Support and Internet Use in Navigating Medicare Plans Among Older Americans

1. Introduction

U.S. counties offer an average of 42 Medicare Advantage plans [1] and 14 stand‐alone prescription drug plans [2], which vary in benefits, provider networks, operations (e.g., prior authorizations), and quality. Insurance decision‐making may be particularly challenging for older adults given their complex health needs [3]. Insurance decision‐making tools and resources are available online, but older adults may have limited access or comfort [4, 5]. To support older adults' enrollment choices—critical decisions with enormous financial and health implications—it is important to understand the resources they use to make insurance‐related decisions.

2. Methods

We used the 2015–2023 National Health and Aging Trends Study (NHATS), a nationally representative, longitudinal survey of Medicare beneficiaries aged 65 or older. Surveys were fielded between May and December each year except in 2023 (March 2023–April 2024), with response rates of 50.8%–96.0%. Respondents were asked whether they changed Medicare plans (i.e., supplemental, prescription drug, Medicare Advantage) in the past year and, if so, whether anyone helped. Respondents were also asked whether they used the internet to handle Medicare insurance matters (e.g., comparing plans, looking up coverage, filing claims).

From 2015 to 2023, we examined prevalences of two outcomes: receiving help changing Medicare plans and using the internet for insurance matters. For cohort comparability across years, we limited the analysis each year to respondents aged 71 or older. We tested the trend using a linear probability model with year included as a continuous variable and additional binary indicators for years 2021, 2022, and 2023 to allow for deviations from the 2015 to 2020 linear trend during the COVID‐19 pandemic; we did not include an additional binary indicator for the year 2020 since the survey asked about respondents' receipt of help and internet use in the past year. Then we used logistic regression to examine factors associated with each outcome, adjusting for respondent characteristics (Table 1). For this analysis, we used the 2022 survey and included respondents aged 65 or older to generate nationally representative estimates for older adults of all ages, calculating predicted probabilities and marginal effects. We used NHATS analytic weights, adjusted for complex survey design. Statistical significance was set at two‐tailed p < 0.05; analyses were performed using Stata, version 18.0 (StataCorp LLC).

TABLE 1.

Factors associated with receipt of help and internet use in navigating medicare plans by adults aged 65 years or older in 2022 ^a .

	Older adults who changed Medicare plans			All older adults
	Total	Receive help changing Medicare plans		Total	Use the internet to handle insurance matters
Respondents, n	711	270		5900	1521
National estimates, millions	7.4	2.8		53.4	17.7
Weighted %	100	38.0		100	32.7
Characteristics		Adjusted probability, % (95% CI)	p		Adjusted probability, % (95% CI)	p
Sex
Male	281 (42.8)	37.9 (29.5, 46.4)	[Reference]	2486 (44.9)	31.9 (28.3, 35.5)	[Reference]
Female	430 (57.2)	38.1 (30.5, 45.6)	0.98	3414 (55.1)	33.5 (30.0, 37.0)	0.34
Age, years
65–74	279 (63.4)	36.6 (28.8, 44.4)	[Reference]	1796 (55.9)	35.6 (31.6, 39.5)	[Reference]
75–84	326 (30.5)	39.6 (31.6, 47.7)	0.45	2700 (33.2)	30.3 (27.4, 33.2)	0.003
85+	106 (6.1)	43.5 (32.6, 54.3)	0.30	1404 (10.9)	23.3 (19.6, 27.0)	< 0.001
Race and ethnicity
Non‐Hispanic White	443 (77.9)	38.2 (30.4, 45.9)	[Reference]	3740 (77.9)	33.4 (29.9, 36.9)	[Reference]
Non‐Hispanic Black	156 (9.0)	26.4 (16.7, 36.1)	0.06	1217 (8.8)	29.8 (25.4, 34.2)	0.12
Hispanic	80 (7.9)	31.2 (19.2, 43.3)	0.27	700 (8.3)	30.7 (25.0, 36.3)	0.37
Other b	32 (5.3)	65.4 (48.8, 82.1)	0.005	243 (5.0)	28.9 (21.9, 35.8)	0.24
Some college or above
No	273 (30.1)	37.2 (29.2, 45.2)	[Reference]	2445 (34.2)	20.4 (17.0, 23.8)	[Reference]
Yes	438 (69.9)	38.4 (30.1, 46.6)	0.82	3455 (65.8)	37.6 (34.0, 41.1)	< 0.001
Total family income, by quartiles c
$0–$24,985	247 (26.1)	32.8 (20.9, 44.6)	[Reference]	1917 (24.7)	18.1 (13.2, 22.9)	[Reference]
$24,986–$49,986	193 (26.2)	34.3 (25.9, 42.7)	0.81	1501 (23.5)	24.9 (20.8, 28.9)	0.003
$49,987–$92,723	162 (25.0)	43.0 (33.6, 52.5)	0.12	1389 (26.8)	36.7 (32.5, 41.0)	< 0.001
$92,724+	109 (22.6)	41.4 (31.9, 50.9)	0.20	1093 (25.0)	45.7 (40.0, 51.4)	< 0.001
Medicaid dual eligibility
No	568 (86.1)	38.1 (30.3, 45.8)	[Reference]	4870 (87.2)	33.1 (29.9, 36.3)	[Reference]
Yes	143 (13.9)	37.7 (22.5, 52.8)	0.96	1030 (12.8)	28.6 (21.9, 35.2)	0.18
Living alone
No	466 (67.7)	37.6 (31.2, 44.1)	[Reference]	3923 (71.7)	32.6 (28.9, 36.2)	[Reference]
Yes	245 (32.3)	39.2 (25.7, 52.6)	0.81	1977 (28.3)	33.3 (28.4, 38.1)	0.80
Living in the community
No	48 (4.8)	37.4 (18.5, 56.4)	[Reference]	388 (4.7)	35.4 (26.5, 44.4)	[Reference]
Yes	663 (95.2)	38.0 (31.3, 44.8)	0.95	5512 (95.3)	32.6 (29.6, 35.7)	0.49
Presence of partner
No	378 (45.2)	31.2 (22.0, 40.5)	[Reference]	3066 (41.5)	32.1 (27.0, 37.2)	[Reference]
Yes	333 (54.8)	42.1 (32.4, 51.8)	0.14	2834 (58.5)	33.0 (29.6, 36.5)	0.75
Any living daughters
No	213 (31.3)	36.8 (28.6, 45.1)	[Reference]	1647 (30.8)	35.0 (30.5, 39.4)	[Reference]
Yes	498 (68.7)	38.5 (30.7, 46.2)	0.73	4253 (69.2)	31.7 (28.4, 35.0)	0.13
Number of family and unpaid caregivers d
0	124 (19.7)	9.5 (2.5, 16.4)	[Reference]	995 (17.7)	28.1 (23.9, 32.3)	[Reference]
1	279 (43.9)	33.9 (25.0, 42.7)	< 0.001	2597 (50.9)	32.3 (28.9, 35.8)	0.09
2	189 (24.2)	50.7 (40.5, 60.9)	< 0.001	1457 (21.5)	36.2 (31.1, 41.4)	0.002
3+	119 (12.2)	66.7 (52.9, 80.4)	< 0.001	851 (9.9)	37.1 (30.4, 43.8)	0.01
General health status
Excellent or very good	277 (46.3)	37.0 (29.1, 44.8)	[Reference]	2222 (45.1)	31.8 (28.1, 35.4)	[Reference]
Good	261 (32.5)	37.3 (28.0, 46.6)	0.95	2145 (33.7)	33.6 (30.1, 37.0)	0.31
Poor or fair	173 (21.2)	41.4 (30.5, 52.2)	0.48	1533 (21.2)	34.0 (28.7, 39.3)	0.41
Functional limitations e
Any mobility activities
No	430 (67.8)	40.3 (32.4, 48.2)	[Reference]	3647 (68.9)	32.8 (29.4, 36.2)	[Reference]
Yes	281 (32.2)	33.6 (24.4, 42.7)	0.21	2253 (31.1)	32.5 (28.8, 36.3)	0.88
Any self‐care activities
No	504 (78.5)	36.3 (29.7, 42.9)	[Reference]	4077 (76.4)	32.6 (29.2, 36.0)	[Reference]
Yes	207 (21.5)	43.8 (32.0, 55.6)	0.18	1823 (23.6)	33.3 (29.0, 37.7)	0.75
Any household activities
No	389 (60.5)	38.8 (31.2, 46.3)	[Reference]	3151 (61.1)	33.5 (30.0, 37.0)	[Reference]
Yes	322 (39.5)	36.9 (26.7, 47.1)	0.74	2749 (38.9)	31.2 (27.4, 35.0)	0.22
Probable dementia f
No	656 (95.7)	37.2 (30.5, 43.8)	[Reference]	5229 (93.4)	33.3 (30.1, 36.6)	[Reference]
Yes	55 (4.3)	55.6 (37.3, 73.9)	0.04	671 (6.6)	19.3 (12.6, 26.0)	< 0.001

^a We used a logistic regression model to examine factors associated with receiving help changing Medicare plans and using the internet to handle insurance matters. Covariates included respondent demographic, socio‐economic, and health characteristics presented in the table. To facilitate interpretation, we calculated the mean predicted probability of receiving help changing Medicare plans and using the internet to handle insurance matters associated with each characteristic and compared whether the predicted probabilities were different between strata of the characteristic by calculating the marginal effect of that characteristic using estimates from the logistic regression. Data were weighted using the National Health and Aging Trends Study (NHATS) analytic weights; Stata svy commands were used to incorporate strata and clustering elements of the sample design when calculating standard errors of the estimates.

^b The other race and ethnicity category includes American Indian, Alaska Native, Asian, Native Hawaiian, Pacific Islander, and other races.

^c Total family income includes respondent and partner earnings, pension and annuity income, government social insurance payments, family capital, and business or farm income.

^d Individuals related to the care recipient such as partner, adult child, grandchild, or sibling, as well as unrelated, unpaid helpers such as friend or neighbor.

^e Self or proxy report of difficulties in performing or needing help with the following tasks: mobility (i.e., getting out of bed, getting around inside and outside), self‐care (i.e., bathing, dressing, eating, and using the toilet), and household activities (i.e., doing laundry, going shopping, preparing meals, handling banking, and managing medications).

^f NHATS‐developed probable dementia (vs. possible or no dementia) was determined based on a doctor's diagnosis of dementia or Alzheimer's disease, a screening instrument administered to proxy respondents, or performance on measures of memory, orientation, and executive function.

3. Results

The proportion of respondents (N = 44,467 respondent‐years) who changed Medicare plans was stable at 10%–12% each year from 2015 to 2022 and increased slightly to 13.5% in 2023 (p = 0.01; Figure 1). Among those who changed, approximately 50% received help until 2020, but this dropped to 40% by 2022 (p = 0.02). The proportion of respondents who used the internet for insurance matters slowly increased to 15% by 2020, then to 30%–31% in 2022 and 2023 (p < 0.001 for all).

FIGURE 1.

Trends in receipt of help and internet use in navigating medicare plans among adults aged 71 years or older from 2015 to 2023. The youngest National Health and Aging Trends Study (NHATS) respondents by 2021 were 71 years old, so we limited the trend analysis to respondents aged 71 or older in each year. The denominators for the proportion of older adults who changed Medicare plans (Green line) and who used the internet to handle insurance matters (Red line) were all older adults (N = 44,467 respondent‐years across all years); the denominators for the proportion of older adults who received help changing Medicare plans (Blue line) were older adults who changed Medicare plans in the past year (N = 4633 respondent‐years). We tested the trend using a linear probability model with year included as a continuous variable and additional binary indicators for years 2021, 2022, and 2023 to allow for deviations from the 2015–2020 linear trend during the COVID‐19 pandemic; we did not include an additional binary indicator for the year 2020 since the survey asked about receipt of help and internet use in the past year. p values for the 2015–2020 linear trend and deviations from the linear trend in 2021, 2022, and 2023: Proportion of older adults who changed Medicare plans (0.66 for 2015–2020, 0.76 for 2021, 0.13 for 2022, and 0.01 for 2023); proportion of older adults who received help changing Medicare plans (0.40 for 2015–2020, 0.13 for 2021, 0.02 for 2022, and 0.13 for 2023); and proportion of older adults who used the internet to handle insurance matters (< 0.001 for 2015–2020, < 0.001 for 2021, < 0.001 for 2022, and < 0.001 for 2023). Data were weighted using NHATS analytic weights; Stata svy commands were used to incorporate strata and clustering around respondents to account for correlation within respondents that were observed over multiple years when calculating standard errors of the estimates. Error bars indicated 95% CIs of the weighted percentages.

Among 5900 respondents aged 65 or older in 2022, 13.4% (N = 711) changed Medicare plans; of these, 38% received help (Table 1), coming from partners (30.5%), other relatives (18.6%), or non‐relatives (53.3%). The probability of receiving help was higher among other race and ethnicity (i.e., Asian, Pacific Islander, Native Hawaiian, American Indian, Alaska Native, or other) relative to non‐Hispanic White and increased with the number of caregivers and among respondents with probable dementia but did not differ by other characteristics. Thirty‐three percent of respondents used the internet for insurance matters, with a higher probability among those who were younger, had at least a college education, had higher income, had more caregivers, and did not have dementia.

4. Discussion

Among U.S. older adults who changed Medicare plans in 2022, only 38% received help with this decision and 33% used the internet to handle Medicare‐related matters. While the latter has grown since 2015, it is unclear whether the steep increase through 2022 and 2023 will continue. Social isolation related to the COVID‐19 pandemic may explain the findings that, from 2021 through 2023, fewer older adults received help changing health plans while more turned to the internet for insurance‐related matters.

In each of the 9 years analyzed, no more than half of older adults reported receiving help changing Medicare plans. Given the large financial implications of this decision and the enormous number of options available, it may not be ideal that so few older adults report help with this complex task, though they may be unaware of resources available (e.g., Medicare‐funded State Health Insurance Assistance Programs) to assist [6]. While insurance companies and government‐funded decision aids (e.g., Medicare Plan Finder) offer insurance‐related resources online, just one‐third of older adults reported using such technology for insurance matters, suggesting a mismatch between where insurance companies and Medicare beneficiaries prefer to conduct business. A Limitation of our study is that we did not know the specific insurance‐related matters older adults dealt with using the internet.

Author Contributions

Study concept and design: Lei, Maust. Acquisition, analysis, or interpretation of data: All authors. Drafting of the manuscript: Lei, Maust. Critical revision of the manuscript for important intellectual content: All authors. Approval of the submitted version: All authors.

Conflicts of Interest

The authors declare no conflicts of interest.

Lei L., Petzold K., Strominger J., and Maust D. T., “Support and Internet Use in Navigating Medicare Plans Among Older Americans,” Journal of the American Geriatrics Society 73, no. 9 (2025): 2923–2927, 10.1111/jgs.19516.

Data Availability Statement

Data from the National Health and Aging Trends Study was publicly available via its online platform.