Telehealth utilization among multi-ethnic patients with obesity during the COVID-19 pandemic

Abstract

Introduction:

The science of telemedicine has shown great advances over the past decade. However, the field needs to better understand if a change in care delivery from in-person to telehealth as a result of the COVID-19 pandemic will yield durable patient engagement and health outcomes for patients with obesity. The objective of this study was to examine the association of mode of healthcare utilization (telehealth versus in-person) and sociodemographic factors among patients with obesity during the COVID-19 pandemic.

Methods:

A retrospective medical chart review identified patients with obesity from a university outpatient obesity medicine clinic and a community bariatric surgery practice. Patients completed an online survey (1 June 2020–24 September 2020) to assess changes in healthcare utilization modality during subsequent changes in infection rates in the geographic area. Logistic regression analysis examined the association of mode of healthcare utilization and key sociodemographic characteristics.

Results:

A total of 583 patients (87% female, mean age 51.2 years (standard deviation 13.0), mean body mass index 40.2 (standard deviation 6.7), 49.2% non-Hispanic white, 28.7% non-Hispanic black, 16.4% Hispanic, 7% other ethnicity, 33.1% completed bariatric surgery) were included. Adjusted logistic regression models showed older age was inversely associated with telehealth use (adjusted odds ratio = 0.58, 95% confidence interval 0.34–0.98) and non-Hispanic black were more likely to use telehealth compared to non-Hispanic white (adjusted odds ratio = 1.72, 95% confidence interval 1.05–2.81).

Conclusions:

The COVID-19 pandemic is impacting access to healthcare among patients with obesity. Telehealth is an emerging modality that can maintain healthcare access during the pandemic, but utilization varies by age and ethnicity in this high-risk population.

Introduction

Severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) is responsible for the novel coronavirus disease 2019 (COVID-19) pandemic that has resulted in over 42 m cases and 1 m deaths worldwide.¹ COVID-19 morbidity and mortality statistics have disproportionately affected vulnerable populations, which include those with pre-existing medical conditions like obesity and severe obesity. Worldwide data have clearly shown that those with obesity, and those with severe obesity complicated by adiposity-associated comorbidities in particular, are at greater risk for COVID-19-related morbidity and mortality versus those of healthy weight.^2–4 Specifically, those with obesity have an elevated risk of hospitalization, serious illness and mortality.⁴ Moreover, studies have shown that non-Hispanic Black (NHB) and Hispanic populations, who tend to have disproportionately greater rates of obesity and related chronic conditions, are more likely to die from COVID-19.^5–7 The social orders and fear of infection during the COVID-19 pandemic have added greater complexity to chronic-condition management, obesity medicine access, and utilization of healthcare providers and services. As such, it is expected that a significant portion of obesity care has moved to a telehealth model since March 2020.

We reported earlier this year in a phase I study that patients with obesity were experiencing significant physical and mental health impacts from COVID-19 stay-at-home orders.⁸ Specifically, 72.8% reported increased anxiety and 83.6% increased depression. In addition, 69.6% reported more difficulty in achieving weight loss goals, which may be associated with an exacerbation of obesogenic behaviours as 47.9% reported less exercise time and 55.8% reported less exercise intensity, 49.6% reported increased stockpiling of food, 61.2% reported stress eating and 61.2% reported following healthy diet plans more challenging during stay-at-home orders. These results present unique research opportunities for obesity care settings (primary care, obesity medicine specialists, and metabolic and bariatric surgery (MBS) programmes) as healthcare systems focus on reengaging patients in their care, risk mitigation and building on the popularity that telehealth now has for healthcare delivery for the duration of the pandemic and beyond. The science of telemedicine has shown great advances over the past decade.⁹ However, the field needs to better understand if a change in care delivery from in-person to telehealth will yield durable patient engagement and health outcomes for patients with obesity, and how to optimise these factors for all people with obesity. Therefore, the purpose of this phase II study was to examine the demographic predictors of engagement in telemedicine among patients with obesity during the pandemic. It was hypothesised that patients with obesity would engage in telehealth-delivered care more than in-person care due to fear of becoming infected if they left their homes.

Methods

Design

A retrospective medical chart review identified patients with obesity from an academic healthcare system’s clinic specialising in obesity medicine and a community-based MBS practice. All patients were receiving healthcare for their obesity diagnosis and included those who were preparing for and those who had completed MBS.

Procedure

An online (non-anonymous) survey was utilised to obtain information about the COVID-19 pandemic’s impact on patients with obesity starting 1 June 2020. This was approximately one month after the Governor of Texas reversed mandated stay-at-home orders, which ended on 30 April 2020. The University of Texas Health System Institutional Review Board approved the study. Patients were asked to respond to a 15-minute survey about their experiences during the COVID-19 pandemic as it pertains to their health, lifestyle and healthcare behaviours. Those that agreed to participate signed an online consent and authorised researchers to contact them for follow-up information. Study data were collected and managed using Research Electronic Data Capture (REDCap) electronic data capture tools hosted at the UT Southwestern Medical Center. REDCap is a secure, Health Insurance Portability and Accountability Act of 1996 (HIPAA) compliant, Web-based software platform designed to support data capture for research studies, providing (a) an intuitive interface for validated data capture; (b) audit trails for tracking data manipulation and export procedures; (c) automated export procedures for seamless data downloads to common statistical packages; and (d) procedures for data integration and interoperability with external sources.^10,11

Measures

Survey respondents were queried on a variety of areas including demographics, COVID-19 infection and health care utilization.

Demographics

Demographic questions were adapted from the validated Behavioral Risk Factor Surveillance System (BRFSS).¹² Respondents were asked their gender, race/ethnicity, age, marital status and basic anthropometrics (height/weight). Household information was collected, including a breakdown of the number of adults (older than 18 years), teenagers (ages 13–18 years), children (ages 5–12 years), and infants/toddlers (ages 0–4 years) living at home. Socioeconomic factors such as annual household income and highest level of education were asked for.

COVID-19

COVID-19 related questions focused on whether participants or their family members were infected with the virus (‘Have any of your family members tested positive for COVID-19?’). Whether infected or not, participants were asked if they were tested (‘Have you been tested for COVID-19?’), had experienced difficulty in receiving a test if desired (‘Have you wanted to get tested for COVID-19 but found it difficult to do so?’) or had symptoms (‘Have you had any symptoms associated with COVID-19?’). Follow-up questions required respondents to select which symptoms (fever, cough, shortness of breath or difficult breathing, tiredness, aches, runny nose, sore throat, loss of smell, loss of taste, rash, nausea/vomiting, diarrhoea, other) they had and the level of severity (no symptoms/asymptomatic, very mild, moderate, severe or very severe). All participants were asked if they had any existing chronic medical conditions (active cancer treatment, asthma/other pulmonary disease, autoimmune disease, diabetes, heart disease, high blood pressure, high cholesterol/hyperlipidaemia, kidney disease, metabolic syndrome, osteoarthritis, polycystic ovarian syndrome, sleep apnoea, other), which would put them at greater risk for COVID-19-related complications beyond their obesity.

Telehealth utilization

Telehealth utilization was our primary dependent variable of interest. Patients were asked to answer, ‘During the STAY-AT-HOME order, how did you see or contact your health care provider?’. Participants were categorised into the telehealth group if they reported using telehealth video, telephone or communicated via an online portal. Other patients who met providers in-person were included in the non-telehealth (in-person) group.

Statistical analysis

Pearson chi-square (n > 5) or Fisher’s exact test (n ≤ 5) was used to compare categorical variables such as age group (<45 years vs 45–64 years vs ≥65 years), sex, race/ethnicity (non-Hispanic white (NHW), NHB, Hispanic and other), education, household incomes, body mass index (BMI) categories (<30 kg/m² vs 30–39.9 kg/m² vs ≥40 kg/m²), insurance coverage, MBS completion and chronic medical conditions by telehealth utilization (yes or no (Y/N)). Age and BMI were also analysed as a continuous variables via two-sample t-test. COVID-19-related information including test results (positive or negative), appearance of symptoms (Y/N) and level of quarantine was also compared by telehealth utilization. In addition, we compared healthcare utilization since COVID-19 by different ethnic groups. Stepwise logistic regression was performed to explore potential significant predictors for telehealth utilization (Y/N). Specifically, a crude logistic regression was run for each potential predictor listed above, and only significant associations with telehealth utilization were selected as independent variables in the adjusted multivariable model. We employed this stepwise method to minimise the risk of multicollinearity, or when two or more independent variables in a multiple regression model were highly correlated. Multicollinearity can cause inaccurate parameter estimates, biased hypothesis tests and, consequently, incorrect inferences about relationships between exposure and outcome variables.¹³ All statistical analyses were performed using SAS v9.4 (SAS Institute, Cary, North Carolina, USA). A two-sided p-value <0.05 was considered significant.

Results

From 1 June 2020 – 28 September 2020, a total of 787 patients have enrolled in this COVID-19 and obesity phase II study. We excluded those who did not complete the patient consent form (n = 7) or did not answer questions regarding telehealth utilization (n = 108). From these 672 respondents, a further 89 patients were excluded either because they did not see or contact healthcare providers or used other methods for healthcare. Thus, the final analytical sample included 583 patients (84.2% female, mean age 53.5 years, standard deviation (SD) 13.2), among which 408 (70%) reported they used telehealth, and 175 (30%) reported having in-person visits. The mean age of telehealth utilisers was significantly less than that of in-person visitors (52.3 (SD 12.9) vs 56.3 (SD 13.5), p<0.001) A total of 16 patients (2.7%) reported they had tested positive for SARS-CoV-2 and the majority (n = 14) had utilised telehealth for contacting providers. However, more patients (12.6%) reported having COVID-19 symptoms and, of those, 13.6% reported telehealth service utilization while 10.6% reported seeing their providers in-person. More than a half (54.2%) were NHW, 22.7% were NHB, 18.7% were Hispanic and 4.4% identified as ‘other’ (multiracial, Asian, etc.). The majority (55.9%) were college graduates, 71.8% were covered by private insurance and about half of the sample (49.5%) had an annual household income > US$75,000. Mean BMI was 35.3 kg/m² (SD 9.3) and 24.8% had completed MBS. Self-reported medical conditions were highly prevalent and included high blood pressure (43.2%), hyperlipidaemia (34.6%), sleep apnoea (29.5%), diabetes (29.3%), asthma (23.7%), heart disease (20.2%) and active cancer treatment (3.4%) (Table 1). Most (72.2%) patients reported only leaving their homes for necessities, followed by 41.9% who went outside for walks or exercise. Notably, more than half of those in the in-person group reported going outside for exercise compared with only 37.5% of patients from the telehealth group (p = 0.001).

Table 1.

Patient characteristics by telehealth utilization status since COVID-19, COVID-19 and Obesity Study Phase 2 (n = 583).

	Total (n = 583)	Telehealtha (n = 408)	In-person (n = 175)	p-Valueb
Age, years, mean (SD)c	53.5 (13.2)	52.3 (12.9)	56.3 (13.5)	<0.001
<45, n (%)	158 (27.2)	117 (28.8)	41 (23.4)	0.004
45–64, n (%)	277 (47.7)	203 (50.0)	74 (42.3)
≥65, n (%)	146 (25.1)	86 (21.2)	60 (34.3)
Sex, n (%)d				0.385
Male	91 (15.8)	67 (16.7)	24 (13.8)
Female	485 (84.2)	335 (83.3)	150 (86.2)
Race, n (%)e				0.072
Non-Hispanic White	308 (54.2)	203 (52.1)	99 (60.7)
Non-Hispanic Black	129 (22.7)	97 (24.9)	26 (16.0)
Hispanic	106 (18.7)	73 (18.7)	30 (18.4)
Other	25 (4.4)	17(4.3)	8 (4.9)
Education, n (%)f				0.807
High school or below	54 (9.3)	40 (9.9)	14 (8.0)
Some college or technical school	202 (34.8)	141 (34.8)	61 (34.9)
College graduate	324 (55.9)	224 (55.3)	100 (57.1)
Annual household income, n (%)g				0.152
<US$25,000	67 (11.7)	54 (13.5)	13 (7.5)
US$25,000–49,999	104 (18.1)	73 (18.2)	31 (17.9)
US$50,000–74,999	119 (20.7)	85 (21.2)	34 (19.7)
≥US$75,000	284 (49.5)	189 (47.1)	95 (54.9)
BMI, mean (SD)h	35.3 (9.3)	35.4 (9.6)	35.2 (8.4)	0.788
<30, n (%)	176 (30.5)	123 (30.5)	53 (30.5)	0.977
30–39.9, n (%)	276 (47.7)	192 (47.5)	84 (48.3)
≥40, n (%)	126 (21.8)	89 (22.0)	37 (21.3)
Insurance coverage, n (%)i				0.173
Medicare	155 (27.7)	100 (25.7)	55 (32.2)
Medicaid	3 (0.5)	3 (0.8)	0 (0)
Private insurance	402 (71.8)	286 (73.5)	116 (67.8)
Completed MBS, n (%)d				0.799
Yes	143 (24.8)	102 (25.1)	41 (24.1)
No	433 (75.2)	304 (74.9)	129 (75.9)
COVID symptoms, n (%)f				0.272
Yes	73 (12.6)	55 (13.6)	18 (10.3)
No	507 (87.4)	350 (86.4)	157 (89.7)
Test positive for COVID, n (%)c				0.168
Yes	16 (2.7)	14 (3.4)	2 (1.2)
No	565 (97.3)	393 (96.6)	172 (98.8)
Medical conditions, n (%)
Active cancer treatment	20 (3.4)	15(3.7)	5 (2.9)	0.618
Asthma/other pulmonary disease	118 (20.2)	75 (18.4)	43 (24.6)	0.088
Diabetes	171 (29.3)	119 (29.2)	52 (29.7)	0.894
Heart disease	55 (9.4)	33 (8.1)	22 (12.6)	0.089
High blood pressure	252 (43.2)	171 (41.9)	81 (46.3)	0.329
High cholesterol/hyperlipidaemia	202 (34.6)	131 (32.1)	71 (40.6)	0.049
Sleep apnoea	172 (29.5)	122 (29.9)	50 (28.6)	0.747
Level of quarantine, n (%)
Not going outside at all	22 (3.8)	19 (4.7)	3 (1.7)	0.088
Going outside for walks or exercise	244 (41.9)	153 (37.5)	91 (52.0)	0.001
Going outside for necessities (food, medications)	421 (72.2)	291 (71.3)	130 (74.3)	0.464
Visiting close family/friends	150 (25.7)	96 (23.5)	54 (30.9)	0.064
Going to work	68 (11.7)	50 (12.3)	18 (10.3)	0.497
Attending religious services	17 (2.9)	12 (2.9)	5 (2.9)	0.956
Attending parties/large social functions	2 (0.3)	1 (0.6)	1 (0.3)	0.511
Going out as normal	2 (0.3)	1 (0.6)	1 (0.3)	0.511

BMI: body mass index; MBS: metabolic and bariatric surgery; SD: standard deviation.

^aInclude telehealth video, telephone, and communicate via online portal

^bChi-square or fisher’s exact test for categorical variables; two sample t-test for continuous variable: age and BMI

^cmissing = 2

^dmissing = 7

^emissing = 15

^fmissing = 3

^gmissing = 9

^hmissing = 5

ⁱmissing = 23.

The majority (53.9%) of the sample reported no difference in terms of the frequency of seeing healthcare providers since COVID-19, while 42.9% reported less often and only a small proportion (3.2%) reported more often. Almost eight in 10 respondents in the telehealth group reported that this was their first time using telehealth. Major reasons for not seeing their healthcare providers included: (a) appointment cancelled (31.7%); (b) fear of infection (26.9%); and (c) office closed (21.6%) (Table 2). Interestingly, fewer NHB (25.6%) and Hispanics (24.5%) had a cancelled appointment compared to NHW (36.0%) (p = 0.042).

Table 2.

Health care utilization by ethnic groups, COVID-19 and Obesity Study Phase 2.

	Total (n = 568)	NHW (n = 308)	NHB (n = 129)	Hispanic (n = 106)	Other (n = 25)	p-Valuea
Seeing or contacting health care						0.451
provider since COVID-19, n (%)b
Less often	242 (42.9)	140 (45.6)	20 (39.7)	45 (42.4)	7 (28.0)
No difference	304 (53.9)	155 (50.5)	73 (57.9)	59 (55.7)	17 (68.0)
More often	18 (3.2)	12 (3.9)	3 (2.4)	2 (1.9)	1 (4.0)
First time use telehealth, n (%)						0.945
Yes	319 (77.2)	176 (78.2)	75 (75.8)	56 (75.7)	12 (80.0)
No	94 (22.8)	49 (21.8)	24 (24.2)	18 (24.3)	3 (20.0)
Avoided seeing or contacting health
care since COVID-19, n (%)
Fear of infection	153 (26.9)	91 (29.5)	30 (23.3)	26 (24.5)	6 (24.0)	0.498
Appointment cancelled	180 (31.7)	111 (36.0)	33 (25.6)	26 (24.5)	10 (40.0)	0.042
Office closed	123 (21.6)	78 (25.3)	19 (14.7)	20 (18.9)	6 (24.0)	0.084
Financial costs/deductible	26 (4.6)	10 (3.3)	11 (8.5)	5 (4.7)	0 (0)	0.096
Loss of health insurance	7 (1.2)	1 (0.3)	3 (2.3)	3 (2.8)	0 (0)	0.117
Busy/lack of time	21 (3.7)	13(4.2)	4(3.1)	2(1.9)	2 (8.0)	0.450

^aPerson Chi-square or Fisher’s exact test to the comparison of NHW vs NHB vs Hispanic vs other

^bmissing = 4.

Table 3 shows the crude and adjusted odd ratios (aORs) for telehealth utilization. In the crude logistic model, patients older than 65 years (odds ratio (OR) = 0.50, 95% confidence interval (CI) 0.31–0.82) or had hyperlipidaemia (OR = 0.69, 95% CI 0.48–0.99) were significantly less likely to use telehealth compared to those <45 years old or did not have hyperlipidaemia. Conversely, the odds of telehealth utilization among NHB were almost twice that of NHW (OR = 1.9, 95% CI 1.17–3.09). After controlling for significant predictors identified from crude models, the final adjusted model indicated that older age was inversely associated with telehealth (aOR = 0.58, 95% CI 0.34–0.98) while NHB were more likely to use telehealth compared to NHW (aOR = 1.72, 95% 1.05–2.81). In comparison with participants who reported seeing or contacting health care provider more often or no difference since COVID-19, those who seeing less often had two times higher odds of using telehealth services (aOR = 2.11, 95% CI 1.43–3.13).

Table 3.

Crude and adjusted odds ratio for telehealth utilization.

Variables	Crude odds (95% CI)a	p-Valuea	Adjusted odds (95% CI)a	p-Valueb
BMI
<30	1.0 (ref)	–	–	–
30–39.9	0.99 (0.65–1.49)	0.857	–	–
>40	1.04 (0.63–1.71)	0.846	–	–
Age (years)
<45	1.0 (ref)	–	1.0 (ref)	–
45–64	0.96 (0.62–1.50)	0.097	1.07 (0.65–1.63)	0.056
≥65	0.50 (0.31–0.82)	0.001	0.55 (0.32–0.95)	0.0053
Sex
Male	1.0 (ref)	–	–	–
Female	0.80 (0.48–1.33)	0.972	–	–
Race
Non-Hispanic White	1.0 (ref)	–	1.0 (ref)	–
Non-Hispanic Black	1.90 (1.17–3.09)	0.048	1.79 (1.09–2.96)	0.046
Hispanic	1.27 (0.78–2.07)	0.983	1.11 (0.67–1.85)	0.727
Education
High school or below	1.28 (0.66–2.45)	0.484	–	–
Some college or technical school	1.03 (0.70–1.51)	0.690	–
College graduate	1.0 (ref)	–	–	–
Annual household income
<US$75,000	1.0 (ref)	–	–	–
≥US$75,000	0.73 (0.51–1.05)	0.088	–	–
Insurance coverage
Medicaid or Medicare	0.76 (0.51–1.12)	0.169	–	–
Private insurance	1.0 (ref)	–	–	–
MBS status
Completed	1.06 (0.70–1.60)	0.799	–	–
Not completed	1.0 (ref)	–	–	–
COVID symptoms
Yes	1.37 (0.78–2.41)	0.274	–	–
No	1.0 (ref)	–	–	–
Hyperlipidaemia
Yes	0.69 (0.48–0.99)	0.049	0.74 (0.49–1.11)	0.141
No	1.0 (ref)	–	1.0 (ref)	–
Other medical conditionsc
Yes	0.86 (0.58–1.28)	0.444	–	–
No	1.0 (ref)	–	–	–
Cancelled appointment
Yes	0.97 (0.66–1.41)	0.867	–	–
No	1.0 (ref)	–	–	–
Seeing or contacting health care
provider since COVID-19
Less often	1.79 (1.23–2.60)	0.002	2.11 (1.43–3.13)	<0.001
No difference or more often	1.0 (ref)	–	1.0 (ref)	–

BMI: body mass index; CI: confidence interval; MBS: metabolic and bariatric surgery; SD: standard deviation.

^aCrude logistic regression.

^bCovariates which were significantly associated with telehealth utilization in univariable logistic regression were further examined by multivariable logistic regression.

^cOther medical conditions include active cancer treatment, asthma/other pulmonary disease, diabetes, heart disease, high blood pressure, sleep apnoea.

During the COVID-19 pandemic, a majority of those who engaged in telehealth reported either no difference in the frequency of seeing doctors (49.5%) or less often (47.5%) while more than half (63.4%) of in-person patient visits reported no difference or occurring less often (p = 0.006). Ethnic groups differences were significant for NHBs (less often 42.5%, no difference 52.5%, more often 2.0% among telehealth users versus less often 18.5%, no difference 77.8%, more often 3.7% among in-person visitors, p = 0.025), but not for other ethnic groups (Figure 1). Additional logistic regression modelling showed that, compared with those who reported ‘no difference’ in seeing providers, patients who reported seeing providers ‘less often’ had a significantly higher odds of using telehealth (OR = 1.84, 95% CI 1.26–2.67). And this effect was even stronger among NHBs with an OR of 3.63 (95% CI 1.27–10.4).

Figure 1.

Degree of engagement with telehealth delivery of care post-COVID-19 lockdown orders (June– September 2020) by ethnic group. Pearson chi-square test or Fishers exact test to compare telehealth vs in-person visitors.

Discussion

The COVID-19 pandemic is an unprecedented situation for healthcare. The epidemic has also disproportionately affected people with obesity, related comorbidities, advanced age and ethnic minorities, with greater risks of severe COVID-19 and mortality.^2,3,5 During the pandemic, telehealth can provide access to vital healthcare services when there is cancellation of in-person appointments, clinic closures and healthcare provider absence or quarantine. Additionally, using telehealth instead of in-person visits can help to limit exposure of patients, healthcare providers, and other essential personnel to potential COVID-19 infection. Our findings highlight several important trends in telehealth utilization and obesity care access during the COVID-19 pandemic, which can be used to optimise telehealth utilization and patient engagement in these high-risk groups. Notably, telehealth utilization was more likely in younger patients and NHB patients, with no significant associations with BMI, sex, education, household income, insurance coverage or MBS status.

Prior to COVID-19, telehealth was only utilised by 8% of Americans in 2019,¹³ despite being shown to produce similar health outcomes, reduce patient travel costs and time away from work compared to in-person visits.¹⁴ This is consistent with our findings that 77% of our subjects utilised telehealth for the first time during the pandemic. Other healthcare systems have had similar experiences in the dynamic adoption of telehealth during the pandemic, with telehealth accounting for >70% of visits within 10 days of expanded telehealth access at one large institution.¹⁵ Previously, barriers to widespread telehealth utilization included limited reimbursement, lack of familiarity with telehealth technologies and little motivation for replacement of in-person care with telehealth (except in rural areas).¹⁵ Currently, a Centers for Medicare and Medicaid Services (CMS) 1135 waiver has permitted telehealth use in broader circumstances during the pandemic.¹⁶ Telehealth has already been used successfully to medically manage patients with obesity in rural populations, nursing home patients and veterans,^17–19 and its wider application beyond the pandemic will no doubt be useful to the growing population of patients with overweight and obesity. Indeed, others have called attention to the sustainment of telehealth services for healthcare in a post-pandemic world.^20,21

Despite the utilization of telehealth to maintain healthcare access during the pandemic, 42.9% of subjects reported seeing their healthcare providers less often. From our sample, 31.7% of patients were affected by cancelled appointments and 26.1% by clinic closures, which were factors beyond their control. Similar levels of people with type-1 diabetes mellitus (~30%), who are also at increased risk for COVID-19, reported seeing providers less often during the pandemic due to cancelled appointments.²² Additional factors that likely contributed to challenges with healthcare provider access include: social orders restricting non-essential visits, physical limitations for social distancing within medical offices, and infrastructure challenges with pivoting to telehealth at the start of the pandemic.²³

Notably, 26.9% of our subjects saw their providers less often due to fear of infection with COVID-19. Other groups have reported significant decreases in patients presenting for acute and preventive, non-emergent, healthcare services during the pandemic.^24,25 This over-abundance of caution by patients was not the intention of public health education campaigns that highlighted groups of at-risk individuals, who are heavily represented in our cohort, including people with obesity, diabetes, hypertension, pulmonary disease, ethnic minorities and individuals over age 65 years.²⁶

People in groups who are known to be high-risk for COVID-19 would certainly benefit from telehealth to limit potential exposures and to maintain access to healthcare providers. In spite of this, our subjects over the age of 65 years were half as likely to use telehealth compared to those under the age of 45 years. Although several studies show that the majority of people over age 65 years view telehealth positively²² and are interested in completing telehealth visits, one study reported that 64% did not have access to the necessary technology and 32% did not feel confident in utilising telehealth.²⁷ This group was able to address these barriers through individualised training and allow for successful completion of telehealth visits.²⁷ According to data from the Centers for Disease Control, people aged 65–74 years have a 90-times higher risk of death due to COVID-19 compared to those aged 18–29 years, which makes older age groups a prime target for telehealth and other risk-mitigation strategies.²⁸ It is possible that our cohort of patients over the age of 65 years may have similar barriers to adopting telehealth, which should be studied further and addressed.

Despite this increased likelihood of NHB patients using telehealth in our study, 42.5% still reported seeing or contacting healthcare providers less often, despite NHBs reporting fewer cancelled appointments. Understanding why NHB patients were more likely to utilise telehealth, and how they were able to do so, could facilitate the creation of telehealth programmes to meet the needs of this and other minority populations beyond the horizon of the current pandemic. There are already recommendations that at-risk groups should be prioritised for COVID-19 vaccines when they become available.²⁹ Potentially, improved healthcare access through telehealth could optimise engagement and health outcomes for minority populations. This is critical as ethnic minorities, especially NHBs and Hispanics, are disproportionately affected by obesity²⁷ and the COVID-19 pandemic has exacerbated these healthcare disparities.²⁸ These inequalities are driven by conditions and non-medical factors that influence access to healthcare services, appropriate foods, conducive physical environments, education, social supports and financial health.^28,29 During the pandemic, NHB and other minority groups were found to have higher rates of infection,³⁰ with increased morbidity and mortality compared to NHWs.^30,31 Specifically, one study reported that NHBs have a 3.4-fold higher mortality rate due to COVID-19 compared to NHWs.³²

Moreover, a large academic institution in New York City reported that, during the pandemic, NHB patients were 40% less likely to use telehealth compared to NHW patients.³³ This contrasts sharply with our findings that NHB patients were 72% more likely to use telehealth than NHW patients. Publication of disparities in outcomes for this ethnic group may have increased utilization of telehealth over in-person visits. This difference in utilization of telehealth by our NHB patients compared to other reports in the literature^34–36 may be partially explained by our patient demographics in that 55.9% of were college graduates, 71.8% had private insurance and half of patients had an annual household income >US$75,000.

Study limitations and strengths

There are several limitations to this study that should be mentioned. First, this was a sample of convenience which can produce selection bias, especially given this was an online survey and would thus limit participants to those who were willing and able to access the survey. Respondents were enrolled from an academic medical centre’s weight management programme and were primarily white women with an average age of 51 years, mean BMI of 35.3 kg/m², and most had a college education and an annual income of ≥US$75,000. This study may not be generalisable to a clinic, or general population with healthy weight. It may also not be generalisable to other populations and as a result may not accurately assess the burden of COVID-19 on obesity-related health and behaviours in lower socioeconomic status. Participants were established weight management patients with secured health insurance, which is not representative of the group of average Americans challenged with obesity in which <2% receive anti-obesity medication^37,38 and <1% undergo MBS.^38,39 Strengths of the study include the capture of important information to inform comprehensive, future healthcare for patients with obesity. Certainly, healthcare access should be maintained in some way during these periods moving forward, especially if the pandemic should cause future stay-at-home orders as there are implications for long-term health as shown from this stay-at-home situation.

Conclusions

Results from this study show that telehealth is an emerging modality that can help maintain healthcare access during the COVID-19 pandemic, but utilization varies by age and ethnicity among patients with obesity. Specifically, utilization of telehealth during the COVID-19 pandemic was not influenced by sex, BMI, bariatric surgery status, education, income or type of insurance in this high-risk patient population. Findings here highlight age and ethnic differences in telehealth adoption that need to be explored further to improve telehealth and healthcare access for high-risk groups during the pandemic and beyond.