Potential Inequities in Access to In-Person SHIP Counseling Services

ABSTRACT

Objectives: Counseling and education on Medicare coverage options are available through the federal State Health Insurance Assistance Program (SHIP), but little is known about the population that SHIP reaches.

Study Design: Cross-sectional study.

Methods: Using a novel data source on SHIP counseling site locations, we characterized the availability of in-person SHIP counseling by zip code tabulation area (ZCTA) and used linear regression and t tests to evaluate whether SHIP counseling sites are disproportionately located in higher-income communities.

Results: Our sample included 1511 SHIP counseling sites. More than half (63%) of the localities in our sample have a SHIP site within the ZCTA or county. Twenty-four percent do not have a SHIP site within the county but have one in an adjacent county. The remaining 13% do not have a nearby SHIP site. There is a disproportionate number of individuals eligible for Medicare in localities without a SHIP site. Moreover, the population living in areas without in-person SHIP sites is more likely to have low income and fewer years of education than the population living in areas with a SHIP site.

Conclusions: These results suggest that there are areas where in-person SHIP service expansion or other additional navigation support may be warranted.

Am J Manag Care. 2024;30(2):e46-e51. https://doi.org/10.37765/ajmc.2024.89500

Takeaway Points

More than half of localities (unique zip code tabulation area–county combinations) have a State Health Insurance Assistance Program (SHIP) site within their county.

• There is a disproportionate number of individuals eligible for Medicare in localities without a SHIP site.
• The population living in areas without in-person SHIP sites is more likely to have low income and fewer years of education than the population living in areas with a SHIP site.
• Our results suggest that there are areas where in-person SHIP service expansion or other additional navigation support may be warranted.

Choices for Medicare coverage are numerous and complex. The growth in Medicare Advantage (MA) enrollment and coverage options amplifies the need to target plan selection counseling services to beneficiaries most in need of them.^1,2 Counseling and education on coverage options for Medicare beneficiaries are available through the federal State Health Insurance Assistance Program (SHIP), but little is known about the population that SHIP reaches. Recent estimates suggest that less than 10% of Medicare beneficiaries use SHIP services, even though 30% of beneficiaries shop for plans each year.^3-5

Overseen by the Administration for Community Living (ACL) within HHS, SHIP program funds are granted to states, which in turn often subgrant funds to local organizations, including counties, Area Agencies on Aging (AAAs), health systems, and community-based organizations.⁶ Within these local organizations, a mix of paid and volunteer counselors provide information on coverage choices, eligibility, and costs to Medicare beneficiaries. Counselors in all states are required to be trained in details of Medicare benefits and coverage, rights to appeal, integrated care programs for individuals dually eligible for Medicare and Medicaid, and programs that facilitate independent living in the community.⁷

The ACL produces annual reports on the SHIP program, but they are at a high level, focusing on state-level metrics of number of contacts with and outreach to Medicare beneficiaries.³ For instance, state SHIP programs are scored on the percentage of “hard-to-reach” individuals that they contact, defined as those who are younger than 65 years, have low income, live in rural areas, or who are not native English speakers. In 2018, the most recent year for which performance data are available, 16 states had a good or excellent rating on this self-reported metric, defined as reaching 6.73% or more of the hard-to-reach population.³ Five states reached less than 1.58% of this population. The low percentages of hard-to-reach individuals who are contacted, even in the states with good or excellent ratings, highlight a potential gap in access to information among individuals who are dually eligible for Medicare and Medicaid. No national data are available on whether SHIP sites are located in areas with a high concentration of Medicare beneficiaries or whether they equitably serve low-income and high-income beneficiaries within a state.

Understanding SHIP service availability for low-income beneficiaries is critical to identifying potential unmet needs for counseling in complex coverage options, including dual eligibility for Medicaid, partial dual eligibility, and the Medicare Low-Income Subsidy. SHIPs are expected to provide this counseling—federal funding for the SHIP program is based in part on the percentage of low-income beneficiaries who reside in each state through the Medicare Improvements for Patients and Providers Act.⁸

Especially for beneficiaries living in low-income neighborhoods, access to SHIP services may be affected by the availability of volunteers and transportation.^9-11 Access also may be affected by the availability of services in languages other than English.^12,13 Although telephone counseling can be accessed nationally, access to in-person counseling and education is of special concern, particularly for adults with hearing loss or limited English proficiency, or who rely on a caregiver or companion to understand information about health care.¹⁴

Using a novel data source on locations of SHIP counseling sites, we characterized the availability of in-person SHIP counseling by zip code tabulation area (ZCTA) and evaluated whether SHIP counseling sites are disproportionately located in higher-income communities. We also assessed the degree to which potential language barriers, population education, and transportation availability vary by SHIP availability. Our goal was to identify areas where beneficiaries may have unmet needs for SHIP in-person services and where SHIP service expansion or other additional navigation support may be warranted.

METHODS

Using state directories of SHIP counseling sites and publicly available data on ZCTA and county-level population characteristics, we examined whether socioeconomic conditions differ across localities that do and do not have SHIP counseling sites. We used ZCTAs and counties as proxy measures of SHIP sites’ likely in-person service areas. We assumed that individuals who reside in a ZCTA or county with a SHIP site were more likely to seek services from that site than individuals who live elsewhere. However, beneficiaries may cross borders to seek information about enrollment choices. To understand characteristics of areas in which in-person SHIP counseling was more or less likely to be available, we compared characteristics across 4 groups: ZCTA included a SHIP; ZCTA did not include a SHIP but was located in a county that included a SHIP; ZCTA was located in a county that did not include a SHIP, but an adjacent county included a SHIP; and ZCTA was located in a county that did not include a SHIP and no adjacent counties included a SHIP.

Data Sources

Data on counseling site locations were obtained from a review of states’ web directories of SHIP service sites. We included listed community partners, providers, office locations, and SHIP sites from state directories as well as SHIP grant subrecipients listed in ACL records from April 2020 to March 2021.⁶ For some states, only the addresses of AAA locations were available. Because AAAs generally work with other local organizations to offer SHIP counseling, we characterized states that only listed AAA locations as not having statewide information on service locations available. For this analysis, we focused on 27 states for which statewide information on service locations was available online. More details about the SHIP directory are available in the eAppendix (available at ajmc.com). Data on socioeconomic characteristics of ZCTAs were obtained from the 2018 Agency for Healthcare Research and Quality’s Social Determinants of Health (AHRQ SDOH) file, which includes data from the American Community Survey (ACS), Area Health Resources Files (AHRF), and Medicare Advantage penetration files.¹⁵ Variables from the SDOH/ACS file reflect 5-year estimates from 2013 to 2018. County-level data on Medicare plan availability are from the 2022 CMS landscape files.¹⁶

Sample

Our sample included ZCTAs established as of 2010 that had a corresponding county Federal Information Processing Standards (FIPS) code and that belonged to the 27 states for which SHIP service location data were available online. (Newer ZCTAs established in 2020 are not included in the 2018 AHRQ SDOH file.) We dropped 111 ZCTAs with fewer than 10 residents. ZCTAs could be wholly contained within 1 county or sit across multiple counties.¹⁷ If a ZCTA spanned 2 counties, it was represented in our data in 2 unique ZCTA-FIPS combinations. After restricting our data set to the 27 states for which SHIP location data were available online, our sample included 24,171 unique ZCTA-FIPS combinations (hereafter referred to as localities). These 27 states include 50% of Medicare beneficiaries and 50% of dually eligible individuals.^18,19

Variables

For each locality, we recorded whether a SHIP site was present in the ZCTA, the county, or an adjacent county. To characterize the size of the population that SHIP sites might serve in each locality, we included the total weighted population of the ZCTA from the ACS, the percentage of the population 65 years or older in the ZCTA, and the number of individuals eligible for Medicare per 1000 residents in the county.

To characterize potential beneficiary income level, we obtained ZCTA-level data on the percentage of households with incomes less than 138% of the federal poverty limit (FPL), between 138% and 199% of FPL, between 200% and 399% of FPL, and 400% or greater of FPL from the SDOH/ACS file. We also included the percentage of the population with Medicaid or other means-tested insurance coverage and the median household income.

We recorded other variables from the SDOH/ACS that may be associated with reduced access to in-person SHIP counseling services: language barriers (percentage of population 5 years or older who did not speak English well or at all), education (percentage of population 25 years or older with a high school diploma as the highest level of education, percentage with at least some college), and transportation availability (percentage of workers 16 years or older who used public transportation, percentage of occupied housing units without a vehicle). We also included race (percentage of population reporting race as Black, White, Asian, American Indian or Alaskan Native, Native Hawaiian or Other Pacific Islander, multiracial, or other) and ethnicity (percentage of population reporting Hispanic ethnicity). From the SDOH/AHRF file, we recorded whether localities were in a county designated as a primary care health professional shortage area (HPSA). We also obtained the rural-urban continuum code (RUCC) and classified localities as belonging to a metropolitan (RUCC = 1-3) or nonmetropolitan (RUCC = 4-9) county.

As a proxy for the complexity of information a SHIP counselor would be required to know, we included measures of the number of Medicare plans (Medicare Advantage, Prescription Drug Plans, Special Needs Plans, and Medicare-Medicaid [dually eligible] plans) available from the CMS landscape files.

Analyses

We conducted our analyses at the ZCTA-FIPS (locality) level. We used linear regression to characterize the degree to which the 4 categories of localities (ZCTA includes SHIP, SHIP in same county, SHIP in adjacent county, no SHIP in adjacent county) differ on income, access-related, and demographic variables. For clarity of presentation, we highlight t test and χ² comparisons between 2 categories of localities: those with a SHIP in the ZCTA or the same county (SHIP present) and those without a SHIP in the county or adjacent county (SHIP absent). Results using the more detailed categories were substantively similar and are available in the eAppendix. Observations with missing values of a variable (0%-6.0% of observations) were dropped from analyses relying on that variable. We repeated these analyses in the localities with greater than the median value of the percentage of the population 65 years and older. We assumed characteristic-dependent null hypotheses (eg, localities with and without SHIPs would not differ significantly on population education status), not a universal null hypothesis (that localities would be equal on all characteristics). As a result, we did not apply a Bonferroni correction for multiple comparisons.²⁰ This study was deemed not human subjects research by the Boston University School of Public Health institutional review board.

RESULTS

Our sample included 1511 SHIP counseling sites in 24,171 localities. More than half (62.9%) of the localities in our sample had a SHIP counseling site within the ZCTA (1267 [5.2%]) or within the county (13,932 [57.6%]). Some ZCTAs included multiple counseling sites. Approximately one-fourth (5780 [23.9%]) did not have a SHIP counseling site within the county but had one in an adjacent county. The remaining 3192 (13.2%) did not have a nearby SHIP counseling site. The distribution of SHIP sites by county is provided in Figure 1.

Distribution of Potential SHIP Clients in Localities With and Without SHIP Sites

The percentage of the population 65 years or older was greater in localities without a nearby SHIP site (mean [SD], 20.5% [10.4%]) than in localities with a SHIP site (18.5% [8.6%]). Similarly, the number of individuals eligible for Medicare per 1000 county residents was greater in localities without a SHIP site (mean [SD], 221 [60]) than in localities with a SHIP (209 [43]). The percentage of the population reporting race as Asian, other, or White was greater in localities with a SHIP site than in areas without a SHIP site, and the percentage of the population reporting race as American Indian or Alaskan Native or Black was greater in localities without a SHIP site than in areas with a SHIP site.

Income in Localities With and Without SHIP Sites

The percentage of the population with income less than 200% of the FPL was higher in localities without SHIP sites than those with a SHIP site (Figure 2). The mean (SD) percentage of the population with income less than 138% of the FPL was 23.9% (13.2%) in localities without a SHIP site and 19.4% (11.9%) in localities with a SHIP site. Similarly, the median household income was higher in localities with a SHIP site than localities without a SHIP site (mean [SD], $61,729 [$24,508] vs $51,260 [$16,501], respectively) (Table 1). We did not find evidence of a relationship between Medicaid enrollment and SHIP availability.

Access-Related Characteristics in Localities With and Without SHIP Sites

Localities without SHIP sites had significantly lower percentages of individuals who do not speak English at all or well than localities with SHIP sites (mean [SD], 1.2% [2.5%] vs 1.4% [2.9%], respectively). The percentage of adults 25 years or older with less than a high school education was significantly greater in localities without SHIP sites than localities with a SHIP site (mean [SD], 13.6 [8.3%] vs 10.8% [7.8%], respectively). Public transportation use was higher in localities with a SHIP site than those without a SHIP site (mean [SD], 2.1% [6.2%] vs 0.5% [1.8%]). Localities with SHIP sites were less likely to be in counties classified as rural than localities without nearby SHIP sites.

Complexity of Coverage Options in Localities With and Without SHIP Sites

Localities with a SHIP site had a greater number of available Medicare plans than areas without a nearby SHIP site (mean [SD], 62.8 [16.9] vs 56.9 [27.3], respectively). Variation in plan count came primarily from the number of MA plans offered in a county (data not shown).

Localities With Relatively High Percentages of Older Residents

In analyses restricted to localities above the median value of the percentage of the population that was 65 years or older, most of the results did not meaningfully differ from our primary analyses (results not shown). The following results changed. The percentage of the population that does not speak English at all or well and the percentage of the population that is Hispanic were slightly higher in localities without a SHIP site than with a SHIP site (do not speak English at all or well: 0.86% vs 0.75%; P = .015; Hispanic: 4.39% vs 3.60%; P < .001), and the percentage of occupied housing units without a vehicle and the percentage with Medicaid or means-tested coverage were higher in localities with a SHIP site vs without a SHIP site (without a vehicle: 5.14% vs 4.60%; P = .001; Medicaid or means-tested coverage: 15.08% vs 13.54%; P < .001). The percentage of the population in the race category “other” and the indicator for HPSA were no longer significantly associated with SHIP availability in this subpopulation analysis focused on higher density of older adults.

DISCUSSION

We created the first multistate database of likely SHIP in-person counseling sites and linked it to national data on Medicare plan availability and population socioeconomic characteristics. With this unique data set, we observed that more than half of localities had a SHIP site within the county, and localities with SHIP sites had the most complex Medicare coverage options. However, there was a disproportionate number of individuals eligible for Medicare in localities without a SHIP site. These results suggest that there are areas where in-person SHIP service expansion or other additional navigation support may be warranted.

Moreover, our results highlight potential inequities in access to in-person SHIP counseling services. The population living in areas without in-person SHIP sites had lower income and fewer years of education than the population living in areas with a SHIP site. The higher proportion of individuals with incomes less than 138% of the FPL and between 138% and 200% of the FPL in areas without a SHIP suggests there may be dually eligible individuals, as well as individuals eligible for Medicare Low-Income Subsidies, at risk of not accessing in-person SHIP services. Following a recent evaluation of California’s SHIP program, one of the key recommendations was to identify strategies to support dually eligible individuals.¹³ As the numbers of dually eligible individuals and integrated care plan options increase, so will the need to help this population navigate the complex coverage landscape.^21-24 Low-income individuals who are 65 years or older may be eligible for low-income subsidies or integrated Medicare-Medicaid plans, including MA dual-eligible special needs plans. Others have incomes just above the FPL and are eligible for but not enrolled in the Medicare Savings Programs that reduce Medicare Part A and Part B cost sharing. Navigation support may be especially important for these individuals. Moreover, dually eligible individuals are permitted to change MA plans quarterly—they have multiple opportunities to switch coverage (and multiple opportunities for coverage questions to arise).

Individuals often do not understand insurance concepts such as coinsurance and co-pays and make poor choices among available coverage options, selecting higher-cost plans than necessary.^25,26 Individuals of lower socioeconomic status tend to have lower health insurance literacy than individuals with higher socioeconomic status, which influences their coverage choices.^21,27 Choosing the best option is more difficult as the number of choices increases.²⁸ However, both additional information on coverage options and regular review of coverage options are associated with better plan choices.²⁷ SHIP counselors can provide coverage information as well as counseling on the benefits of a regular review.

Proximity to in-person SHIP services is only 1 metric of potential access to SHIP counseling services.²⁹ At the population level, neither language barriers nor overall transportation availability appears to be systematically associated with SHIP site availability. However, the degree to which these factors affect access to SHIP services within areas containing a SHIP site remains to be examined on a national scale. California’s SHIP evaluation (the only state-level evaluation of which we are aware) highlighted the need to maintain an adequate supply of bilingual volunteers.¹³ An exploration of drivers of access to SHIP services in areas containing a SHIP site may identify additional opportunities to improve access.

Efforts to monitor and expand access to SHIP services will be most useful if they are accompanied by efforts to ensure that SHIP counseling services are of high quality. To our knowledge, the accuracy and completeness of information provided by SHIP counselors have not yet been systematically examined. Accurate and complete information from SHIP counselors may help combat low understanding of health insurance terminology and plan choices and instances of misinformation³⁰ from advertisers about available coverage options.

Limitations

Our results should be interpreted in light of some limitations. This analysis provides information on associations only. Our goal was to identify characteristics of areas that do and do not have a SHIP site; a causal analysis of factors that lead a SHIP site to be established in a given locality is beyond the scope of this paper, given the endogenous nature of state interest in the program and available local organizational capacity to support its delivery. In addition, ZCTAs and counties are imperfect measures of service area and vary in size—a SHIP site could be located in the same county as an individual but still require a long travel distance for in-person services. Our data do not allow us to quantify the average distance traveled to SHIP counseling sites. As noted above, we include measures of SHIP availability in adjacent counties to account for spillover of SHIP services to neighboring areas. In addition, we cannot always determine whether our directory reflects true service locations or business offices. In some states, beneficiaries must contact a phone number or contractor before being linked to an in-person counselor. We do not have national data on the location of in-person volunteers or other organizations reachable only via a third party. However, to our knowledge, this is the first systematic multistate characterization of SHIP sites and examination of potential SHIP clients. ACL itself does not maintain a list of physical SHIP locations. By characterizing areas with and without SHIPs, we can identify areas that may benefit from additional SHIP services. The fact that online information about service location could not be located for nearly half of the states also highlights opportunities to improve availability of SHIP service information.

CONCLUSIONS

Areas without in-person SHIP sites have a disproportionate share of low-income individuals and Medicare-eligible individuals. Further expansion of in-person SHIP sites may be a way to improve availability of navigation services to dually eligible and other vulnerable individuals.

Author Affiliations: VA Boston Healthcare System (MMG, ABF), Boston, MA; Boston University School of Public Health (MMG, AD, DB, PRS, ABF), Boston, MA; National Association of Dental Plans (MA), Dallas, TX; Harvard T.H. Chan School of Public Health (ABF), Boston, MA.

Source of Funding: This study was funded by The Commonwealth Fund (#20223884) and Arnold Ventures (#21-06541).

Author Disclosures: Dr Garrido reports receiving grant funding from The Commonwealth Fund and Arnold Ventures. Mr Adelberg was the federal division director responsible for the SHIP program while at CMS more than a decade ago and was previously a public policy consultant at Faegre Drinker, which received fees from Boston University to fund advisory services. Dr Shafer reports grant funding from The Commonwealth Fund, Renova Health, Veterans Health Administration, Robert Wood Johnson Foundation, the Kate B. Reynolds Charitable Trust, and Starbucks Coffee Company; consulting fees from Patient Funding Associates and the Wesleyan Media Project; and honoraria from Johns Hopkins University and the Oklahoma City-County Health Department. 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 (MMG, MA, PRS); acquisition of data (MMG, AD, DB); analysis and interpretation of data (MMG, AD, MA, DB, PRS, AF); drafting of the manuscript (MMG); critical revision of the manuscript for important intellectual content (MMG, MA, DB, PRS, AF); statistical analysis (MMG, AD); obtaining funding (MMG, MA, AF); and supervision (MMG).

Address Correspondence to: Melissa M. Garrido, PhD, Boston University School of Public Health, Talbot 2W, 715 Albany St, Boston, MA 02118. Email: garrido@bu.edu.

REFERENCES

1. Neuman P, Jacobson GA. Medicare Advantage checkup. N Engl J Med. 2018;379(22):2163-2172. doi:10.1056/NEJMhpr1804089

2. Meyers DJ, Mor V, Rahman M, Trivedi AN. Growth in Medicare Advantage greatest among Black and Hispanic enrollees. Health Aff (Millwood). 2021;40(6):945-950. doi:10.1377/hlthaff.2021.00118

3. Administration for Community Living. State Health Insurance Assistance Program (SHIP) Report to Congress GY 2018. Administration for Community Living; 2018. Accessed June 17, 2023. https://acl.gov/sites/default/files/about-acl/2021-07/ACL_2018%20SHIP%20Report%20To%20Congress.pdf

4. CMS fast facts. CMS. December 2021. Accessed January 5, 2024. https://data.cms.gov/fact-sheet/cms-fast-facts

5. Koma W, Freed M, Cubanski J, Neuman T. Seven in ten Medicare beneficiaries did not compare plans during past open enrollment period. Kaiser Family Foundation. October 13, 2021. Accessed June 17, 2023. https://web.archive.org/web/20211016152127/https://www.kff.org/medicare/issue-brief/seven-in-ten-medicare-beneficiaries-did-not-compare-plans-during-past-open-enrollment-period/

6. National Annual SHIP Subrecipient Report GY 2020. Administration for Community Living. 2021. Accessed June 17, 2023. https://acl.gov/sites/default/files/programs/2021-10/
National%20Annual%20SHIP%20Subrecipient%20Report%20GY%202020.xlsx

7. Administration for Community. 2020 State Health Insurance Assistance Program (SHIP) base grant HHS-2020-ACL-CIP-SAPG-0363. Grants.gov. August 30, 2019. Accessed January 5, 2024. https://www.grants.gov/search-results-detail/319734

8. State Health Insurance Assistance Program (SHIP). Every CRS Report. Updated October 23, 2023. Accessed January 5, 2024. https://www.everycrsreport.com/reports/IF10623.html

9. Benenson J, Stagg A. An asset-based approach to volunteering: exploring benefits for low-income volunteers. Nonprofit Volunt Sect Q. 2016;45(suppl 1):131S-149S. doi:10.1177/0899764015604739

10. Syed ST, Gerber BS, Sharp LK. Traveling towards disease: transportation barriers to health care access. J Community Health. 2013;38(5):976-993. doi:10.1007/s10900-013-9681-1

11. Phillips MA, Rivera MD, Shoemaker JA, Minyard K. Georgia’s Utilization Minigrant Program: promoting Medicaid/CHIP outreach. J Health Care Poor Underserved. 2010;21(4):1282-1291. doi:10.1353/hpu.2010.0914

12. Mathieson KM, Kronenfeld JJ. Barriers to enrollment and successful outreach strategies in CHIP: reflections on the Arizona experience. J Health Care Poor Underserved. 2003;14(4):465-477. doi:10.1353/hpu.2010.0684

13. Janoski BE. Modernizing California’s Health Insurance Counseling & Advocacy Program (HICAP): Strategy Recommendations. ATI Advisory; 2021. Accessed June 17, 2023. https://atiadvisory.com/resources/wp-content/uploads/2022/11/Modernizing-Californias-HICAP-Strategy-Recommendations.pdf

14. Ladin K, Porteny T, Perugini J, et al. Perceptions of telehealth vs in-person visits among older adults with advanced kidney disease, care partners, and clinicians. JAMA Netw Open. 2021;4(12):e2137193. doi:10.1001/jamanetworkopen.2021.37193

15. Social determinants of health database. Agency for Healthcare Research and Quality. December 2020. Accessed June 17, 2023. https://www.ahrq.gov/sdoh/data-analytics/sdoh-data.html

16. PDP state/county penetration 2022 01. CMS. January 2022. Accessed June 17, 2023. https://www.cms.gov/research-statistics-data-and-systemsstatistics-trends-and-reportsmcradvpartdenroldatapdp-state/pdp-statecounty-penetration-2022-01

17. Explanation of the 2010 ZIP code tabulation area (ZCTA) to county relationship file. US Census Bureau. Accessed November 30, 2022. https://www2.census.gov/geo/pdfs/maps-data/data/rel/explanation_zcta_county_rel_10.pdf

18. Number of dual-eligible individuals, based on Medicare claims data. Kaiser Family Foundation. Accessed June 17, 2023. https://www.kff.org/other/state-indicator/number-of-medicare-medicaid-enrollees-based-on-medicare-claims-data/?currentTimeframe=0&sortModel=%7B%22colId%22:%22Location%22,%22sort%22:%22asc%22%7D

19. Total number of Medicare beneficiaries by type of coverage. Kaiser Family Foundation. Accessed June 17, 2023. https://www.kff.org/medicare/state-indicator/total-medicare-beneficiaries/?currentTimeframe=0&sortModel=%7B%22colId%22:%22Location%22,%22sort%22:%22asc%22%7D

20. Schulz KF, Grimes DA. Multiplicity in randomised trials II: subgroup and interim analyses. Lancet. 2005;365(9471):1657-1661. doi:10.1016/S0140-6736(05)66516-6

21. McCormack L, Bann C, Uhrig J, Berkman N, Rudd R. Health insurance literacy of older adults. J Consum Aff. 2009;43(2):223-248. doi:10.1111/j.1745-6606.2009.01138.x

22. Medicare Payment Advisory Commission; Medicaid and CHIP Payment and Access Commission. Data Book: Beneficiaries Dually Eligible for Medicare and Medicaid. Medicaid and CHIP Payment and Access Commission; 2022. Accessed October 19, 2022. https://www.macpac.gov/wp-content/uploads/2022/02/Beneficiaries-Dually-Eligible-for-Medicare-and-Medicaid-February-2022.pdf

23. Grabowski DC. Special Needs Plans and the coordination of benefits and services for dual eligibles. Health Aff (Millwood). 2009;28(1):136-146. doi:10.1377/hlthaff.28.1.136

24. Roberts ET, Mellor JM. Differences in care between Special Needs Plans and other Medicare coverage for dual eligibles. Health Aff (Millwood). 2022;41(9):1238-1247. doi:10.1377/hlthaff.2022.00463

25. Loewenstein G, Friedman JY, McGill B, et al. Consumers’ misunderstanding of health insurance. J Health Econ. 2013;32(5):850-862. doi:10.1016/j.jhealeco.2013.04.004

26. Bhargava S, Loewenstein G, Sydnor J. Choose to lose: health plan choices from a menu with dominated option. Q J Econ. 2017;132(3):1319-1372. doi:10.1093/qje/qjx011

27. Park S, Langellier BA, Meyers DJ. Association of health insurance literacy with enrollment in traditional Medicare, Medicare Advantage, and plan characteristics within Medicare Advantage. JAMA Netw Open. 2022;5(2):e2146792. doi:10.1001/jamanetworkopen.2021.46792

28. Botti S, Iyengar SS. The dark side of choice: when choice impairs social welfare. J Public Policy Mark. 2006;25(1):24-38. doi:10.1509/jppm.25.1.24

29. Tsui J, Hirsch JA, Bayer FJ, et al. Patterns in geographic access to health care facilities across neighborhoods in the United States based on data from the National Establishment Time-Series between 2000 and 2014. JAMA Netw Open. 2020;3(5):e205105. doi:10.1001/jamanetworkopen.2020.5105

30. US Senate Committee on Finance. Deceptive Marketing Practices Flourish in Medicare Advantage. US Senate Committee on Finance; 2022. Accessed June 17, 2023. https://www.finance.senate.gov/imo/media/doc/
Deceptive%20Marketing%20Practices%20Flourish%20in%20Medicare%20Advantage.pdf