to 2017 could reflect changes in MA plan structure that would negatively affect beneficiaries’ access to care and affordability of care, such as increased utilization management, higher cost sharing, or narrower networks. In addition, the rebate reductions over this period could have harmed beneficiaries through higher cost sharing, although these reductions were small on average. This study sought to explore whether plan responses to the ACA’s payment reductions to MA plans were accompanied by changes in MA beneficiaries’ access to and affordability of care relative to TM beneficiaries.
Our data are from the NHIS, which is a nationally representative household survey conducted by the US Census Bureau.7 The survey collects information from approximately 87,500 noninstitutionalized residents across 35,000 households each year. All respondents provide health insurance coverage and healthcare affordability information for all household members, but only 1 adult per household is sampled for access-to-care questions. Following the passage of the ACA, the NHIS added several questions on access to and affordability of care in 2011.
We focused primarily on data from the 2009 and 2017 rounds of the survey to allow us to best compare access and affordability pre- and post ACA payment changes. We also compared 2011 and 2017 NHIS data to allow us to use the additional questions on access to and affordability of care added to the NHIS in 2011. As the ACA’s payment cuts were first implemented in 2011 and would not affect enrollees until the 2012 plan year, we were still able to compare access and affordability pre- and post ACA. We opted to use the 2009 and 2017 surveys for our primary analysis, as the passage of the ACA froze payments at 2009 levels during early implementation, potentially affecting the MA market.
Our sample included Medicare beneficiaries of all ages who reported both Medicare Part A and Part B coverage or MA coverage at the time of the survey. We excluded any beneficiaries who also reported Medicaid coverage. In 2009, our sample included 8867 Medicare beneficiaries, with 4832 responding to the sample adult questionnaire. For 2011, our sample included 10,646 Medicare beneficiaries, with 5954 responding to the sample adult questionnaire. Finally, for 2017, our sample included 11,398 Medicare beneficiaries, with 6311 responding to the sample adult questionnaire.
We opted not to use the Medicare Current Beneficiary Survey (MCBS) for our analysis, despite its larger sample size for the Medicare population and focus on access to care in Medicare. Changes were made in the MCBS sample design between 2009 and 2015, which may lead to inaccurate estimates of changes in access and affordability over time. Additionally, the 2009 MCBS included a limited income measure, making adjustment for differences in income between MA and TM enrollees, as well as adjustment for changes in income over time, imprecise. This is a problem for accurately measuring changes in affordability of care, as income is a strong predictor of respondent perceptions of healthcare affordability. Overall, the NHIS provides more current data, better socioeconomic control variables, and more consistent sample design over time than the MCBS, making it the superior option for assessing changes in access and affordability for the Medicare population between 2009 and 2017.
Key Outcome Measures
For our primary analysis using 2009 and 2017 NHIS data, our 3 measures for access to care were (1) did not have a usual source of care at the time of the survey, (2) did not have any healthcare visit during the last 12 months, and (3) had a delay in care because of wait time (either for an appointment or at a provider’s location). These measures were drawn from the sample adult questionnaire and may indicate whether MA plans inhibited enrollees’ access to care in order to reduce costs following payment cuts. Our affordability measures were (1) had a delay in care for cost reasons and (2) did not get needed healthcare for cost reasons, both in the past 12 months. These questions were asked for all family members, not just sample adults, leading to a larger sample size for our affordability measures than for our access-to-care measures. These measures may indicate whether MA plans increased cost sharing in response to payment cuts.
For our additional analyses using the 2011 and 2017 NHIS, our 3 access-to-care measures were (1) had trouble finding a general doctor with availability, (2) were told that they would not be accepted as a new patient, and (3) were told that their health insurance was not accepted, all during the last 12 months. Changes in these measures may indicate that MA plans narrowed networks or otherwise more aggressively managed physician access in response to payment cuts. Our 2 affordability measures were (1) did not get needed specialist care during the last 12 months for cost reasons and (2) had problems paying medical bills during the last 12 months, which may indicate that MA plans increased cost sharing in response to payment cuts.
We conducted all analyses using Stata 15 (StataCorp; College Station, Texas) with the applicable survey weights to account for each of the 2 sample populations. We first conducted descriptive analyses to determine the demographic, socioeconomic, geographic, and health characteristics of MA and TM beneficiaries in each survey year (Table 1 and eAppendix Table 1 [eAppendix available at ajmc.com]), as well as the raw share of beneficiaries experiencing each access and affordability difficulty in each year. Because the characteristics of enrollees in MA and TM differed in each year and changed over the study period, we used ordinary least squares regression to adjust estimates of within-year differences and changes over time for each outcome variable between MA and TM. Our regression model adjusted for observable characteristics including age, gender, race and ethnicity, education, family income relative to the federal poverty level (FPL), marital status, geographic region, and self-reported health status.
Between 2009 and 2017, both the MA and TM populations changed as baby boomers aged into the Medicare program, with higher shares reporting they were aged between 65 and 74 years, were of very good or excellent health status, had graduated college, and had family incomes at or above 300% FPL, with a greater decrease in the share at or below 150% FPL for MA (Table 1). The changes from 2011 to 2017 for age, health status, educational attainment, and income were similar to those from 2009 to 2017, with slightly lower magnitudes and no significant difference between MA and TM in income changes over time (eAppendix Table 1). The demographic, socioeconomic, and health characteristics of beneficiaries in MA and TM in 2017 were fairly similar, although MA enrollees were more likely to be Hispanic, be nonwhite, and have family incomes below 300% FPL (Table 1).
Changes in access to care were comparable for MA and TM beneficiaries. MA enrollees were more likely than TM beneficiaries to have interacted with the healthcare system in both 2009 and 2017, with little change between the 2 years. In both years, MA beneficiaries were less likely than TM beneficiaries to report not having a usual source of care and not having a visit to a healthcare professional during the past 12 months after adjusting for differences in demographic, socioeconomic, and health characteristics (Table 2). The changes over time between MA and TM beneficiaries were minimal for these 2 measures, with magnitudes less than 0.5 percentage points, and were not statistically significant. Over time, TM beneficiaries did report increased rates of delay in care because of wait times (change of 2.3 percentage points; P <.10), but the changes in delay in care for wait times in MA and TM were not statistically different.
The higher level of connection to the healthcare system for MA enrollees did not seem to result in fewer problems obtaining care, however. MA enrollees reported rates similar to TM beneficiaries for having trouble finding a general doctor, being told their health insurance was not accepted, and being told they would not be accepted as a new patient in both 2011 and 2017, even after adjustment for differences in demographic, socioeconomic, and health characteristics (Table 3). Between 2011 and 2017, there were no statistically significant changes in these measures for either MA or TM. There were no statistically significant differences between access-to-care changes in MA compared with TM over either the 2009 to 2017 period or the 2011 to 2017 period after regression adjustment.
There was also no evidence that the ACA’s payment cuts were associated with declines in MA beneficiaries’ affordability of care. MA enrollees were more likely than TM beneficiaries to report delays in care for cost reasons in both 2009 (7.4% vs 5.6%; P <.05) and 2017 (5.6% vs 4.6%; P <.10), and these differences remained statistically significant after regression adjustment (Table 2). However, changes in delay in care for cost reasons were not statistically significant in either MA or TM, and there was no statistically significant difference between the MA change and the TM change over the study period. Respondents reported similar changes in not obtaining specialist care due to cost reasons: Beneficiaries in MA had higher rates than those in TM in both 2011 (3.0% vs 2.4%; P <.10) and 2017 (4.2% vs 2.6%; P <.10), but there was no statistically significant change in either MA or TM over the study period (Table 3). Other affordability measures, including not getting needed care for cost reasons and problems paying medical bills, were not statistically different between MA and TM in any year and did not change significantly over time (Tables 2 and 3). There were no statistically significant differences between MA changes over time and TM changes over time in any affordability measures.
Overall, there were no statistically significant changes in access to or affordability of care for the MA population relative to TM from 2009/2011 to 2017 (Tables 2 and 3). Sensitivity analyses focusing exclusively on changes from 2011 to 2017 also showed no statistically significant differences between MA and TM after regression adjustment (eAppendix Table 2).
The ACA’s cuts to MA payments were expected to reduce the attractiveness of MA to both plans and beneficiaries, primarily through reductions in rebates, thereby decreasing enrollment in the MA plan program.7-9 Instead, MA enrollment steadily increased as the ACA was implemented. Prior research indicates that MA plans significantly reduced costs (as reflected by their bids) in response to ACA payment pressure, lowering their bids from 102% of TM to 90% of TM.4,5 This preserved room for rebates,6 which may partially explain the program’s continued attractiveness. However, such bid reductions could have been achieved through significant narrowing of networks, increased utilization management, or increases in cost sharing, which could harm beneficiary access to and affordability of care even as rebates remained stable.
Our findings show that the ACA’s payment reductions and plans’ resulting reductions in costs were not associated with declines in access to or affordability of care in MA relative to TM. Between 2009 and 2017, as the ACA’s changes to MA payments were fully phased in, we found no statistically significant changes in healthcare access and affordability among MA beneficiaries after adjusting for demographic, socioeconomic, region, and health status characteristics. More importantly, we found no statistically significant differences between changes in access and affordability in MA compared with TM over this period. These results suggest that MA plans could reduce costs under the ACA without negatively affecting the average beneficiary’s healthcare access and affordability.
Bid reductions were only one way that plans maintained revenue and rebates, however, and other revenue sources available to plans depend partially on maintaining a high level of beneficiary access to care. For example, MA plans also preserved revenue and rebates by increasing quality scores,10 which increase both benchmarks and rebate percentages. Quality scores take into account beneficiary experiences accessing care, as well as beneficiaries’ receipt of preventive services, creating an incentive for plans to maintain high beneficiary access to care. In addition, MA may have preserved revenue through the risk adjustment system by more comprehensive coding of enrollees’ diagnoses,11 either through health assessments or when seen by their physicians. Overall, continued MA enrollment growth may reflect that premiums, extra benefits,1,10 and access and affordability have remained relatively stable despite overall payment reductions, making those payment changes largely invisible to beneficiaries.
Our analysis has several limitations. First, our regression adjustments cannot fully control for changes in the MA and TM beneficiary populations over the study period or for changes in local healthcare markets that could affect access and affordability. Second, MA payment cuts were not uniform across the country, but we are unable to isolate those geographic areas or plans most exposed to payment cuts using the NHIS. We are also unable to observe MA and TM beneficiaries’ county of residence, so we cannot verify whether MA and TM respondents are exposed to similar healthcare markets. Third, we focus on the average beneficiary, but it is possible that beneficiaries with more significant healthcare needs have fared differently in MA compared with TM over this period. Fourth, the NHIS captures health insurance coverage at the time of the survey, so we are unable to isolate beneficiaries who changed coverage during the year. These beneficiaries may be particularly likely to experience access and affordability disruptions. Finally, we used survey results from after the ACA was passed to measure pre—payment cut levels of access and affordability in our second model, although MA plans may have already started to implement changes to prepare for the payment reductions.
The ACA’s reductions to MA plan payments were not associated with declines in healthcare access or affordability for MA enrollees. In fact, as payment cuts were phased in, MA plans reduced costs without diminishing healthcare access or affordability for enrollees relative to TM beneficiaries. Despite contrary projections, MA plans experienced steady, robust enrollment growth from 2009 to 2017, implying that MA plans became more attractive to beneficiaries during a period of increasing payment pressure.Author Affiliations: Urban Institute (LS, JA, SZ), Washington, DC.
Source of Funding: National Institute for Health Care Reform.
Author Disclosures: Ms Skopec received a grant for unrelated work from Anthem Inc. The remaining authors report no relationship or financial interest with any entity that would pose a conflict of interest with the subject matter of this article.
Authorship Information: Concept and design (LS, JA, SZ); acquisition of data (JA); analysis and interpretation of data (LS, JA, SZ); drafting of the manuscript (LS, JA); critical revision of the manuscript for important intellectual content (LS, JA, SZ); statistical analysis (LS, JA, SZ); obtaining funding (LS, SZ); and supervision (LS, SZ).
Address Correspondence to: Laura Skopec, MS, Urban Institute, 500 L’Enfant Plaza SW, Washington, DC 20024. Email: email@example.com.REFERENCES
1. Song Z, Pelech DM. Payments to Medicare Advantage plans and plan generosity before and after the Affordable Care Act. Presented at: 7th Conference of the American Society of Health Economists; June 11, 2018; Atlanta, GA. ashecon.confex.com/ashecon/2018/webprogram/Paper6269.html. Accessed October 2, 2018.
2. Song Z, Landrum MB, Chernew ME. Competitive bidding in Medicare Advantage: effect of benchmark changes on plan bids. J Health Econ. 2013;32(6):1301-1312. doi: 10.1016/j.jhealeco.2013.09.004.
3. McGuire TG, Newhouse JP, Sinaiko AD. An economic history of Medicare Part C [erratum in Milbank Q. 2013;91(1):210]. Milbank Q. 2011;89(2):289-332. doi: 10.1111/j.1468-0009.2011.00629.x.
4. The Medicare Advantage program. In: Medicare Payment Advisory Commission. Report to the Congress: Medicare Payment Policy. Washington, DC: Medicare Payment Advisory Commission; 2009. medpac.gov/docs/default-source/reports/mar09_ch03.pdf. Accessed October 2, 2018.
5. Status report on the Medicare Advantage program. In: Medicare Payment Advisory Commission. Report to the Congress: Medicare Payment Policy. Washington, DC: Medicare Payment Advisory Commission; 2017. medpac.gov/docs/default-source/reports/mar17_medpac_ch13.pdf. Accessed October 2, 2018.
6. Guterman S, Skopec L, Zuckerman S. Do Medicare Advantage plans respond to payment changes? a look at the data from 2009 to 2014. The Commonwealth Fund website. commonwealthfund.org/publications/issue-briefs/2018/mar/do-medicare-advantage-plans-respond-to-payment-changes. Published March 14, 2018. Accessed October 2, 2018.
7. Comparison of projected enrollment in Medicare Advantage plans and subsidies for extra benefits not covered by Medicare under current law and under reconciliation legislation combined with H.R. 3590 as passed by the Senate. Congressional Budget Office website. cbo.gov/sites/default/files/111th-congress-2009-2010/costestimate/macomparisons.pdf. Published March 19, 2010. Accessed October 2, 2018.
8. Foster RS. Estimated financial effects of the “Patient Protection and Affordable Care Act,” as amended. CMS website. cms.gov/Research-Statistics-Data-and-Systems/Research/ActuarialStudies/Downloads/PPACA_2010-04-22.pdf. Published April 22, 2010. Accessed October 2, 2018.
9. Nicholas LH. Will Medicare Advantage payment reforms impact plan rebates and enrollment? Am J Manag Care. 2014;20(11):917-924.
10. Jacobson G, Damico A, Neuman T, Gold M. Medicare Advantage 2017 spotlight: enrollment market update. Kaiser Family Foundation website. kff.org/medicare/issue-brief/medicare-advantage-2017-spotlight-enrollment-market-update. Published June 6, 2017. Accessed October 1, 2018.
11. The Medicare Advantage program: status report. In: Medicare Payment Advisory Commission. Report to the Congress: Medicare Payment Policy. Washington, DC: Medicare Payment Advisory Commission; 2018. medpac.gov/docs/default-source/reports/mar18_medpac_ch13_sec.pdf. Accessed October 2, 2018.