IntroductionBackground

In the United States, telemedicine use has risen over the years. According to the 2019 American Hospital Association annual survey, the percentage of telehealth programs implemented across hospitals increased from 43.1% in 2015 to 61.2% in 2017 []. Since the COVID-19 pandemic began, we have seen rapid and unprecedented growth in the demand for, and use of, telemedicine. A recent report from the Centers for Disease Control and Prevention has documented that the frequency of telehealth visits increased by 50% from 2019 to 2020 []. The increase in the use of telemedicine is also observed in populations of patients with cancer; for example, several studies conducted during 2020 and 2021 estimated that the prevalence of telemedicine use ranges from 34.9% to 64.9% among patients with breast or other cancers [-]. In 2020, the Centers for Medicare & Medicaid Services introduced policies that offered regulatory waivers and flexible reimbursement to Medicaid and Medicare providers for telehealth, contributing in part to the observed increase in telemedicine use and implementation [,]. In 2021, the American Society of Clinical Oncology performed a systematic literature review on telemedicine and published standards and recommendations for telehealth services and practices in the oncology setting []. Telemedicine helps facilitate access to health care and services for patients with cancer and their caregivers or family members. However, telehealth care and services are not equally distributed, and not all patients with cancer have equal access to telehealth care and services across different US regions. There are notable gaps in existing literature regarding the influence of neighborhood-level socioeconomic status (SES) on telemedicine use in patients with breast cancer and oncology services offered through telehealth versus in-person visits.

Neighborhood-level SES is a fundamental component of the social determinants of health framework [,]. Neighborhood socioeconomic disadvantage has been shown to negatively affect health outcomes [,], access to care and preventive services [,], survival outcomes [,], and quality of life [] among patients with cancer []. Previous investigations have also found neighborhood socioeconomic disadvantage to be associated with a lower likelihood of telemedicine use among patients in primary care and hematology or oncology clinics, as well as among outpatients [,-]. A study of 627 patients with cancer experiencing financial distress during the COVID-19 pandemic reported a 3% decrease in the rate of telemedicine use per 10-unit increase in the Area Deprivation Index (ADI) [], a validated composite measure of neighborhood-level SES [,]. Fassas et al [] conducted a univariate analysis of 64 patients with head and neck cancer, revealing no significant differential interest in telehealth visits based on the ADI. Another study noted a higher percentage of telehealth visits among patients residing in the least socioeconomically deprived neighborhoods (54%) than those in the most deprived neighborhoods (46.1%) in a large cohort of patients with hematologic malignancies and patients with cancer from Kaiser Permanente []. These prior studies either lacked significant sample sizes or included heterogeneous populations of patients with cancer. Therefore, these findings may not be generalizable to the population of patients with breast cancer.

In addition, whether provisions or receipts of oncology services differ between telemedicine and in-person office visits among patients with breast cancer is unclear. A recent retrospective analysis of 311 patients with cancer indicated that clinical practices, such as molecular test ordering and palliative care referrals, conducted through telehealth visits achieve similar efficiency to in-person visits []. A pilot study of 45 patients with advanced cancer in Mexico has suggested the feasibility of supportive care delivery via telemedicine []. Multiple studies have found telehealth provisions or visits to be feasible, effective, and safe for patient follow-ups after ambulatory or breast surgeries [-]. Earlier research has also demonstrated that, when comparing telehealth to in-person visits, patients with cancer in the United States, Canada, and Europe reported similar communication experiences or satisfaction with the oncology care consultations they received [,,]. Moreover, telemedicine-based cancer genetic counseling has been shown to be feasible and effective and to achieve high degrees of satisfaction among providers as well as patients with colorectal, breast, or gynecologic cancer residing in remote or rural areas [-]. Although prior studies have elucidated the successful implementation of telemedicine and shown that certain types of cancer care and services delivered through telemedicine are equivalent to in-person office settings in mixed cohorts of patients with cancer, the results may not be applicable to patients with breast cancer. Furthermore, most of these studies were not able to examine the distributions of oncology services comparing telehealth and in-person visits because of small sample sizes and therefore are primarily descriptive. Understanding these associations can help oncology programs identify telehealth disparities and patient’s unmet needs, improve telemedicine practice and care delivery, reduce health disparities and inequities, and provide optimal support to patients with breast cancer.

Objectives

To fill these critical gaps in the literature, we undertook this study primarily seeking to evaluate (1) the association between neighborhood socioeconomic disadvantage and telemedicine use and (2) the differences in provisions of oncology services comparing telehealth and in-person office visits. The secondary objectives of this study were to describe (1) common perceived challenges or concerns related to telehealth visits and (2) patient satisfaction with oncology services delivered via telemedicine in this cohort of patients with breast cancer.

MethodsStudy Design and Population

This study used a cross-sectional design and analyzed data from patients with breast cancer enrolled in the ongoing Chicago Multiethnic Epidemiologic Breast Cancer Cohort (ChiMEC), which is a hospital-based cohort established at the University of Chicago Medicine in 1993 []. From July to September 2022, a total of 1868 questionnaires were sent to ChiMEC participants who consented to follow-up surveys, of whom 1236 (66.17%) responded. The study survey is provided in . For this analysis, of the 1236 respondents, we included 1163 (94.09%) patients who reported having had either telemedicine or in-person visits in the past 12 months.

Ethical Considerations

The University of Chicago Institutional Review Board reviewed and approved this study (approval 16352A). All participants provided written informed consent before taking part in the ChiMEC study and follow-up surveys.

Measures

Neighborhood socioeconomic disadvantage was defined by the ADI, a composite measurement of patients’ neighborhood-level income, education, employment, and housing quality based on linked zip codes and US Census block groups [,]. For this study, participants’ residential addresses were geocoded to census block groups and then linked with the corresponding ADI national ranking percentile, which ranks neighborhoods by socioeconomic disadvantage at the national level in the United States. ADI scores range from 1 to 100, with higher scores reflecting higher levels of neighborhood socioeconomic deprivation. We further categorized ADI scores into quartiles. The first quartile represented the least socioeconomically deprived neighborhoods, whereas the fourth quartile represented the most deprived neighborhoods.

Telemedicine use was defined as having had a telehealth visit with a physician or other health providers in any specialty in the past 12 months and dichotomized as yes or no. For patients who used telemedicine, we asked whether their visits were conducted through telephone, videoconferencing, or both. Similarly, in-person visits were assessed by asking participants whether they had had an in-person office visit with a physician or other health providers in the past 12 months. Furthermore, participants were asked whether their telemedicine or in-person visits were related to 6 different types of oncology services: treatment consultation; review of laboratory, screening, and pathology test results; management of cancer symptoms and treatment side effects; cancer genetic counseling; cancer clinical trial follow-up; and informed consent for a cancer clinical trial. Common cancer symptoms and treatment side effects discussed during telehealth or in-person visits were also assessed, including hot flashes; chemotherapy-induced neuropathy, nausea, and vomiting; pain related to cancer or cancer treatment; depressive symptoms or mood changes; fatigue or tiredness; anxiety or stress; lymphedema; and insomnia or sleep problems.

In addition, we asked participants to report any challenges or concerns when using telemedicine, such as technology difficulty or lack of comfort with technology, lack of electronic device (eg, desktop computer, laptop computer, smartphone, or iPad), lack of high-speed internet or slow internet connection at home, compromised patient-provider communication, compromised patient-provider relationship, telemedicine not being offered at the clinic or by a provider, cost, and telemedicine not being covered by health insurance. We then asked how satisfied participants were with their telehealth or in-person visits, using a 5-point Likert scale (ie, not at all, a little, somewhat, very, and extremely satisfied). Participants were also asked how likely they were to continue using telemedicine, using another 5-point Likert scale (ie, very unlikely, unlikely, neutral, likely, and very likely).

Individual-level sociodemographic and clinicopathologic characteristics included age at survey, race, ethnicity, highest level of education, marital status, type of health insurance coverage, duration from cancer diagnosis to survey, Charlson comorbidity index (excluding breast cancer diagnoses), histologic type, American Joint Committee on Cancer stage group, molecular subtype, tumor grade, receipt of cancer treatment (chemotherapy, hormone therapy, or radiotherapy), and type of surgery. We obtained patients’ clinicopathologic information from electronic health records and the hospital cancer registry. Distance from residence to hospital (in miles) was geocoded and calculated by taking the differences of coordinates (longitudes or latitudes) between the patient’s address and the University of Chicago Medicine Comprehensive Cancer Center’s address based on the Haversine formula.

Statistical Analysis

We described patients’ characteristics using summary statistics. Means and SDs or medians and IQRs were calculated for continuous variables, and we used 2-tailed t tests, Wilcoxon rank sum tests, or Kruskal-Wallis tests to conduct bivariate analyses. For nominal data, we tabulated frequencies and percentages and compared the distributions using Pearson chi-square or Fisher exact tests. To examine the association between neighborhood socioeconomic disadvantage (continuous ADI scores) and telemedicine use, we fitted 3 separate multivariable logistic regression models. For modeling, we implemented a stepwise regression approach. Potential confounders were selected and adjusted for in the models based on a P value of <.10 from bivariate analyses or a priori knowledge. Model 1 included ADI, age at survey, race, ethnicity, duration from cancer diagnosis to survey, highest level of education, marital status, type of health insurance coverage, Charlson comorbidity index, and distance from residence to hospital. Model 2 was controlled for histologic type, American Joint Committee on Cancer stage, molecular subtype, and tumor grade, in addition to the covariates in model 1. Model 3 contained all variables in model 2 plus receipt of chemotherapy, hormone therapy, or radiotherapy, as well as type of surgery. Adjusted odds ratios (AORs) and corresponding 95% CIs were calculated. To evaluate the differences in types of oncology services between telemedicine and in-person office visits, we conducted McNemar tests on match-paired data of patients having both visit modalities. P values (2-tailed) <.05 were considered statistically significant. All statistical analyses were performed using Stata 17 (StataCorp LLC).

ResultsPatient Characteristics

Overall, the 1868 study surveys received 1236 (66.17%) responses. Of the 1236 participants who responded, 1163 (94.09%) had had either telemedicine or in-person visits in the past 12 months. These participants’ mean age was 61.8 (SD 12.0) years; 4.48% (52/1161) identified as Asian, 19.72% (229/1161) as Black, 3.01% (35/1161) as Hispanic, and 72.78% (845/1161) as White. Furthermore, 69.94% (747/1068) were married, 38.73% (450/1162) had a graduate or professional degree, 70.77% (823/1163) were privately insured, and 22.96% (267/1163) were on Medicaid or Medicare. The median distance from residence to hospital was 19.9 (IQR 9.5-32.3) miles, and the median duration from cancer diagnosis to survey was 6.5 (IQR 3.6-11.0) years. By ADI quartile, patients with breast cancer living in the most socioeconomically disadvantaged neighborhoods (fourth quartile) tended to be older, Black, at a lower level of education, and on Medicaid or Medicare ().

Table 1. Characteristics of patients with breast cancer overall and by neighborhood socioeconomic disadvantage (n=1163).VariableTotal (n=1163)Area Deprivation IndexaP valueb

First quartile (n=381)Second quartile (n=376)Third quartile (n=252)Fourth quartile (n=99)
Age (y) at survey, mean (SD)61.8 (12.0)60.9 (11.5)61.6 (11.7)62.0 (12.9)64.2 (12.5).68Age (y) at survey, n (%).03
<45107 (10.2)33 (9.5)37 (10.8)26 (11.4)7 (7.8)

45-54179 (17)64 (18.4)59 (17.3)38 (16.7)12 (13.3)

55-64308 (29.2)116 (33.3)99 (28.9)58 (25.4)22 (24.4)

≥65460 (43.6)135 (38.8)147 (43)106 (46.5)49 (54.4)
Race and ethnicity, n (%)<.001
Asian52 (4.5)26 (6.8)14 (3.7)6 (2.4)3 (3)

Black229 (19.7)16 (4.2)40 (10.7)98 (38.9)56 (56.6)

Hispanic35 (3)5 (1.3)20 (5.3)5 (2)4 (4)

White845 (72.8)333 (87.6)301 (80.3)143 (56.7)36 (36.4)
Highest level of education, n (%)<.001
High school, GEDc, or less115 (9.9)12 (3.1)45 (12)37 (14.7)16 (16.2)

Associate’s degree or some college259 (22.3)52 (13.6)86 (22.9)70 (27.8)44 (44.4)

Bachelor’s degree338 (29.1)127 (33.3)102 (27.2)69 (27.4)20 (20.2)

Graduate or professional degree450 (38.7)190 (49.9)142 (37.9)76 (30.2)19 (19.2)
Marital status, n (%)<.001
Married747 (69.9)282 (80.3)259 (73.4)136 (59.6)36 (40.4)

Single or not married192 (18)44 (12.5)53 (15)59 (25.9)30 (33.7)

Divorced, separated, or widowed129 (12.1)25 (7.1)41 (11.6)33 (14.5)23 (25.8)
Type of health insurance, n (%)<.001
Private823 (70.8)302 (79.3)276 (73.4)162 (64.3)49 (49.5)

Medicaid50 (4.3)5 (1.3)8 (2.1)17 (6.7)15 (15.2)

Medicare217 (18.7)54 (14.2)74 (19.7)55 (21.8)24 (24.2)

Other or unknown73 (6.3)20 (5.2)18 (4.8)18 (7.1)11 (11.1)
Distance from residence to hospital (miles)d, median (IQR)19.9 (9.5-32.3)20.5 (10.9-31.9)22.5 (13.3-33.2)16.4 (4.6-30.5)11.9 (3.3-27.6)<.001Duration (y) from cancer diagnosis to survey, median (IQR)6.5 (3.6-11.0)6.8 (3.7-10.9)6.2 (3.6-10.3)6.5 (3.6-11.5)8.3 (4.2-11.6).61Duration (y) from cancer diagnosis to survey, n (%).63
≤3199 (17.1)58 (15.2)68 (18.1)48 (19)14 (14.1)

4-6319 (27.4)107 (28.1)107 (28.5)67 (26.6)23 (23.2)

≥7645 (55.5)216 (56.7)201 (53.5)137 (54.4)62 (62.6)
Charlson comorbidity index, n (%).03
0994 (88.5)333 (90.5)335 (91.5)209 (85)77 (82.8)

162 (5.6)19 (5.2)11 (3)21 (8.5)6 (6.5)

≥267 (6.0)16 (4.3)20 (5.5)16 (6.5)10 (10.8)
Histologic type, n (%).08
Ductal742 (80.2)247 (77.9)238 (79.9)159 (81.1)64 (88.9)

Lobular92 (10)38 (12)34 (11.4)15 (7.7)1 (1.4)

Ductal and lobular55 (6)19 (6)18 (6)12 (6.1)3 (4.2)

Other36 (3.9)13 (4.1)8 (2.7)10 (5.1)4 (5.6)
AJCCe stage group, n (%).002
0200 (18.1)51 (14.2)69 (19)51 (21.1)21 (22.6)

I515 (46.5)189 (52.5)160 (44.1)104 (43)36 (38.7)

II271 (24.5)88 (24.4)91 (25.1)58 (24)24 (25.8)

III112 (10.1)31 (8.6)42 (11.6)24 (9.9)10 (10.8)

IV10 (0.9)1 (0.3)1 (0.3)5 (2.1)2 (2.2)
Molecular subtype, n (%).06
HRf+/HER2g−571 (66.2)208 (69.3)180 (66.2)120 (65.6)35 (53.8)

HR+/HER+98 (11.4)34 (11.3)36 (13.2)15 (8.2)8 (12.3)

HR−/HER2+51 (5.9)12 (4)17 (6.2)19 (10.4)3 (4.6)

TNBCh142 (16.5)46 (15.3)39 (14.3)29 (15.8)19 (29.2)
Tumor grade, n (%).047
1149 (14.3)59 (17.3)47 (13.8)27 (11.9)9 (10.3)

2471 (45.3)159 (46.6)146 (42.9)99 (43.8)42 (48.3)

3420 (40.4)123 (36.1)147 (43.2)100 (44.2)36 (41.4)
Receipt of chemotherapy, n (%).92
No572 (54.3)190 (54.6)182 (53.2)125 (54.8)48 (53.3)

Yes482 (45.7)158 (45.4)160 (46.8)103 (45.2)42 (46.7)
Receipt of hormone therapy, n (%).03
No341 (32.4)100 (28.7)113 (33)74 (32.5)39 (43.3)

Yes713 (67.7)248 (71.3)229 (67)154 (67.5)51 (56.7)
Receipt of radiation therapy, n (%).08
No394 (37.4)140 (40.2)125 (36.5)83 (36.4)26 (28.9)

Yes660 (62.6)208 (59.8)217 (63.5)145 (63.6)64 (71.1)
Type of surgery received, n (%).006
None13 (1.3)5 (1.5)3 (0.9)2 (0.9)3 (3.4)

Lumpectomy615 (59.3)185 (53.8)194 (57.4)146 (66.1)61 (68.5)

Mastectomy307 (29.6)116 (33.7)107 (31.7)50 (22.6)25 (28.1)

Bilateral mastectomy102 (9.9)38 (11.0)34 (10.1)23 (10.4)0 (0)

aThe Area Deprivation Index (national ranking percentile) is a composite measure consisting of the domains of income, education, employment, and housing quality. It ranks neighborhoods by socioeconomic disadvantage at the national level and is scored from 1 to 100, with higher scores representing greater neighborhood socioeconomic deprivation.

bP values were calculated using Kruskal-Wallis tests.

cGED: General Educational Development Test.

dDistance from residence to hospital was calculated by taking the differences of coordinates (longitudes or latitudes) between the patient’s address and the University of Chicago Medicine Comprehensive Cancer Center’s address based on the Haversine formula.

eAJCC: American Joint Committee on Cancer.

fHR: hormone receptor.

gHER2: human epidermal growth factor receptor 2.

hTNBC: triple-negative breast cancer.

Telemedicine Use and Association With ADI

Overall, 35.95% (416/1157) of the patients with breast cancer had a telehealth visit in the past 12 months (). By modality of telemedicine, 65% (266/409) of the clinic visits were conducted through videoconferencing only, followed by 22.7% (93/409) through telephone only and 12.2% (50/409) through both videoconferencing and telephone. The mean ADI score for telemedicine users was 37.7 (SD 24.2) compared to 39.5 (SD 24.0) for nonusers (). By ADI quartile, 38.3% (145/379) of the patients living in the least socioeconomically disadvantaged neighborhoods (first quartile) used telemedicine, followed by 37.9% (58/153), 35.1% (132/356), and 32.5% (81/249) in the fourth, second, and third quartiles, respectively. On multivariable regression analysis (model 3), higher ADI scores (per 10-unit increase) were associated with lower odds of telemedicine use (AOR 0.89, 95% CI 0.82-0.97; ).

Table 2. Characteristics of patients with breast cancer by telehealth visit (n=1157).VariableHad a telehealth visit in the past 12 monthsP valuea
No (n=741)Yes (n=416)

Modality of telemedicine (n=409), n (%)—b
Telephone or audio call—93 (22.7)

Videoconference—266 (65)

Both—50 (12.3)
Area Deprivation Indexc, mean (SD)39.5 (24.0)37.7 (24.2).18Area Deprivation Index, n (%).13
First quartile234 (61.7)145 (38.3)

Second quartile224 (64.9)132 (35.1)

Third quartile168 (67.5)81 (32.5)

Fourth quartile95 (62.1)58 (37.9)
Age (y) at survey, mean (SD)62.2 (11.9)60.9 (12.2).09Age (y) at survey, n (%).04
<4556 (52.3)51 (47.7)

45-54121 (67.6)58 (32.4)

55-64201 (66.3)102 (33.7)

≥65299 (65.1)160 (34.9)
Race and ethnicity, n (%).08
Asian37 (71.2)15 (28.9)

Black136 (60.2)90 (39.8)

Hispanic17 (48.6)18 (51.4)

White550 (65.3)292 (34.7)
Highest level of education, n (%).002
High school, GEDd, or less92 (80)23 (20)

Associate’s degree or some college163 (63.4)94 (36.6)

Bachelor’s degree208 (61.9)128 (38.1)

Graduate or professional degree277 (61.8)171 (38.2)
Marital status, n (%).70
Married481 (64.7)263 (35.4)

Single or not married116 (61.4)73 (38.6)

Divorced, separated, or widowed83 (64.3)46 (35.7)
Type of health insurance, n (%).25
Private515 (63)302 (37)

Medicaid28 (56)22 (44)

Medicare147 (67.7)70 (32.3)

Other or unknown51 (69.9)22 (30.1)
Distance (miles) from residence to hospitale, median (IQR)19.9 (9.8-32.3)20.4 (9.3-32.3).96Duration (y) from cancer diagnosis to survey, median (IQR)6.8 (3.7-0.9)6.3 (3.5-11.0).22Duration (years) from cancer diagnosis to survey, n (%).009
≤3109 (55)89 (45)

4-6217 (68)102 (32)

≥7415 (64.8)225 (35.2)
Charlson comorbidity index, n (%).31
0635 (64.3)353 (35.7)

134 (54.8)28 (45.2)

≥244 (65.7)23 (34.3)
Histologic type, n (%).27
Ductal459 (62.1)280 (37.9)

Lobular63 (68.5)29 (31.5)

Ductal and lobular38 (69.1)17 (30.9)

Other19 (52.8))17 (47.2)
AJCCf stage group, n (%).26
0135 (68.5)62 (31.5)

I333 (64.9)180 (35.1)

II161 (59.6)109 (40.4)

III75 (61.5)47 (38.5)

IV5 (50)5 (50)
Molecular subtype, n (%).91
HRg+/HER2h−358 (62.8)212 (37.2)

HR+/HER+64 (65.3)34 (34.7)

HR−/HER2+34 (66.7)17 (33.3)

TNBCi87 (62.1)53 (37.9)
Tumor grade, n (%).10
187 (59.6)59 (40.4)

2316 (67.4)153 (32.6)

3258 (61.6)161 (38.4)
Receipt of chemotherapy, n (%).19
No377 (66.4)191 (33.6)

Yes300 (62.5)180 (37.5)
Receipt of hormone therapy, n (%).92
No217 (64.4)120 (35.6)

Yes460 (64.7)251 (35.3)
Receipt of radiation therapy, n (%).68
No255 (65.4)135 (34.6)

Yes422 (64.1)236 (35.9)
Type of surgery received, n (%).35
None6 (46.2)7 (53.8)

Lumpectomy404 (66)208 (34)

Mastectomy190 (62.5)114 (37.5)

Bilateral mastectomy68 (66.7)34 (33.3)

aP values were calculated using 2-tailed t tests or Wilcoxon rank sum, Pearson chi-square, or Fisher exact tests, as appropriate.

bNot applicable.

cThe Area Deprivation Index (national ranking percentile) is a composite measure consisting of the domains of income, education, employment, and housing quality. It ranks neighborhoods by socioeconomic disadvantage at the national level and is scored from 1 to 100, with higher scores representing greater neighborhood socioeconomic deprivation.

dGED: General Educational Development Test.

eDistance from residence to hospital was calculated by taking the differences of coordinates (longitudes or latitudes) between the patient’s address and the University of Chicago Medicine Comprehensive Cancer Center’s address based on the Haversine formula.

fAJCC: American Joint Committee on Cancer.

gHR: hormone receptor.

hHER2: human epidermal growth factor receptor 2.

iTNBC: triple-negative breast cancer.

In the same model (model 3), patients with breast cancer aged 45 to 54 years had lower odds of having a telehealth visit than those aged <45 years (AOR 0.49, 95% CI 0.27-0.91). Patients aged 55 to 64 years (AOR 0.63, 95% CI 0.36-1.12) or ≥65 years (AOR 0.63, 95% CI 0.34-1.18) also had a lower likelihood, but these differences were not statistically significant. Black (AOR 2.38, 95% CI 1.41-4.00) or Hispanic (AOR 2.65, 95% CI 1.07-6.58) patients had greater odds of telemedicine use than White patients. Compared to patients with high school or less education, those with an associate’s (AOR 2.67, 95% CI 1.33-5.35), bachelor’s (AOR 2.75, 95% CI 1.38-5.48), or graduate (AOR 2.57, 95% CI 1.31-5.04) degree had higher odds of telemedicine use in the past 12 months. Longer distance from residence to hospital (per 10-mile increase) was associated with greater odds of use of telemedicine, although this was not statistically significant (AOR 1.02, 95% CI 0.96-1.09; ). Clinicopathologic and treatment factors were not significantly associated with telemedicine use (Table S1 in ). In subgroup analyses, ADI scores were not significantly different between videoconference and telephone visits (AOR 0.88, 95% CI 0.73-1.07). We also observed that patients with a graduate or professional degree had greater odds of using videoconference visits (AOR 5.78, 95% CI 1.03-32.55), and patients on Medicare had lower odds of videoconference visit use than privately insured patients (AOR 0.26, 95% CI 0.07-0.91; Table S2 in ).

Table 3. Association between neighborhood socioeconomic disadvantage and telemedicine use in patients with breast cancer.VariableModel 1, adjusted odds ratioa (95% CI)P valueModel 2, adjusted odds ratiob (95% CI)P valueModel 3, adjusted odds ratioc (95% CI)P valueArea Deprivation Indexd (continuous)e0.93 (0.87-0.99).030.89 (0.82-0.96).0050.89 (0.82-0.97).004Distance from residence to hospitale1.04 (0.99-1.10).131.03 (0.97-1.09).401.02 (0.96-1.09).48Age (y) at survey
<451.0 (reference)
1.0 (reference)
1.0 (reference)

45-540.55 (0.33-0.94).030.53 (0.29-0.97).040.49 (0.27-0.91).02
55-640.57 (0.35-0.93).020.64 (0.37-1.11).110.63 (0.36-1.12).11
≥650.65 (0.39-1.09).100.62 (0.34-1.13).120.63 (0.34-1.18).16Race and ethnicity
Asian0.55 (0.26-1.17).120.50 (0.20-1.22).130.50 (0.20-1.23).16
Black1.86 (1.21-2.86).0052.50 (1.48-4.20).0012.38 (1.41-4.00).001
Hispanic2.12 (1.02-4.41).042.85 (1.17-6.91).022.65 (1.07-6.58).03
White1.0 (reference)
1.0 (reference)
1.0 (reference)
Highest level of education
High school, GEDf, or less1.0 (reference)
1.0 (reference)
1.0 (reference)

Associate’s degree or some college2.66 (1.47-4.81).0012.76 (1.40-5.44).0032.67 (1.33-5.35).006
Bachelor’s degree