• Center on Health Equity & Access
  • Clinical
  • Health Care Cost
  • Health Care Delivery
  • Insurance
  • Policy
  • Technology
  • Value-Based Care

Insurance Coverage and Health Care Spending by State-Level Medigap Regulations

Publication
Article
The American Journal of Managed CareApril 2022
Volume 28
Issue 4

Despite their intention to protect against coverage denial and/or premium increases, additional state-level Medigap regulations are correlated with lower Medigap enrollment and stronger moral hazard.

ABSTRACT

Objectives: Medigap protects traditional Medicare (TM) beneficiaries against catastrophic expenses. Federal regulations around Medigap enrollment and pricing are limited to the first 6 months after turning 65 years old. Eight states institute regulations that apply to later enrollment; half use community rating (charging everyone the same premium) and half use both community rating and guaranteed issue (requiring insurers to accept any beneficiary irrespective of health conditions). We examined the impact of state-level Medigap regulations on insurance coverage and health care spending for Medicare beneficiaries.

Study Design: We used a retrospective cohort study design. Using the 2010-2016 Medicare Current Beneficiary Survey, we identified beneficiaries with TM only, TM + Medigap, or Medicare Advantage (MA) by state-level Medigap regulations.

Methods: Outcomes were insurance coverage and health care spending. We used an instrumental variable approach to address endogenous insurance choice. We conducted 2-stage least squares regression while controlling for individual-level characteristics and area-level demographic characteristics. Then we used the recycled prediction methods to predict enrollment and spending outcomes for the 3 state-level Medigap regulation scenarios.

Results: Although enrollment in TM only was consistent across regulation scenarios, the scenario with community rating and guaranteed issue had lower Medigap enrollment and higher MA enrollment than the no-regulation scenario. Despite negligible health differences, TM + Medigap beneficiaries had higher Medicare spending than TM-only beneficiaries, suggesting moral hazard.

Conclusions: Our findings suggest a link between additional regulations and lower Medigap and higher MA enrollment. Policy makers should consider the potential effects on insurance coverage, premiums, financial protection, and moral hazard when designing Medigap regulations.

Am J Manag Care. 2022;28(4):172-179. https://doi.org/10.37765/ajmc.2022.88860

_____

Takeaway Points

Despite their intention to protect against coverage denial and/or premium increases, additional state-level Medigap regulations are correlated with lower Medigap enrollment and stronger moral hazard.

  • Although enrollment in traditional Medicare (TM) only was consistent across regulation regimes, states with additional regulations had lower Medigap enrollment and higher Medicare Advantage enrollment than states without regulations.
  • Despite negligible observable health differences, TM beneficiaries with Medigap had higher Medicare spending than TM-only beneficiaries, suggesting moral hazard.
  • Differences in both enrollment and spending were notable in states with community rating and guaranteed issue.
  • Policy makers should consider how to design Medigap regulations to ensure financial protections while minimizing moral hazard.

_____

Medigap is private supplemental insurance for traditional Medicare (TM) beneficiaries to protect against catastrophic expenses because TM has significant co-pays and deductibles and no limit on out-of-pocket spending. In 2018, 33.7% of TM beneficiaries purchased Medigap.1 Medigap fills in these gaps of TM (parts A and B) insurance by paying co-pays or deductibles, often reducing out-of-pocket spending on Medicare-covered services to zero. However, Medigap does not cover long-term care, dental care, or prescription drugs.

Medigap pricing is regulated at both the federal and state levels.2 Under federal regulations, TM beneficiaries who purchase Medigap within 6 months after turning 65 years old cannot be denied coverage and pricing is based on a limited set of factors such as age and sex. Those who purchase Medigap during this period are guaranteed renewal at the same price offered to others each year. However, only 25% of TM beneficiaries purchase Medigap at age 65 years when they are subject to these federal regulations.3 For those who enroll after age 65, Medigap insurers may deny coverage or charge higher premiums based on preexisting health conditions. This lack of regulatory protection could expose TM beneficiaries to lifelong consequences in the form of coverage denial and/or premium increases.

Eight states institute additional state-level Medigap regulations beyond the age-65 enrollment period to address this risk. Currently, 4 states (Alaska, Minnesota, Vermont, and Washington) require community rating (charging all beneficiaries the same premium regardless of health conditions), regardless of age of first purchase. Four other states (Maine, Connecticut, Massachusetts, and New York) require both community rating and guaranteed issue (in which all beneficiaries must be offered coverage irrespective of their health conditions), regardless of age of first purchase.

Prior research on Medigap has mainly focused on examining selection and moral hazard in Medigap at the aggregate level,4-8 and thus limited information exists on the association between state-level Medigap regulations and insurance coverage or health care spending for Medicare beneficiaries. State-level Medigap regulations are designed to protect TM beneficiaries against coverage denial and/or premium increases, but these regulations may distort the market, limiting its ability to provide financial protection.9 This has critical implications for Medicare beneficiaries as well as for the federal government.

For Medicare beneficiaries, differences in state-level Medigap regulations may lead to differential insurance coverage. Low-risk beneficiaries in states with Medigap regulations may delay Medigap coverage until they develop a chronic condition, leading a disproportionately high Medigap enrollment of high-risk beneficiaries, further increasing Medigap premiums.10 Consequently, this may lead to higher enrollment in Medicare Advantage (MA) because MA must cover all of the services that TM covers (parts A and B) and may offer additional benefits11,12 and/or lower prices13,14 compared with TM. Prior research has shown a positive association between Medigap premiums and MA enrollment,15,16 suggesting that MA plans could be an alternative to Medigap coverage.

For the federal government, these Medigap regulations may increase financial burden as TM administered by the federal government pays for a large portion of care including both the care offered on the margin and the excess care. Evidence suggests that Medigap coverage may lead to excessive health care utilization, suggesting moral hazard.8,17 However, it is important to disentangle from moral hazard because although moral hazard is associated with enrollment of low-risk beneficiaries,8,17 state-level Medigap regulations may result in a disproportionately high enrollment of high-risk beneficiaries.10

In this study, we estimated the associations between state-level Medigap regulations and insurance coverage and health care spending for Medicare beneficiaries.

METHODS

Data and Study Sample

We used data from multiple sources for 2010 to 2016. First, we used the Medicare Current Beneficiary Survey (MCBS), which provides a nationally representative sample of the Medicare population. The data combine information from Medicare claims and administrative data with an interview survey. We could not use the 2014 MCBS data because they were never released. Second, we used the Area Health Resources Files to obtain county-level demographic data. Finally, we used the data collected by the National Association of Insurance Commissioners to obtain state-specific monthly Medigap premiums.

We identified Medicare beneficiaries 65 years or older with 12 calendar months of continuous enrollment in parts A and B benefits. We excluded beneficiaries based on insurance (those who were eligible for Medicaid; whose original Medicare eligibility was due to disability or end-stage renal disease; who had employer-sponsored health insurance; or who had switched coverage within the year), reporting problems (those who had missing survey questions; or who reported both Medigap and MA, which is illegal in practice), and who died within the year. Eight states were excluded because the survey questions used in this analysis were missing (Alaska, Delaware, Hawaii, Idaho, Maine, Montana, North Dakota, and Oregon).

Outcomes

We included 2 types of outcomes: insurance enrollment and health care spending. First, we created a binary indicator of whether a beneficiary enrolled in TM only, TM + Medigap, or MA. We determined these 3 mutually exclusive types of insurance coverage based on 12 calendar months of continuous enrollment. We defined Medigap as self-purchased private health insurance. Second, we measured 4 types of health care spending: Medicare spending, out-of-pocket spending, service-specific spending (inpatient spending, outpatient spending, and medical provider spending), and out-of-pocket spending as a share of the sum of out-of-pocket spending and Medicare spending. Following prior research,18 we calculated out-of-pocket spending for health care services and premiums. We used administrative data for out-of-pocket spending on premiums for TM and self-reported survey data for out-of-pocket spending on premiums for Medigap and MA plans. All health care spending was inflation adjusted to 2019 US$.

Instrumental Variables

We used an instrumental variable approach to control for unmeasured confounding due to selection bias. Prior research has shown that beneficiaries with TM + Medigap tend to be healthier than beneficiaries with TM only,3,8,17 suggesting favorable selection into Medigap. However, state-level differences in Medigap regulatory regimes may lead to differential selection into Medigap enrollment, with some states having healthier populations in MA than others. We used state-specific Medigap regulations as an instrument for the average monthly Medigap premiums (eAppendix [available at ajmc.com]). It is hypothesized that more regulations in the Medigap market are positively related to higher Medigap premiums, and it is presumed that the regulations themselves are not directly related to insurance coverage and health care spending, but are only related through the increase in premiums. Three types of state-specific Medigap regulation regimes are in our data: 37 (36 states plus the District of Columbia) without Medigap regulations, 3 states with community rating only, and 3 states with both community rating and guaranteed issue. Eight states including 2 states with Medigap regulations were excluded due to missing data in survey questions for beneficiaries in those states. No changes in state-level Medigap regulations occurred during the study period.

Covariates

To control for differences in individual-level and area-level characteristics, we included individual-level factors (age, gender, race/ethnicity, education, income, marital status, residence in metro area, Census region of residence, comorbidities, number of chronic conditions, number of activities of daily living [ADL] limitations, and general health status) and county-level factors (number of hospital beds per capita, median household income, percentage of residents with incomes below poverty level, unemployment rate, total population size, percentage of the population 65 years and older, number of primary care physicians, and number of specialist physicians). We also included the county-level MA enrollment rate because the MA penetration rate in a county is positively related to MA enrollment, possibly leading to lower enrollment in Medigap.

Statistical Analysis

We estimated a 2-stage least squares regression model. In the first stage, we obtained the predicted monthly Medigap premiums based on state Medigap regulations. In the second stage, we estimated the association between the predicted monthly Medigap premiums estimated from the first stage and the outcomes of interest. For each binary measure of insurance coverage, we conducted a linear probability model for enrollment and a generalized linear model (GLM) with gamma distribution and log link function for health care spending. Both stages adjusted for the covariates described earlier.

We used the recycled prediction methods to estimate and compare the enrollment and spending outcomes in 3 hypothetical scenarios: one in which all beneficiaries are in states without Medigap regulations, another in which all beneficiaries are in states with community rating only, and a third in which all beneficiaries are in states with both community rating and guaranteed issue. We predicted the mean outcome holding constant all other variables, allowing us to compare the outcome of interest across all 3 scenarios for all beneficiaries. We conducted this exercise to produce interpretable and comparable results for 2 reasons. First, the estimate from the second stage measures the extent of an increase in outcome associated with a $1 increase in monthly Medigap premiums. However, the finding of interest in this study is the absolute level of the outcome by state-level Medigap regulations. Second, GLMs produce less easily interpretable coefficients for health care spending. This requires us to rescale the findings so that they can be interpreted as dollar values.

For our analysis for insurance coverage, we examined whether the relationship between the state-level Medigap regulations and insurance coverage differs by income and health, the main sources of selection into Medigap.7 We estimated enrollment rates for Medicare beneficiaries with incomes lower than $25,000, those with more than 3 chronic conditions, and those who reported their health status as good or excellent.

For all analyses, we adjusted the standard errors for clustering within individuals. We used survey weights and included year-fixed effects.

RESULTS

We included a total of 22,898 Medicare beneficiaries (Table 1 [part A and part B]). There were several differences in sample characteristics among TM-only, TM + Medigap, and MA beneficiaries, and these differences were consistently observed across the Medigap regulation categories. TM-only beneficiaries were more likely to have income less than $25,000, to be unmarried, to have more than 3 ADL limitations, and to report health status as very poor or poor than TM + Medigap and MA beneficiaries. TM + Medigap beneficiaries were more likely to be White, to have more than a college degree, to have income more than $50,000, and to have cancer than TM + Medigap and MA beneficiaries. However, these characteristics were similar between TM + Medigap and MA beneficiaries. Unadjusted outcomes are presented in the eAppendix Table.

Our first-stage regression analysis showed that state-specific Medigap regulations were significantly and strongly predictive of the monthly Medigap premiums and F statistics indicate strong instruments19 (Table 2). Monthly Medigap premiums were $11.24 and $52.26 higher in states with community rating only and in states with both community rating and guaranteed issue, respectively, compared with states without regulation. Most individual-level covariates were balanced across values of the instrument.

Our second-stage regression analysis showed that monthly Medigap premiums were significantly associated with enrollment changes in TM + Medigap and MA (Table 3). A $1 increase in monthly Medigap premiums was associated with a 0.28% decrease in TM + Medigap enrollment and a 0.24% increase in MA enrollment. There was no significant association between monthly Medigap premiums and enrollment in TM only.

There were limited differences in health care spending associated with a $1 increase in monthly Medigap premiums (Table 3). A $1 increase in monthly Medigap premiums was associated with decreases in Medicare spending and medical provider spending among TM-only beneficiaries (coefficients = –1.36 and –0.73, respectively) and outpatient spending among TM + Medigap beneficiaries (coefficient, –0.69). No significant associations were observed in other outcomes or groups.

Our analysis of insurance coverage under hypothetical Medigap regulation scenarios showed that if everyone was in a state adopting both community rating and guaranteed issue, one would expect to have lower TM + Medigap enrollment and higher MA enrollment than if all states eliminated Medigap regulations or adopted community rating only (36.3%, 37.22%, and 24.80% for TM + Medigap enrollment and 39.73%, 39.03%, and 50.34% for MA enrollment in the no-regulation scenario, the community rating–only scenario, and the scenario with both community rating and guaranteed issue, respectively) (Table 4). However, we found relatively constant enrollment rates for TM-only beneficiaries across the 3 regulation scenarios (23.90%, 23.75%, and 24.80%). Furthermore, enrollment rates differed by certain characteristics. Those with income less than $25,000 or those who reported health status as good or excellent were more likely to enroll in MA than in TM + Medigap or TM only.

Our analysis of health care spending under hypothetical Medigap regulation scenarios showed that TM + Medigap beneficiaries were expected to have higher Medicare spending than TM-only beneficiaries across all regulation scenarios ($11,801 and $10,170, respectively, for those in the scenario with no Medigap regulation; $11,332 and $8572 in the community rating–only scenario; and $12,336 and $10,142 in the scenario with both community rating and guaranteed issue) (Table 5). Also, differences in Medicare spending between TM beneficiaries and TM + Medigap beneficiaries were greater in the community rating–only scenario and the scenario with both community rating and guaranteed issue than in the scenario with no Medigap regulation ($2760 and $2194 vs $1631). A similar pattern was observed in Medicare inpatient, outpatient, and medical provider spending. Further, TM + Medigap beneficiaries had higher out-of-pocket spending, despite the higher insurance coverage, than TM-only beneficiaries across all regulation scenarios ($9085 and $5752, respectively, in the no-regulation scenario; $9744 and $5752 in the community rating–only scenario; and $10,839 and $5942 in the scenario with both community rating and guaranteed issue). However, there were marginal differences in out-of-pocket share between TM-only beneficiaries and TM + Medigap beneficiaries across all regulation scenarios. MA beneficiaries had relatively lower Medicare spending and out-of-pocket spending across all regulation scenarios.

DISCUSSION

Our study provides new evidence of the consequences of the state-level Medigap regulation on insurance coverage and health care spending for Medicare beneficiaries. The analyses allow us to address unmeasured confounding due to selection into MA, address the skewness of medical spending, and put our results into easily interpretable and policy-relevant terms. Although enrollment in TM only was consistent across states, states with both community rating and guaranteed issue had lower TM + Medigap enrollment and higher MA enrollment than states without Medigap regulations. We also found evidence of moral hazard, which might be more pronounced in states with Medigap regulations. Despite negligible differences in observable health status, TM + Medigap beneficiaries had higher Medicare spending than TM-only beneficiaries, consistent with findings of prior research.8,17

Our findings suggest that community rating and guaranteed issue regulations for Medigap may not necessarily improve access to Medigap coverage. In states with both community rating and guaranteed issue regulations, which enable Medicare beneficiaries to switch out of MA and still purchase Medigap after a diagnosis of a costly medical condition, we found that MA plans serve as an important insurance alternative. The high premiums coupled with the reduced financial risk of delaying Medigap enrollment led to low TM + Medigap enrollment and high MA enrollment. This was more pronounced among those with income less than $25,000 and those who reported their health status as good or excellent. However, in states with only community rating regulations, which protect against pricing on individual health conditions, the uncertainty about the ability to gain Medigap coverage in the future led to a relatively small impact overall. It is important to note that there were negligible differences in Medigap enrollment rates at age 65 years, based on state-level regulations protecting beneficiaries who enroll at later ages.3 This suggests that beneficiaries are likely unaware of these regulations and pricing until they face the decision to enroll themselves.

Despite the heterogeneous consequences of the regulatory regimes on insurance coverage, we found a similar level of impact on health care spending for Medicare beneficiaries in community-rating states, with and without guaranteed issue. The first notable finding is that, compared with TM-only beneficiaries, TM + Medigap beneficiaries had similar health status, as measured by comorbidities, but had higher Medicare spending, suggesting evidence of moral hazard.8,17 Our estimated moral hazard effect ($1631 as measured by differences in Medicare spending between TM-only beneficiaries and TM + Medigap beneficiaries if all states had no regulation) was close to the estimate from prior research ($1615).8 Our findings provide suggestive evidence that state-level Medigap regulations may amplify the moral hazard effects. This may be due to the Medigap regulations increasing premiums, possibly resulting in discretionary high health care needs and encouraging excessive health care utilization. Although we found few differences in observable health status between TM-only and TM + Medigap beneficiaries, there may be unobservable differences that we cannot capture. On the other hand, the level of health care spending for MA beneficiaries was relatively low. Because MA providers are paid on a capitated basis rather than for each service performed, this creates the incentive for them to be efficient in their approach to care.20-22

Our findings have important policy implications for designing Medigap regulations. Although the state-level Medigap regulations are designed to protect against catastrophic expenses, these led to unintended consequences by potentially amplifying moral hazard. Indeed, there are several proposals that are intended to improve efficiency of the Medigap program. Some proposals intend to apply a premium surcharge or excise tax on Medigap premiums.17 However, policy makers should consider how these changes to the Medigap program also affect the financial protection provided to Medicare beneficiaries. The additional state-level Medigap regulations allow Medicare beneficiaries to have access to Medigap, whereas these also lead to higher Medigap premiums, which may place further limits on insurance coverage for Medicare beneficiaries with low income. Furthermore, this may induce additional adverse consequences by continuing to raise Medigap premiums and consequently falling into a death spiral, which may in turn eliminate an important safety valve for Medicare beneficiaries to protect against catastrophic expenses.

Limitations

Our study had limitations. The MCBS provides a relatively small sample size, especially for small states, and thus this may limit our ability to perform a representative cross-state comparison of Medigap regulations. However, prior research found similar results using large administrative or claims data sets.3,23 Moreover, we merely assessed the associations of state-level Medigap regulations with insurance coverage and health care spending at the time of being fully implemented. Thus, it is less known whether the Medigap regulation passed because Medigap played a smaller role in the states historically or whether Medigap played a smaller rolebecause of the regulation. Furthermore, states with different Medigap regulatory regimes may differ in various ways. For example, Medigap premiums differ by plan and insurance provider. Also, state-level differences may exist in insurance markets, which might lead to biased results.

CONCLUSIONS

Our findings suggest that additional state-level Medigap regulations are correlated with lower TM + Medigap enrollment and higher MA enrollment. Policy makers should consider the potential effects on insurance coverage, premiums, financial protection, and moral hazard when designing Medigap regulations.

Author Affiliations: Department of Health Management and Policy, Dornsife School of Public Health, Drexel University (SP), Philadelphia, PA; Department of Health Convergence, College of Science and Industry Convergence, Ewha Womans University (SP), Seoul, Republic of Korea; Department of Medical Ethics and Health Policy, Perelman School of Medicine, University of Pennsylvania (NBC), Philadelphia, PA.

Source of Funding: This work was supported by grant R01 AG049815 from the National Institutes of Health.

Author Disclosures: The 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 (SP, NBC); acquisition of data (NBC); analysis and interpretation of data (SP, NBC); drafting of the manuscript (SP); critical revision of the manuscript for important intellectual content (SP, NBC); statistical analysis (SP, NBC); obtaining funding (NBC); administrative, technical, or logistic support (NBC); and supervision (NBC).

Address Correspondence to: Sungchul Park, PhD, Department of Health Management and Policy, Dornsife School of Public Health, Drexel University, Nesbitt Hall, 3215 Market St, Philadelphia, PA 19104. Email: smp462@drexel.edu.

REFERENCES

1. State of Medigap: trends in enrollment and demographics. AHIP. June 2020. Accessed September 1, 2021. https://www.ahip.org/documents/AHIP_State_of_Medigap-2020.pdf

2. Boccuti C, Jacobson G, Orgera K, Neuman T. Medigap enrollment and consumer protections vary across states. Kaiser Family Foundation. July 11, 2018. Accessed September 1, 2021. https://www.kff.org/medicare/issue-brief/medigap-enrollment-and-consumer-protections-vary-across-states/

3. Curto V. Pricing regulations in individual health insurance: evidence from Medigap. Becker Friedman Institute for Economics at the University of Chicago. June 13, 2018. Accessed September 1, 2021. https://bfi.uchicago.edu/working-paper/pricing-regulations-in-individual-health-insurance-evidence-from-medigap/

4. Wolfe JR, Goddeeris JH. Adverse selection, moral hazard, and wealth effects in the Medigap insurance market. J Health Econ. 1991;10(4):433-459. doi:10.1016/0167-6296(91)90024-h

5. Ettner SL. Adverse selection and the purchase of Medigap insurance by the elderly. J Health Econ. 1997;16(5):543-562. doi:10.1016/s0167-6296(97)00011-8

6. Starc A. Insurer pricing and consumer welfare: evidence from Medigap. RAND J Econ. 2014;45(1):198-220.

7. Fang H, Keane MP, Silverman D. Sources of advantageous selection: evidence from the Medigap insurance market. J Polit Econ. 2008;116(2):303-350. doi:10.1086/587623

8. Keane MP, Stavrunova O. Adverse selection, moral hazard and the demand for Medigap insurance. J Econom. 2016;190(1):62-78. doi:10.1016/j.jeconom.2015.08.002

9. Lo Sasso AT, Lurie IZ. Community rating and the market for private non-group health insurance. J Public Econ. 2009;93(1-2):264-279. doi:10.1016/j.jpubeco.2008.07.001

10. Bundorf MK, Simon KI. The effects of rate regulation on demand for supplemental health insurance. Am Econ Rev. 2006;96(2):67-71. doi:10.1257/000282806777212279

11. Meyers DJ, Durfey SNM, Gadbois EA, Thomas KS. Early adoption of new supplemental benefits by Medicare Advantage plans. JAMA. 2019;321(22):2238-2240.

12. Park S, Langellier BA, Burke RE. Telehealth benefits offered by Medicare Advantage plans in 2020. Med Care. 2021;59(1):53-57. doi:10.1097/MLR.0000000000001408

13. Trish E, Ginsburg P, Gascue L, Joyce G. Physician reimbursement in Medicare Advantage compared with traditional Medicare and commercial health insurance. JAMA Intern Med. 2017;177(9):1287-1295.
doi:10.1001/jamainternmed.2017.2679

14. Baker LC, Bundorf MK, Devlin AM, Kessler DP. Medicare Advantage plans pay hospitals less than traditional Medicare pays. Health Aff (Millwood). 2016;35(8):1444-1451. doi:10.1377/hlthaff.2015.1553

15. Rice T, Snyder RE, Kominski G, Pourat N. Who switches from Medigap to Medicare HMOs? Health Serv Res. 2002;37(2):273-290. doi:10.1111/1475-6773.023

16. McLaughlin CG, Chernew M, Taylor EF. Medigap premiums and Medicare HMO enrollment. Health Serv Res. 2002;37(6):1445-1468. doi:10.1111/1475-6773.10832

17. Cabral M, Mahoney N. Externalities and taxation of supplemental insurance: a study of Medicare and Medigap. Am Econ J Appl Econ. 2019;11(2):37-73. doi:10.1257/app.20160350

18. Cubanski J, Koma W, Damico A, Neuman T. How much do Medicare beneficiaries spend out of pocket on health care? Kaiser Family Foundation. November 4, 2019. Accessed September 1, 2021. https://www.kff.org/medicare/issue-brief/how-much-do-medicare-beneficiaries-spend-out-of-pocket-on-health-care/

19. Staiger D, Stock JH. Instrumental variables regression with weak instruments. Econometrica. 1997;65(3):557-586. doi:10.2307/2171753

20. Huckfeldt PJ, Escarce JJ, Rabideau B, Karaca-Mandic P, Sood N. Less intense postacute care, better outcomes for enrollees in Medicare Advantage than those in fee-for-service. Health Aff (Millwood). 2017;36(1):91-100. doi:10.1377/hlthaff.2016.1027

21. Park S, White L, Fishman P, Larson EB, Coe NB. Health care utilization, care satisfaction, and health status for Medicare Advantage and traditional Medicare beneficiaries with and without Alzheimer disease and related dementias. JAMA Netw Open. 2020;3(3):e201809. doi:10.1001/jamanetworkopen.2020.1809

22. Park S, Larson EB, Fishman P, White L, Coe NB. Health care utilization, process of diabetes care, care satisfaction, and health status in patients with diabetes in Medicare Advantage vs traditional Medicare. Med Care. 2020;58(11):1004-1012. doi:10.1097/MLR.0000000000001390

23. Meyers DJ, Trivedi AN, Mor V. Limited Medigap consumer protections are associated with higher reenrollment in Medicare Advantage plans. Health Aff (Millwood). 2019;38(5):782-787. doi:10.1377/hlthaff.2018.05428

Related Videos
dr carol regueiro
Screenshot of Adam Colborn, JD during an interview
dr carol regueiro
Screenshot of an interview with James Chambers, PhD
Screenshot of an interview with Susan Wescott, RPh, MBA
Stuart Staggs
Lalan Wilfong, MD, during a video interview
dr ian neeland
dr manisha jhamb
Related Content
© 2024 MJH Life Sciences
AJMC®
All rights reserved.