Attitudes towards Telemedicine among Urban and Rural Residents

Abstract

Introduction:

Adoption of telemedicine by healthcare facilities has dramatically increased since the start of coronavirus pandemic; yet, major differences exist in universal acceptance of telemedicine across different population groups. The goal of this study was to examine population-based factors associated with current and/or future use of telemedicine in Alabama.

Methods:

A cross-sectional survey was administered to 532 participants online or by phone, in four urban and eight rural counties in Alabama. Data were collected on: demographics, health insurance coverage, medical history, access to technology, and its use in accessing healthcare services. Generalized logit regression models were used to estimate the odds of choosing “virtual visit” and “phone communication” compared to “in-person visit” for the preferred choice of visit with the healthcare provider; as well as odds for willingness to participate in “virtual visit” in the future.

Results:

Our study sample had a mean age of 43 (±15) years, 72.9% women, 45.9% Black or African American; 59.4% population living in an urban county. The odds of “phone communication” were higher compared to the odds of “in-person visit”, with a unit increase in age (odds ratio: 1.02, 95% confidence interval: 1.00–1.03), after adjusting for other covariates. Among participants with past experience of virtual communications, the odds for choosing “virtual visit” were significantly higher compared to choice of in-person visit (odds ratio for virtual visit: 3.23, 95% confidence interval: 2.01–5.18), adjusted for other covariates. Further, people with college or more education were 71% less likely to choose “No” compared to those with high school or lower general education development education for future virtual visit [odds ratio for college or more: 0.29, 95% confodence interval: 0.10–0.87). Likewise, participants residing in rural counties were 57% less likely to choose “No” compared to urban counties for future virtual visit (odds ratio for rural participants: 0.43, 95% confidence interval:0.19–0.97).

Discussion:

Our study found notable differences in age, education, and rurality for use and/or preference for telemedicine. Medical institutions and healthcare providers will need to account for these differences to ensure that the implementation of telemedicine does not exacerbate existing health disparities.

INTRODUCTION

The increased availability of personal technology and Internet connectivity offers healthcare professionals the opportunity to utilize real-time virtual communication to enhance healthcare access for patients with transportation challenges, busy schedules, and/or physical disabilities in any geography.^1–3 Telemedicine has also gained wider acceptance in healthcare systems as studies have found that telemedicine can reduce healthcare spending by decreasing medication misuse, unnecessary emergency department visits, and prolonged hospitalizations.⁴ Telemedicine is defined as “the practice of medicine using electronic communication, information technology, or other means between a physician in one location, and a patient in another location, with or without an intervening health care provider.”^5–8

While currently, more than 60% of all healthcare institutions in the Unted States use telemedicine and the use of telemedicine has steadily increased in the last decade from 35% to 76% across the country, such adoption of telemedicine is not uniform.^9,10 Data prior to 2020 indicate that the interest in using telemedicine was generally high (66%) among patients, but only about 8%–10% of all healthcare encounters were happening through virtual platform; patients between ages 18 and 35 years were three times more likely to have had a video visit with a doctor compared to the other demographics.^9,11 There are several barriers that affect the use of telemedicine. Some of these barriers pertain to policy, including insurance coverage and reimbursement as well as regulatory and legal issues.^4,12 At the provider level barriers may include lack of training, unwillingness or uncertainty of the value of telemedicineto provide adequate care, and initial investment costs.^12,13 At the patient level, digital access and digital literacy is a major obstacle; 97% of Americans in urban areas have access to high-speed fixed service and only 65% and 60% have access to rural areas and tribal lands, respectively.^14,15 Patient characteristics are also critical factors affecting this relationship, whereby access and engagement with telemedicine vary by age, sex, socioeconomic status, and cultural acceptance.^15–18

The coronavirus (COVID-19) pandemic led to dramatic changes in the adoption of telemedicine by healthcare facilities and hospitals.^9,11 Within this context, the federal and state governments implemented public health emergency policy changes to increase telemedicine use to deliver acute, chronic, primary, and specialty care in many settings including inpatient and outpatient.¹⁹ During the pandemic, telemedicine was used as a critical part of the coronavirus response to “forward triage” patients.²⁰ As a result, larger proportion of patients and providers who previously were reluctant to use telemedicine or had no knowledge of telemedicine have used video calls or some other virtual platform for their clinical encounters.^21,22

But before the widespread adoption of telemedicine, it is important to understand if people have the ability and/or are willing to access telemedicine. There is a chance that it may accentuate health disparities if this technology does not reach medically underserved populations. That is, it will reach the ones who already have access but “prefer” online for convenience and leave the ones who “need” it behind. It is also not clear whether the upward trend in telemedicine visits would sustain once the pandemic is over. A recent report indicated that the telemedicine visits decreased by 25% from 35.8% in June 2020 to 26.9% in November 2020.^22,23 The report also found a significantly higher overall average percentage of telemedicine visits in urban centers compared to rural health centers (p < 0.01). The disparity was more pronounced in the Southern US.

Clearly, having access to the technology may not be sufficient for the adoption of telemedicine; people should be willing to use. In this paper, we present the results from a population-based survey in Alabama that assessed access to telemedicine technology in urban and rural populations, history of telemedicine visits, and willingness to use telemedicine in the future. We further examined whether access to telemedicine and subsequently, willingness to use telemedicine differed by sociodemographic characteristics of the participants.

METHODS

Overview

The current study is part of a consortium of 17 Comprehensive Cancer Centers to study the impact of coronavirus on the cancer continuum (Impact of COVID-19 on the Cancer Continuum Consortium—IC-4). In addition to the primary outcomes, the survey also collected information regarding changes in the use of health services, including primary care visits, dental visits, and access to prescription medications. Details about the study setting, recruitment, and data collection tools are described in detail in our prior publication.²⁴

With our site, a total of 12 rural and urban counties were selected and recruitment of participants per county was weighted with respect to geographic regions (urban vs. rural),²⁵ race,²⁶ and percent of population living in poverty.²⁷

Survey Instrument

The survey was developed in collaboration with 4 of the 17 comprehensive cancer centers but each cancer center also included additional survey questions specific to the respective site. All data were captured using Health Insurance Portability and Accountability Act (HIPAA) compliant Research Electronic Data Capture (REDCap) tools.²⁸ The survey was pilot tested to assess for readability levels, appropriateness, and comprehension.

Detailed descriptions of the questionnaire and its variables are discussed in our companion paper.²⁴ In brief, data were collected on demographic variables (e.g. urban/rural residence, age, marital status, educational attainment, income), health insurance coverage, and medical history. In our survey, we questioned the participants about their access to an electronic device with an Internet connection—“Do you have a device (cell phone, laptop, tablet, or desktop with webcam) that would allow you to video conference with your doctor or healthcare provider?” Among the participants who said “yes” to having an electronic device with an Internet connection, we asked about their preferred choice in view of cancelation of in-person visit with their primary care physician (“If you were scheduled for a routine, non-urgent clinic appointment and your primary doctor was not able to see you, which of the following would you prefer? I would prefer to:”. Response options were: “(1) Wait until my doctor is available and reschedule an in-person visit; (2) Reschedule an in-person visit with a different doctor; (3)Talk to my doctor by phone for advice; (4) Send in a photo and message for advice through a secure online portal; (5) Set up a video-visit with my doctor.” For purpose of this paper, participants who chose options (1) or (2) were grouped together and categorized as “prefer in-person visit”. Similarly, participants who chose options (4) or (5) were combined and categorized as “prefer virtual communication”. Participants who chose (3) remained in the same group. Thus, the dependent variable was reframed as “preferred choice of visit with the healthcare provider” with three response categories (“1- prefer in-person visit; 2-prefer to speak by phone; 3-prefer to use virtual mode of communication”).

Among those who had access, we also enquired whether they had any past virtual interaction with their doctor or other healthcare providers (“Have you engaged in a virtual visit with any of your doctor or healthcare providers?”). For those who answered yes, we asked about their perception about the visit (“Did you feel comfortable communicating with your healthcare provider in a virtual format?”). Among those who answered “no”, we assessed their willingness to participate in a virtual visit in the future (“Would you feel comfortable communicating with your healthcare provider in a virtual format?”).

A total of 616 participants completed the survey. Of these, 584 (94.5%) participants had a device with which they could communicate with their doctor virtually. The 34 participants who did not have any device were more likely to be African American (58.8%), female (76.5%), had a high school degree or lower (55.9%), with income <$19,000, and live in a rural county (88%). Since only 5.4% of the participants did not have an electronic device with an Internet connection, we decided to exclude them from the main analyses. We further excluded participants in “other race category” since they constituted a very small percentage (7.2%) of the total sample. Five participants with missing data on the primary outcome variable were excluded. Thus, the study analytic sample was comprised of 532 participants (Figure 1).

Figure 1.

Final number of participants included in e-health analyses

Statistical Analysis

Summary statistics were computed for each of the three responses for the preferred choice of visit with the healthcare provider. Unadjusted comparisons were performed between groups using chi-square tests (or Fisher’s exact tests) for categorical variables and using analysis of variance (ANOVA) for continuous variables.

A generalized logit regression model was used to evaluate the association of demographics, health insurance coverage, and history of past virtual visit, with a preferred choice of visit with the healthcare provider. The estimated odds ratios (OR) were probabilities of choosing “virtual visit” and “phone communication” compared to “in-person visit” for the preferred choice of visit with the healthcare provider. In additional models, we evaluated the association of demographicvariables and health insurance coverage with the willingness to participate in virtual visit in the future. The estimated OR were probabilities of responding “No” and “Don’t know/not sure” compared to “Yes”. We made the decision to include the “Don’t know/Not sure” response as a separate choice because a “don’t know” response may not indicate response ambivalence but rather a lack of knowledge about a topic.²⁹ All statistical analyses were performed by utilizing SAS 9.4.³⁰ Statistical significance was defined at p < 0.05.

RESULTS

Demographics

Table 1 provides the distribution of demographic characteristics by comfort of virtual consultation. The mean (SD) age for the total sample of 532 was 43.0 (±15) years, majority were female (72.9%) and almost 45.9% were Black or African Americans. About 32.1% of the participants had a combined annual income of $75,000 or higher, 48.5% of participants had at least a college degree, 59.4% lived in an urban county, and 89.3% had health insurance coverage. Almost 35% of the participants had previously engaged with their doctor or healthcare provider on a virtual platform.

Table 1.

Distribution of demographic characteristics by comfort of virtual consultation (N=532)^*

			Response†‡
	All		A		B		C
	N	%	N	%	N	%	N	%	P-value
Total	532		232	43.6	132	24.8	168	31.9
Age									0.0070a
Mean (SD)	43 (15)		42(15)		47(15)		42(15)
Median (Min-Max)	44(18–82)		40(18–79)		48(18–82)		40(18–77)
Age categories (in years)									0.0193
18–44	270	51	123	53	53	40	94	56
45–64	219	41	96	41	62	47	61	36
65+	42	7.9	12	5.2	17	12.9	13	7.7
Gender									0.2272b
Male	143	26.9	66	28.4	40	30.3	37	22
Female	388	72.9	166	71.6	92	69.7	130	77.4
Race									0.8200b
White	288	54.1	122	52.6	73	55.3	93	55.4
Black	244	45.9	110	47.4	59	44.7	75	44.6
Education level									0.0566c
High school or lower	97	18.2	47	20.3	31	23.5	19	11.3
Some college	176	33.1	76	32.8	42	31.8	58	34.5
College and plus	258	48.5	109	47	59	44.7	90	53.6
Income									0.4468c
$0 to $19,999	65	12.2	24	10.3	22	16.7	19	11.3
$20,000 to $49,999	114	21.4	47	20.3	26	19.7	41	24.4
$50,000 to $74,999	117	22	51	22	31	23.5	35	20.8
$75,000 and higher	171	32.1	76	32.8	39	29.5	56	33.3
Rurality									0.3624b
Rural	216	40.6	97	41.8	58	43.9	61	36.3
Urban	316	59.4	135	58.2	74	56.1	107	63.7
Health Insurance Coverage									0.3473b
Yes	475	89.3	205	88.4	116	87.9	154	91.7
No	53	10	26	11.2	15	11.4	12	7.1
Engaged in virtual visit in the past									<0.001
Yes	187	35.2	58	25	42	31.8	87	51.8
No	343	64.5	174	75	90	68.2	79	47

^*Excluded those with no electronic device (n=34); other race (n=45); excluded missing values of response variable (n=4);

^†‡Outcome variable computed from responses to the question about preference of consultation with the physician during COVID-19 pandemic. ‘Response A’ = Wait for in-person appointment or see another physician; ‘Response B’ = Talk to the physician on the phone; ‘Response C’= Share or email pictures; ‘Response D’= consult the physician on a virtual platform; regarding their health problem; Row percentages

^aP-value associated with ANOVA;

^bP-value associated with test statistic for general association; P-value associated with extended Mantel-Haenszel mean score statistic

Of the 532 participants, 232 participants (43.6%) preferred to have an in-person doctor’s visit; 132 preferred communication by phone (24.8%), and 168 chose virtual visit (31.9%). Among the 187 participants who had previous experience of engaging with healthcare providers on a virtual platform, 170 people responded (93.1%) that they felt comfortable with their virtual visit; 11 responded “No” (5.9%) and only 2 responded that they “didn’t know/not sure” how they felt about their virtual visit. Almost 50% of those who had prior experience chose virtual visit when asked about their choice of preferred communication with their health care provider. Of the 350 people who never experienced virtual visit (Table 1), 73.6% responded “Yes” to whether they will feel comfortable with a future virtual visit; 12.0% responded “No” and 14.3% responded “Don’t know/not sure”.

In bivariate analyses, age and previous history of virtual visit were significantly associated with preferred choice of visit with the healthcare provider. No notable difference were found in participants’ race, gender, educational attainment, income, urban/rural residence, and health insurance coverage (Table 1). Among those who never experienced virtual visit, only education attainment was significantly associated with willingness to participate in virtual visit in the future (Table 2).

Table 2.

Characteristics of participants comfortable with future virtual visit with healthcare provider (n=349)

	All		Yes		No		Don’t know/not sure		P-value
	N		N	%	N	%	N	%
All	343		253	74	39	11	50	14
Age									0.5089a
Mean (SD)	41.9(16)		41.7(15)		40(16)		43.8(17)
Median (Min-Max)	39.5(17–82)		38(18 −75)		33(21–82)		42.5(19–75)
Age categories (in years)									0.9014b
18–44	191	56	142	56	23	59	26	52
45–64	123	36	92	36	12	30.8	19	38
65+	28	8.2	19	7.5	3	7.7	5	10
Gender									0.5477b
Male	96	28	72	29	8	20.5	15	30
Female	247	72	181	71.5	31	79.5	35	70
Race									0.8151b
White	191	56	138	55	23	59	29	58
Black	152	44	115	46	16	41	21	42
Education level									0.0253c
High school or lower	64	19	39	15	10	26	15	30
Some college	122	36	89	35	15	39	17	34
College and plus	157	46	125	49	14	36	18	36
Income									0.4079c
$0 to $19,999	46	13	32	13	5	13	9	18
$20,000 to $49,999	71	21	52	21	11	28	8	16
$50,000 to $74,999	75	22	57	23	8	21	9	18
$75,000 and higher	109	32	86	34	11	28	12	24
Rurality									0.2973b
Rural	154	45	118	47	13	33.3	22	44
Urban	189	55	135	53	26	67	28	56
Health Insurance Coverage									0.9932b
Yes	299	87	220	87	34	87.2	44	88
No	42	12	31	12	5	12.8	6	12

^*Only participants who had no experience of virtual communication with healthcare provider

Outcome is responses by the participants to question whether they would feel comfortable communicating with your healthcare provider in a virtual format

^aP-value associated with ANOVA;

^bP-value associated with test statistic for general association;

^cP-value associated with extended Mantel-Haenszel mean score statistic

In multivariable models, results of the generalized logit model evaluating relationship between the sociodemographic factors and preference of visit with the healthcare provider found that the odds of “phone communication” were higher compared to the odds of “in-person visit”, with a unit increase in age (OR for preference for phone communication: 1.02, 95% confidence interval (CI): 1.00–1.03) (Tables 2 and 3). Also, among participants with past experience in virtual communications, odds for “virtual visit” were significantly higher compared to in-person visit (OR for virtual visit: 3.23, 95% CI: 2.01–5.18).

Table 3.

Generalized Logit Regression Model of factors associated with preferred choice of visit with the healthcare provider, in relation to those who preferred “in-person visit”.

Variable	By Phone OR (95% CI)	By Video or Photos OR (95% CI)
	OR (95% CI)	OR (95% CI)
Age	1.02 (1.00–1.03)*	0.99 (0.98–1.01)
Gender
Ref: Male
Female	1.23 (0.74–2.06)	0.79 (0.47–1.32)
Race
Ref: Non-Hispanic Black or African American
White	1.31 (0.83–2.07)	1.31 (0.81–2.10)
Education attainment
Ref: High School or lower / GED
Some College	0.69 (0.35–1.37)	1.84 (0.87–3.89)
College/College Plus	0.71 (0.35–1.45)	1.93 (0.88–4.21)
Income
Ref: $0 to $19,000
$20,000 to $49,999	0.65 (0.29–1.46)	0.79 (0.35–1.78)
$50,000 to $74,999	0.66 (0.28–1.54)	0.54 (0.22–1.3)
$75,000 and higher	0.55 (0.23–1.31)	0.56 (0.24–1.34)
Rurality
Ref:Urban
Rural	1.02 (0.6–1.73)	1.03 (0.62–1.71)
Insurance
Ref: No
Yes	1.09 (0.5–2.41)	1.56 (0.67–3.63)
History of past virtual visit with a healthcare provider
Ref: No
Yes	1.29 (0.76–2.17)	3.23 (2.01–5.18)*

CI: confidence interval; OR: odds ratio; GED: genera education degree

^*P-value <0.05

Generalized logit model (Table 4) assessing the relationship between sociodemographic factors and willingness to use telemedicine in the future among participants who never experienced virtual visit found that people with college or more education were 71% less likely to choose “No” compared to high school education or lower, in relation to “Yes” for future virtual visit (OR for college or more: 0.29, 95% CI: 0.10–0.87). Likewise, participants residing in rural counties were 57% less likely to choose “No” compared to those in urban counties in relation to “Yes” for future virtual visit OR for rural participants: 0.43, 95% CI: 0:19–0.97).

Table 4.

Generalized Logit Regression Model of Factors associated with being comfortable with virtual visit with healthcare provider in future among participants who have not a virtual healthcare visit in the past

Variable	No	Don’t know/Not Sure
Age	0.998(0.97–1.02)	1.01(0.99–1.03)
Gender
Ref: Male
Female	1.66 (0.68–4.05)	0.9(0.41–1.96)
Race
Ref: Non-Hispanic Black or African American
White	0.85 (0.4–1.82)	1.23 (0.59–2.55)
Education attainment
Ref: High School or lower / GED
Some College	0.69 (0.25–1.91)	0.55 (0.21–1.43)
College/College Plus	0.29 (0.09–0.91) *	0.38 (0.13–1.11)
Income
Ref: $0 to $19,000
$20,000 to $49,999	1.81 (0.5–6.6)	0.64 (0.21–1.94)
$50,000 to $74,999	1.28 (0.31–5.35)	0.64 (0.2–2.11)
$75,000 and higher	1.25 (0.3–5.23)	0.59 (0.18–1.95)
Region
Ref:Urban
Rural	0.32 (0.13–0.78) *	0.7 (0.31–1.57)
Insurance
Ref: No
Yes	1.28 (0.38–4.31)	1.72 (0.52–5.71)

GED: genera education degree

^*P-value <0.05

DISCUSSION

The obtained results indicated that the majority of participants preferred to seek care in person as compared to aninteraction on phone or a virtual visit. Increasing age was significantly associated with opting for a phone visit, and higher education and residence in rural counties were associated with willingness to participate in future telemedicine services among participants who had not had a virtual healthcare visit in the past.

Use of telemedicine is certainly not a novel concept, particularly in rural areas and other low-resource settings. It has been an alternative solution for those with transportation issues, for hospitals with limited access to specialty care, for the closing of rural practices and hospitals, and for the dwindling rural health care workforce. A recent review of 38 meta-analyses representing about 928 primary studies reported that telemedicine within several specialties such as psychology or psychiatry and endocrinology disciplines can be equivalent or more clinically effective when compared to usual care.³¹ Our study has indicated a positive attitude toward use of telemedicine among those who had past experience with telemedicine. These findings suggest that, moving forward, people who were exposed to telemedicine first time during the pandemic, may continue to use telemedicine in the future, if given a choice. A consumer survey in May 2020 indicated that 76% of their respondents were likely to use telehealth going forward and almost 74% of telehealth users reported high satisfaction.³² These changes may remove barriers to telemedicine and create an important opportunity to rethink its role in the healthcare system.

However, the expansion of telemedicine in healthcare is not without its challenges. The major challenge is the “Digital Divide”. Digital divide refers not only to the difficulty in accessing Internet, but also access to computers, tablets, laptops, and smartphones.³³ In our survey, we observed a high proportion of people with access to digital sevices with which they could interact virtually with their providers. The high prevalence of internet access (95%) could be attributed to the fact that our survey was conducted online or on phone. Compared to the participants in our survey, only 76.8% of the state’s population has access to Internet and about 85.5% has access to a computer or digital device. This still leaves 24.6% of the population without proper access to Internet, which rises to up to 59% in some areas. Majority of these populations are concentrated in rural and semi-urban areas overlapping with medically underserved areas and areas with higher minority populations.

Digital literacy is another challenge. Populations with low digital literacy are less capable of utilizing technological devices, and navigating online platforms to obtain health information.¹⁵ In our study, higher education was significantly associated with higher preference for virtual visit. However, increasing age was associated with higher preference for phone visit but not with preference for virtual visit. This suggests that older adults and individuals with low educational attainment are less likely to engage in online patient portals. It may also point to the lack of confidence and trust in using computer or mobile devices for medical reasons, particularly among older adults.³⁴ Such findings are consistent with reports from other recent studies that have found that patients 65 years and older, African American, Hispanic, Spanish-speaking, and from areas with low broadband Internet connections were less likely to choose video visits.^34–36 While race and gender, were not associated with any preferred choice for seeking care in our study, other studies have reported that racial/ethnic minorities are less likely to engage in telemedicine.³⁵ Both digital divide and digital literacy can further increase existing health disparities.³⁷

To realize the full potential of telemedicine and telehealth, several things need to happen. These primarily include (1) expansion of robust and effective broadband structure in rural America;^36,38,39 providing government subsidies for broadband subscriptions and data plans for mobile health applications to help reduce cost barriers to individual patients; (2) improving access to electronic devices such as computers, smartphones, tablets, and laptops. High prevalence of digital access in our study population clearly indicates the readiness of the rural populations to use electronic devices. Policy changes in insurance coverage related to telemedicine services may address issues of equitable access to telemedicine; and (3) increasing comfort with using devices for telemedicine by ensuring adequate digital literacy, particularly among older, rural, and low-income communities.

While a lot of these solutions may address gaps in technologyand digital inequities, further investigation is needed to better understand potential behavioral drivers to increase utilization among some population groups. In addition, there is a need for regular evaluation of telemedicine programs to ensure equitable distribution of resources and to make sure that vulnerable populations have access and opportunities to take advantage of telemedicine services.

Study strengths and limitations

Our study has a few limitations. It was conducted between August and December 2020 during the peak of the coronavirus pandemic. Due to social distancing measures, we had to rely on social networks, online church groups, virtual social meet-ups to promote recruitment, and individual in-person recruitment at shopping areas. This resulted in a convenience sample who were more likely to have access to Internet and digital devices and possibly left out populations that had limited access to Internet or even phone. Although we facilitated phone-based data collection wherein the research staff administered the survey on the phone uptake for phone surveys was limited. Despite these efforts, it is likely that we missed out on participation from people who preferred in-person data collection. Also, the sample is limited to the population in Alabama and this may limit the generalization of these results only to populations in the southern states in the United States with similar demographic distribution. Another limitation of the study was self-report. However, it is most likely that participants could easily recall if they had previous experience with telemedicine and whether they felt comfortable with their virtual visit or not; thus, limiting information bias. The key strengths of this study are its representation of urban and rural populations with adequate representation by race and gender and external validity to compare results with other comparable survey populations from within the consortium.

Conclusion

In our study, we noted differences in age, education, and rurality to use and/or preference for telemedicine. There is a lot of excitement among the research communities to bring telemedicine into mainstream healthcare system as a result of the coronavirus pandemic. However, medical institutions and healthcare providers will need to take into account digital inequalities in the target populations to make sure that widespread implementation of telemedicine does not exacerbate existing health disparities.