Assessment of Disparities in Digital Access Among Medicare Beneficiaries and Implications for Telemedicine

Abstract

This cross-sectional study uses data from the 2018 American Community Survey to assess disparities in digital access among Medicare beneficiaries by demographic and socioeconomic characteristics.

In response to the coronavirus disease 2019 (COVID-19) pandemic, Medicare temporarily expanded its coverage of telemedicine to all beneficiaries, included visits in the patient’s home, and began paying for audio-only visits at the same rate as video and in-person visits.^1,2 Previously, Medicare (with a few exceptions) limited telemedicine coverage to video visits for rural beneficiaries and required video visits to take place at a medical facility, such as a physician’s office, rather than at a patient’s home.³

Access to technology at home and the ability to use technology may affect use of video or audio-only telemedicine visits by Medicare beneficiaries. Although evidence on the efficacy of video vs audio-only visits is lacking,⁴ audio-only visits might be inadequate in some situations, such as when visual monitoring or diagnosis is important for care. We examined disparities in digital access (ie, access at home to technology that enables video telemedicine visits) among Medicare beneficiaries by socioeconomic and demographic characteristics.

Methods

For this cross-sectional study, we analyzed public use respondent- and household-level data files from the 2018 American Community Survey (ACS; from January 1 2018, to December 31, 2018), a nationally representative survey of the US population. We selected respondents to the ACS who lived in the community (excluding those in nursing homes) and indicated that they were Medicare beneficiaries at the time of the survey. The University of Pittsburgh Institutional Review Board waived study review because this study used deidentified data and was determined to be non–human subjects research.

Among Medicare beneficiaries, we assessed the proportion who did not have (1) a desktop or laptop computer with a high-speed internet subscription, (2) a smartphone with a wireless data plan, or (3) either means of digital access. We examined how access limitations differed by, age, sex, race/ethnicity, marital status, educational level, language, income, enrollment in Medicaid, and disability status. We adjusted for person-level survey weights in the ACS to make our estimates representative of the national Medicare population. Analyses were performed using Stata, version 16 (StataCorp LLC). Reported P values were 2-sided and considered to be statistically significant at P < .05. The eAppendix in the Supplement provides more details about the methods.

Results

The study sample consisted of 638 830 surveyed individuals. When weighted, this sample represented 54 749 082 individuals in the community-dwelling Medicare population.

Overall, 41.4% (95% CI, 40.4%-42.4%) of Medicare beneficiaries lacked access to a desktop or laptop computer with a high-speed internet connection at home, and 40.9% (95% CI, 40.0%-41.8%) lacked a smartphone with a wireless data plan (Table). The proportion of beneficiaries without either form of digital access was 26.3% (95% CI, 25.5%-27.1%), and this proportion varied across demographic and socioeconomic groups. For example, a 50.1% (95% CI, 49.3%-50.9%) of beneficiaries with income of 100% below the federal poverty level lacked digital access compared with 11.5% (95% CI, 11.0%-11.9%) of those with income 400% or more above the federal poverty level (P < .001). The proportion of Medicare beneficiaries with digital access was lower among those who were 85 or older, were widowed, had a high school education or less, were Black or Hispanic, received Medicaid, or had a disability.

Table. Limitations in Computer and Internet Access Among Community-Dwelling Medicare Beneficiaries in 2018^a.

Characteristic	Without desktop or laptop computer with high-speed internetb		Without smartphone with a data plan for wireless internetc		Without any digital accessd
	Proportion (95% CI), %	P valuee	Proportion (95% CI), %	P valuee	Proportion (95% CI), %	P valuee
Among Medicare beneficiaries	41.4 (40.4-42.4)	NA	40.9 (40.0-41.8)	NA	26.3 (25.5-27.1)	NA
Sex
Male	39.2 (38.1-40.2)	<.001	38.6 (37.6-39.7)	<.001	24.0 (23.2-24.9)	<.001
Female	43.3 (42.4-44.2)		42.8 (41.9-43.7)		28.1 (27.3-28.8)
Age, y
<64	46.8 (45.8-47.8)	<.001	35.2 (34.2-36.1)	<.001	24.4 (23.6-25.2)	<.001
65-69	33.5 (32.5-34.3)		29.8 (28.8-30.7)		17.1 (16.4-17.8)
70-74	36.2 (35.1-37.3)		36.1 (35.0-37.2)		21.1 (20.3-22.0)
75-59	42.0 (40.8-43.1)		46.1 (44.9-47.3)		28.6 (27.5-29.6)
80-84	49.9 (48.7-51.1)		56.9 (55.6-58.1)		38.4 (37.2-39.6)
≥85	59.1 (57.9-60.2)		66.5 (65.3-67.7)		50.0 (48.7-51.2)
Race/ethnicity
Non-Hispanic White	38.6 (37.4-39.8)	<.001	40.7 (39.7-41.7)	<.001	24.5 (23.6-25.3)	<.001
Non-Hispanic Black	56.3 (55.0-57.5)		47.9 (46.7-49.0)		37.3 (36.1-38.5)
Hispanic	51.8 (50.7-53.0)		40.1 (38.9-41.2)		31.6 (30.5-32.7)
Other	35.5 (33.4-37.5)		31.2 (29.3-33.1)		20.7 (18.9-22.6)
Marital status
Married	32.4 (31.4-33.5)	<.001	33.5 (32.5-34.5)	<.001	17.9 (17.2-18.7)	<.001
Widowed	54.3 (53.3-55.4)		54.5 (53.4-55.5)		40.6 (39.6-41.7)
Divorced or separated	49.2 (48.1-50.3)		44.8 (43.8-45.8)		31.2 (30.3-32.2)
Never married	51.7 (50.6-52.9)		47.6 (46.4-48.7)		34.3 (33.2-35.4)
Educational attainment
Less than high school	62.3 (61.2-63.4)	<.001	54.8 (53.6-56.1)	<.001	44.8 (43.7-46.0)	<.001
High school	49.9 (48.9-50.8)		50.1 (49.2-51.0)		34.2 (33.5-35.0)
Some college or higher	30.3 (29.5-31.1)		31.4 (30.7-32.2)		16.1 (15.5-16.6)
Language spoken at home
English	41.0 (39.9-42.1)	<.001	41.9 (40.9-42.8)	<.001	26.3 (25.4-27.1)	.01
Spanish	50.2 (49.0-51.4)		38.1 (37.0-39.2)		29.7 (28.6-30.9)
Other	36.7 (35.5-37.9)		34.6 (33.3-35.9)		22.5 (21.3-23.6)
Household income, % of FPLf
<100	67.5 (66.7-68.2)	<.001	61.9 (61.1-62.7)	<.001	50.1 (49.3-50.9)	<.001
100 to <200	59.3 (58.5-60.1)		58.5 (57.5-59.4)		43.3 (42.4-44.2)
200 to <300	44.1 (43.2-45.0)		45.5 (44.5-46.4)		27.9 (27.1-28.6)
300 to <400	35.9 (34.9-36.8)		37.1 (36.2-38.0)		20.3 (19.6-21.0)
≥400 FPL	25.0 (24.2-25.8)		24.5 (23.9-25.2)		11.5 (11.0-11.9)
Enrolled in Medicaid
Yes	54.4 (53.4-55.3)	<.001	47.3 (46.2-48.4)	<.001	36.1 (35.2-37.0)	<.001
No	38.5 (37.5-39.5)		39.5 (38.5-40.4)		24.0 (23.2-24.8)
Has disabilityg
Yes	48.9 (48.0-49.9)	<.001	48.1 (47.1-49.0)	<.001	33.6 (32.8-34.4)	<.001
No	36.9 (36.0-37.9)		36.7 (35.7-37.6)		21.8 (21.1-22.6)

Abbreviations: FPL, federal poverty level; NA, not applicable.

^aAnalyses based on 638 830 observations in the 2018 American Community Survey. When weighted, this sample represented 54 749 082 individuals in the community-dwelling Medicare population.

^bMedicare beneficiaries in households that did not have a desktop or laptop computer with high-speed internet provided via a cable, digital subscriber line, or fiber-optic connection. The eAppendix in the Supplement gives variable definitions.

^cMedicare beneficiaries in households that did not have a smartphone or other mobile device with a data plan for wireless internet service. The eAppendix in the Supplement gives variable definitions.

^dMedicare beneficiaries who not have access at home to either (1) a laptop or internet computer with a high-speed wireline internet connection or (2) a smartphone with a data plan for wireless internet service.

^eP values are for differences between groups of Medicare beneficiaries categorized according to the demographic and socioeconomic variables shown and are adjusted for clustering within public use microdata areas.

^fThe FPL that applied to the individual’s household size and state in 2018.

^gDetails of the assessment of disability status using the American Community Survey are given in the eAppendix in the Supplement.

Discussion

Using data from 2018, we found that 26.3% of Medicare beneficiaries lacked digital access at home, making it unlikely that they could have telemedicine video visits with clinicians. The proportion of beneficiaries who lacked digital access was higher among those with low socioeconomic status, those 85 years or older, and in communities of color. Although Medicare’s payment for audio-only visits at the same rate as video and in-person visits may be associated with improved access to care for those without digital access, the inability to have a video visit may be associated with increased disparities in access to care. Moreover, some Medicare beneficiaries are unable to use technology for video or even audio visits. Limitations of our study include the lack of data in the ACS on beneficiaries’ ability to use technology or community-level broadband internet availability.

During the COVID-19 pandemic, federal telemedicine policy has focused on reimbursement and clinicians’ capacity to deliver care remotely.¹ Our results underscore a need to address disparities in digital access among patients. Expanding programs such as Lifeline, a program of the Federal Communications Commission that provides reduced-cost phone or internet service to families with incomes 135% or more below the federal poverty level,⁵ may help reduce disparities. However, Lifeline does not pay for devices, and patients may also need assistance using technology for video visits. Addressing these factors associated with digital access in populations with low socioeconomic status will be important as the use of telemedicine increases.

Supplement.

eAppendix