Abstract
Objective
In response to the coronavirus disease 2019 (COVID-19) pandemic, the Veterans Health Administration (VA) rapidly expanded virtual care (defined as care delivered by video and phone), raising concerns about technology access disparities (ie, the digital divide). Virtual care was somewhat established in primary care and mental health care prepandemic, but video telehealth implementation was new for most subspecialties, including cardiology. We sought to identify patient characteristics of virtual and video-based care users in VA cardiology clinics nationally during the first year of the COVID-19 pandemic.
Materials and Methods
Cohort study of Veteran patients across all VA facilities with a cardiology visit January 1, 2019–March 10, 2020, with follow-up January 1, 2019–March 10, 2021. Main measures included cardiology visits by visit type and likelihood of receiving cardiology-related virtual care, calculated with a repeated event survival model.
Results
416 587 Veterans with 1 689 595 total cardiology visits were analyzed; average patient age was 69.6 years and 4.3% were female. Virtual cardiology care expanded dramatically early in the COVID-19 pandemic from 5% to 70% of encounters. Older, lower-income, and rural-dwelling Veterans and those experiencing homelessness were less likely to use video care (adjusted hazard ratio for ages 75 and older 0.80, 95% confidence interval (CI) 0.75–0.86; for highly rural residents 0.77, 95% CI 0.68–0.87; for low-income status 0.94, 95% CI 0.89–0.98; for homeless Veterans 0.85, 95% CI 0.80–0.92).
Conclusion
The pandemic worsened the digital divide for cardiology care for many vulnerable patients to the extent that video visits represent added value over phone visits. Targeted interventions may be necessary for equity in COVID-19-era access to virtual cardiology care.
Keywords: telemedicine, VA, specialty care, access to care
BACKGROUND AND SIGNIFICANCE
The coronavirus disease 2019 (COVID-19) pandemic prompted an abrupt shift from in-person to virtual patient encounters within the United States Veterans Administration (VA), as across the country and the world.1,2 Beginning March 11th, 2020, when the World Health Organization formally declared COVID-19 a pandemic,3 Veterans Health Administration (VA) explicitly encouraged providers to substitute virtual care, defined as phone or video visits, for in-person patient visits when feasible.4–6
When in-person care is disrupted, whether because of a pandemic or other personal or widespread catastrophe, virtual care is an essential tool for ensuring continuity of care.7,8 In the case of an infectious disease pandemic like COVID-19, virtual care can additionally decrease patient and clinician exposure to potential infection. Compared to in-person care, virtual care has been shown to offer important benefits with respect to access to treatment, patient satisfaction, and reduced cost to patients, including within VA.9–15
Yet because virtual care is so necessary and beneficial and because it relies on access to technology, its rapid and widespread expansion raises concerns about disparities in access to care—a dynamic frequently referred to as the digital divide.16,17 Increasing evidence suggests that older, rural, and low-income individuals face larger barriers to care when in-person care is limited, such as while sheltering in place, and thus may be susceptible to the digital divide.18,19 Pandemic-era studies in VA suggest that certain patient populations (eg, older adults and those in rural areas) were less likely to have a video visit early in the pandemic.6 However, current evidence regarding the expansion of virtual care within specific specialties, even outside VA, is limited to smaller studies.20 Users of specialty care comprise distinct patient populations with varying vulnerabilities and capabilities; different comorbidities may predispose patients toward preferring virtual care or in-person care or toward a certain modality of virtual care (ie, video vs phone), and patients requiring specialty care may be more medically complex on average. Cardiology exemplifies a subspecialty field posited as ripe for expansion of virtual care, given the existence of various monitoring technologies relevant for cardiovascular comorbidities and the possibility of performing some critical physical exam maneuvers via video.21–23
OBJECTIVE
We sought to identify patient characteristics of virtual and video-based care users in VA cardiology clinics nationally during the first year of the COVID-19 pandemic in order to inform policies and interventions related to virtual care expansion.
MATERIALS AND METHODS
Data sources and patient population
All VA patient and encounter data came from VA’s Corporate Data Warehouse repository.24 We identified a cohort of patients 18 and over with at least 1 VA cardiology encounter with an associated cardiovascular diagnosis between January 1, 2019 and March 10, 2020 and followed them for virtual care cardiology encounters from this first encounter through the end of March 10, 2021. Since this analysis focused on pandemic-era follow-up care, patients were excluded from the analysis if they died prior to March 11th, 2020; patients who died during the pandemic were included in the analysis. Veterans receiving care in the US territories (US Virgin Islands, Guam, Palau, and the Northern Marianas Islands) were also excluded from analysis (<0.1%). The final cohort included 416 609 Veterans with a total of 1 693 540 cardiology visits over the study period.
Classification of VA outpatient cardiology encounters
Cardiology encounters were identified using VA’s Managerial Cost Accounting 3-digit stop code for cardiology visits as either the primary code for the visit, or as the secondary code for encounters where the primary code indicated a telephone visit; these codes characterize all VA outpatient encounters and their clinical work units to inform resource allocation. We included only visits using Evaluation and Management (E&M) Current Procedural Terminology (CPT) codes. Cardiovascular diagnoses are included in Supplementary Table S1. Encounters were categorized as telephone, video, or in-person visits (as indicated by primary or secondary stop code) (Supplementary Table S2) and aggregated into weekly counts, starting on Sunday.
Primary outcome
The main outcome of interest for both the encounter-level and Veteran-level analyses was outpatient cardiology visits, defined by the stop codes and CPT codes described above, delivered by phone or by video.
Veteran characteristics
Patient-level data included sociodemographic and clinical characteristics as described below. Age was categorized as <50, 50–64, 65–74, and ≥75 years, roughly corresponding to quartiles for the study population. Race and ethnicity data were based on the most frequently recorded race and ethnic identification in the electronic health record. Distinct categories were created for missing data for race or ethnicity. Number of chronic conditions was out of a set of 47 possible International Statistical Classification of Disease 10 diagnosis groups, defined based on prior research in VA populations25–28 (Supplementary Table S4). Urban and rural were defined by US Census Bureau criteria,29 with highly rural defined as areas with population density less than 7 people per square mile.30 Drive time to secondary care, which would encompass cardiology care, was divided into categories of ≤30 minutes, 31–60 minutes, or >60 minutes. Homelessness was based on VA diagnosis codes and outpatient stop codes denoting use of homeless services (Supplementary Tables S2 and S3).
The VA priority-based enrollment system uses service-connected disability, income, recent military service, and other factors to categorize VA patients. As in Ferguson et al 2020,6 the eight groups comprising VA’s priority-based enrollment system30 were condensed to form four categories representing different levels of social risk: high disability (either >50% service-connected disability or VA catastrophically disabled), low-moderate disability (10%–20% or 30%–40% service-connected disability or military exposure), low income (annual income below area-adjusted mean), or without special VA enrollment priority (0% service-connected disability).
Utilization of noncardiology care prior to the pandemic was defined in terms of primary care utilization in the calendar year prior to each patient’s first visit in the study period, where patients were categorized into tertiles as low-, medium-, or high-frequency users if they had ≤4, 5–8 or ≥9 visits in that year, respectively. Binary variables for use of mental health care and emergency department or urgent care in the same time period were also constructed.
We defined the geographic region of each Veterans by the particular VA Veterans Integrated Service Network location (VISN) (N = 18 geographic regions) where the patient received specialty care.
Analysis
We examined the above characteristics by encounter type and timing (before vs during the pandemic), as well as among Veteran users of any cardiology care, video care, and virtual care.
Our analysis followed this cohort of patients over time from their initial visit in the study period beginning January 1, 2019 through March 10th, 2021. We opted to use a Prentice, Williams, and Peterson repeated event survival model rather than for example, multivariable logistic regression to account for a differential time under observation across patients as well as to allow for censoring in this relatively high-risk, high-mortality population. Due to limits on computational power, we applied the Breslow method for handling event ties for these models, which tends to produce a less precise approximation of the marginal likelihood than the Efron method;31–33 we therefore performed a sensitivity analysis with a 10% random sample using the Efron method. To account for potential differences in the baseline hazard of using virtual care prior and during the COVID-19 pandemic, the baseline hazard for all models contained a strata statement for prior or post March 10, 2020.34 Results were considered significant if p < 0.05.
Statistical analyses and graphical output were conducted in Stata 15 (StataCorp, LLC, College Station, TX).
This study was part of the Virtual Access Quality Enhancement Research Initiative and was designated as nonresearch quality improvement by VA’s Office of Rural Health. This report follows the STROBE reporting guideline for observational studies.35
RESULTS
Changes in cardiology care delivery patterns
1 689 595 video, phone, or face-to-face encounters took place over the study period, 634 778 of which occurred during the pandemic. Cardiology care shifted dramatically from face-to-face to virtual care in the initial weeks of the pandemic; Figure 1 shows weekly visits by modality. There were 158 752 patients (38.1% of total cohort patients) with no visits after the onset of the pandemic (Supplementary Table S5). This is reflected in the total number of cardiology encounters, which dropped in the first months of the pandemic from 73 493 visits in February 2020 to 50 603 in April 2020 and remained persistently below the prepandemic baseline at between 62.6% and 76.9% of the average monthly visits for the 3 months preceding the pandemic (December 2019–February 2020). By April 2020, the proportion of face-to-face visits of all cardiology care fell from prepandemic levels of 95.2% to 29.0%, while phone visits increased from 4.8% to 71.0%. Very few Veterans used video care for cardiology prior to the pandemic, with only 189 total video cardiology visits taking place for this population between January 1, 2019 and March 11, 2020. In contrast, 29 415 cardiology video visits occurred in the first year of the pandemic. While video visits have remained a minority of virtual care, the share of virtual visits comprised of video visits rose steadily from 1.0% immediately before the pandemic to an average of 15.1% during the first 3 months of 2021.
Figure 1.
VA cardiology encounters by week, January 2020–March 2021.
Patient characteristics associated with virtual care use
Table 1 shows patient characteristics by encounter type and timing before versus during the pandemic, while Table 2 compares characteristics of all patients, video care users, virtual care users, and virtual care nonusers.
Table 1.
Characteristics of encounters by modality (video, phone, and face-to-face) (N = 1 689 595) before and during the first year of the COVID-19 pandemic
| Video encounters |
Phone encounters |
Face-to-face encounters |
||||
|---|---|---|---|---|---|---|
| Prior to pandemic, No. (%)a,b | During pandemic, No. (%) | Prior to pandemic, No. (%) | During pandemic, No. (%) | Prior to pandemic, No. (%) | During pandemic, No. (%) | |
| Total encounters | N = 189 | N = 26 415 | N = 44 087 | N = 271 998 | N = 1 010 541 | N = 336 365 |
| Age, years (mean [SD]) | 67.5 (11.2) | 68.6 (10.5) | 70.2 (10.1) | 70.5 (9.8) | 70.1 (10.3) | 70.3 (9.6) |
| Age, years, categorical | ||||||
| 18–49 | 17 (9.0) | 1281 (4.8) | 1542 (3.5) | 7872 (2.9) | 38 285 (3.8) | 9698 (2.9) |
| 50–64 | 40 (21.2) | 6513 (24.7) | 8994 (20.4) | 53 262 (19.6) | 205 668 (20.4) | 67 216 (20.0) |
| 65–74 | 95 (50.3) | 12 358 (46.8) | 20 751 (47.1) | 130 791 (48.1) | 475 042 (47.0) | 162 878 (48.4) |
| 75+ | 37 (19.6) | 6263 (23.7) | 12 800 (29.0) | 80 073 (29.4) | 291 546 (28.9) | 96 573 (28.7) |
| Race | ||||||
| American Indian or Alaska Native | 1 (0.5) | 187 (0.7) | 404 (0.9) | 2217 (0.8) | 8782 (0.9) | 2940 (0.9) |
| Asian | 2 (1.1) | 202 (0.8) | 451 (1.0) | 1753 (0.6) | 7176 (0.7) | 2241 (0.7) |
| Black or African American | 25 (13.2) | 4566 (17.3) | 6317 (14.3) | 47 716 (17.5) | 186 890 (18.5) | 63 904 (19.0) |
| Native Hawaiian or other Pacific Islander | 2 (1.1) | 294 (1.1) | 428 (1.0) | 2346 (0.9) | 8786 (0.9) | 2783 (0.8) |
| Unknown | 11 (5.8) | 1474 (5.5) | 1935 (4.4) | 11 314 (4.2) | 42 107 (4.2) | 12 632 (3.8) |
| White | 148 (78.3) | 19 709 (74.6) | 34 552 (78.4) | 206 652 (76.0) | 756 800 (74.9) | 251 865 (74.9) |
| Ethnicity | ||||||
| Hispanic or Latino | 6 (3.2) | 2330 (8.8) | 2131 (4.8) | 16 551 (6.1) | 53 783 (5.3) | 14 395 (4.3) |
| Not Hispanic or Latino | 180 (95.2) | 23 143 (87.6) | 40 542 (92.0) | 247 480 (91.0) | 928 153 (91.8) | 313 207 (93.1) |
| Unknown | 3 (1.6) | 942 (3.6) | 1414 (3.2) | 7967 (2.9) | 28 605 (2.8) | 8763 (2.6) |
| Gender | ||||||
| Female | 7 (3.7) | 1182 (4.5) | 1688 (3.8) | 9316 (3.4) | 37 529 (3.7) | 10 575 (3.1) |
| Male | 182 (96.3) | 25 233 (95.5) | 42 399 (96.2) | 262 682 (96.6) | 973 012 (96.3) | 325 790 (96.9) |
| Rurality | ||||||
| Highly rural | 6 (3.2) | 492 (1.9) | 1742 (4.0) | 10 711 (3.9) | 37 171 (3.7) | 12 650 (3.8) |
| Rural | 53 (28.0) | 5296 (20.1) | 13 434 (30.5) | 82 555 (30.4) | 296 471 (29.4) | 97 814 (29.2) |
| Urban | 130 (68.8) | 20 558 (78.0) | 28 849 (65.5) | 178 108 (65.6) | 673 942 (66.9) | 224 746 (67.0) |
| Drive time to secondary care, minutes (mean [SD]) | 50.9 (37.7) | 37.8 (28.1) | 43.7 (33.6) | 42.6 (33.0) | 42.2 (33.8) | 41.1 (33.3) |
| Drive time to secondary care, categorical | ||||||
| Short | 55 (29.3) | 12 799 (48.8) | 18 602 (42.5) | 120 984 (44.8) | 471 334 (47.0) | 162 196 (48.5) |
| Medium | 80 (42.6) | 9284 (35.4) | 14 745 (33.7) | 87 640 (32.4) | 304 841 (30.4) | 99 884 (29.9) |
| Long | 53 (28.2) | 4167 (15.9) | 10 372 (23.7) | 61 549 (22.8) | 227 522 (22.7) | 72 173 (21.6) |
| Enrollment priority | ||||||
| No service disability | 33 (17.5) | 3138 (11.9) | 5516 (12.5) | 32 731 (12.1) | 121 612 (12.0) | 38 202 (11.4) |
| Low income | 41 (21.7) | 5997 (22.7) | 11 373 (25.8) | 70 137 (25.8) | 262 368 (26.0) | 87 346 (26.0) |
| Low/moderate disability | 37 (19.6) | 5044 (19.1) | 8168 (18.6) | 48 851 (18.0) | 184 158 (18.2) | 60 255 (17.9) |
| High disability | 78 (41.3) | 12 236 (46.3) | 19 030 (43.0) | 120 274 (44.1) | 442 832 (43.8) | 150 549 (44.8) |
| Primary care visits in year prior to analysis period (mean [SD]) | 8.0 (7.4) | 7.9 (7.6) | 8.1 (7.7) | 7.5 (7.1) | 7.7 (7.5) | 8.0 (7.8) |
| Primary care visits in year prior to analysis period, categorical | ||||||
| 0–4 | 69 (36.5) | 9990 (37.8) | 16 079 (36.5) | 109 144 (40.1) | 402 917 (39.9) | 129 222 (38.4) |
| 5–8 | 58 (30.7) | 7903 (29.9) | 13 307 (30.2) | 83 432 (30.7) | 303 309 (30.0) | 101 074 (30.0) |
| 9+ | 62 (32.8) | 8522 (32.3) | 14 701 (33.3) | 79 422 (29.2) | 304 315 (30.1) | 106 069 (31.5) |
| Number of chronic medical conditions (mean [SD]) | 7.2 (3.4) | 8.1 (3.7) | 8.3 (3.8) | 8.1 (3.6) | 8.1 (3.7) | 8.4 (3.8) |
| Number of chronic medical conditions, categorical | ||||||
| 0–3 | 26 (13.8) | 2236 (8.5) | 3471 (7.9) | 21 037 (7.7) | 87 414 (8.7) | 23 905 (7.1) |
| 4–7 | 76 (40.2) | 10 466 (39.6) | 16 799 (38.1) | 110 684 (40.7) | 405 080 (40.1) | 128 342 (38.2) |
| 8–11 | 66 (34.9) | 9270 (35.1) | 15 298 (34.7) | 95 001 (34.9) | 341 720 (33.8) | 119 011 (35.4) |
| 12+ | 21 (11.1) | 4443 (16.8) | 8519 (19.3) | 45 276 (16.6) | 176 327 (17.4) | 65 107 (19.4) |
| At least 1 emergency visit in the year prior to analysis period | ||||||
| No | 112 (59.3) | 13 584 (51.4) | 20 992 (47.6) | 136 866 (50.3) | 487 710 (48.3) | 155 097 (46.1) |
| Yes | 77 (40.7) | 12 831 (48.6) | 23 095 (52.4) | 135 122 (49.7) | 522 831 (51.7) | 181 268 (53.9) |
| At least 1 mental health visit in the year prior to analysis period | ||||||
| No | 129 (68.3) | 16 644 (63.0) | 28 999 (65.8) | 180 853 (66.5) | 660 766 (65.4) | 221 413 (65.8) |
| Yes | 60 (31.7) | 9764 (37.0) | 15 088 (34.2) | 91 145 (33.5) | 349 775 (34.6) | 114 952 (34.2) |
Note: Prepandemic study period January 1, 2019–March 10, 2020; pandemic study period March 11, 2020–March 10, 2021.
For continuous variables, figures outside of parentheses represent means and figures in parentheses represent standard deviations.
Some categories may not sum to overall total due to missing data.
Table 2.
Characteristics of study population (N = 416 587) and of video and virtual care users and nonusers, patient level
| All patients, No. (%)a,b | Video users, No. (%) | Virtual carec users, No. (%) | Virtual care nonusers, No. (%) | |
|---|---|---|---|---|
| Total patients | N = 416 587 | N = 18 478 | N = 167 646 | N = 248 941 |
| Age, years (mean [SD]) | 69.6 (10.8) | 68.7 (10.4) | 70.3 (9.9) | 69.2 (11.3) |
| Age, years, categorical | ||||
| 18–49 | 19 489 (4.7) | 885 (4.8) | 5288 (3.2) | 14 201 (5.7) |
| 50–64 | 87 358 (21.0) | 4444 (24.1) | 33 297 (19.9) | 54 061 (21.7) |
| 65–74 | 192 573 (46.2) | 8744 (47.3) | 81 015 (48.3) | 111 558 (44.8) |
| 75+ | 117 167 (28.1) | 4405 (23.8) | 48 046 (28.7) | 69 121 (27.8) |
| Race | ||||
| American Indian or Alaska Native | 3636 (0.9) | 136 (0.7) | 1373 (0.8) | 2263 (0.9) |
| Asian | 3228 (0.8) | 142 (0.8) | 1071 (0.6) | 2157 (0.9) |
| Black or African American | 71 370 (17.1) | 2978 (16.1) | 28 110 (16.8) | 43 260 (17.4) |
| Native Hawaiian or other Pacific Islander | 3785 (0.9) | 219 (1.2) | 1538 (0.9) | 2247 (0.9) |
| Unknown | 17 962 (4.3) | 1008 (5.5) | 6884 (4.1) | 11 045 (4.4) |
| White | 316 639 (76.0) | 13 995 (75.7) | 128 670 (76.8) | 187 969 (75.5) |
| Ethnicity | ||||
| Hispanic or Latino | 20 887 (5.0) | 1350 (7.3) | 8793 (5.2) | 12 094 (4.9) |
| Not Hispanic or Latino | 383 326 (92.0) | 16 433 (88.9) | 153 853 (91.8) | 229 473 (92.2) |
| Unknown | 12 374 (3.0) | 695 (3.7) | 5000 (3.0) | 7374 (3.0) |
| Gender | ||||
| Female | 17 815 (4.3) | 820 (4.4) | 5818 (3.5) | 11 997 (4.8) |
| Male | 398 772 (95.7) | 17 658 (95.6) | 161 828 (96.5) | 236 944 (95.2) |
| Rurality | ||||
| Highly rural | 16 668 (4.0) | 338 (1.8) | 6488 (3.9) | 10 180 (4.1) |
| Rural | 128 462 (31.0) | 3799 (20.6) | 51 178 (30.6) | 77 284 (31.2) |
| Urban | 269 787 (65.0) | 14 293 (77.6) | 109 597 (65.5) | 160 190 (64.7) |
| Drive time to secondary care, minutes (mean [SD]) | 44.4 (35.2) | 38.0 (28.5) | 43.0 (33.1) | 45.4 (36.5) |
| Drive time to secondary care, categorical | ||||
| Short | 182 255 (44.1) | 8974 (48.9) | 73 886 (44.4) | 108 369 (43.9) |
| Medium | 127 611 (30.9) | 6433 (35.0) | 53 962 (32.4) | 73 649 (29.8) |
| Long | 103 708 (25.1) | 2951 (16.1) | 38 625 (23.2) | 65 083 (26.3) |
| Enrollment priority | ||||
| No service disability | 53 805 (12.9) | 2237 (12.1) | 21 045 (12.6) | 32 760 (13.2) |
| Low income | 103 463 (24.8) | 4165 (22.5) | 42 119 (25.1) | 61 344 (24.6) |
| Low/moderate disability | 79 457 (19.1) | 3562 (19.3) | 31 076 (18.5) | 48 381 (19.4) |
| High disability | 179 818 (43.2) | 8514 (46.1) | 73 401 (43.8) | 106 417 (42.8) |
| Primary care visits in year prior to analysis period (mean [SD]) | 6.9 (6.6) | 7.7 (7.3) | 7.2 (6.7) | 6.8 (6.5) |
| Primary care visits in year prior to analysis period, categorical | ||||
| 0–4 | 184 228 (44.2) | 7212 (39.0) | 70 859 (42.3) | 113 369 (45.5) |
| 5–8 | 124 847 (30.0) | 5613 (30.4) | 51 100 (30.5) | 73 747 (29.6) |
| 9+ | 107 512 (25.8) | 5653 (30.6) | 45 687 (27.3) | 61 825 (24.8) |
| Number of chronic medical conditions (mean [SD]) | 7.5 (3.5) | 7.9 (3.6) | 7.8 (3.5) | 7.3 (3.5) |
| Number of chronic medical conditions, categorical | ||||
| 0–3 | 44 063 (10.6) | 1572 (8.5) | 14 188 (8.5) | 29 875 (12.0) |
| 4–7 | 185 729 (44.6) | 7599 (41.1) | 71 654 (42.7) | 114 747 (29.6) |
| 8–11 | 132 196 (31.7) | 6419 (34.7) | 57 045 (34.0) | 75 151 (30.2) |
| 12+ | 54 599 (13.1) | 2888 (15.6) | 24 759 (14.8) | 29 840 (12.0) |
| At least 1 emergency visit in the year prior to analysis period | ||||
| No | 229 698 (55.1) | 9794 (53.0) | 89 720 (53.5) | 139 978 (56.2) |
| Yes | 186 889 (44.9) | 8684 (47.0) | 77 926 (46.5) | 108 963 (43.8) |
| At least 1 mental health visit in the year prior to analysis period | ||||
| No | 276 971 (66.5) | 11 738 (63.8) | 112 668 (67.2) | 164 303 (66.0) |
| Yes | 139 613 (33.5) | 6665 (36.2) | 54 978 (32.8) | 84 638 (34.0) |
| Video visits per patient (mean [SD]) | 0.1 (0.4) | 1.4 (1.1) | 0.2 (0.6) | 0 (0) |
| Virtual visits per patient (mean [SD]) | 0.8 (1.7) | 2.6 (2.6) | 1.9 (1.7) | 0 (0) |
For continuous variables, figures outside of parentheses represent means and figures in parentheses represent standard deviations.
Some categories may not sum to overall total due to missing data.
Virtual care comprises phone and video care.
Table 3 presents the results of repeated event survival models adjusted for patient and facility characteristics for all virtual care use and video care use specifically; Figure 2 demonstrates a subset of these results in graphical form.
Table 3.
Adjusted hazard ratios and 95% confidence intervals for virtual and video cardiology care use between January 1, 2019 and March 10, 2021 in Veterans in Active Care at the Veterans Health Administration
| Virtual carea |
Video care |
|||
|---|---|---|---|---|
| Hazard ratiob | 95% CI | Hazard ratio | 95% CI | |
| Age (years) (Refc: 18–49) | ||||
| 50–64 | 1.15 | 1.13–1.18 | 1.06 | 0.99–1.14 |
| 65–74 | 1.16 | 1.13–1.18 | 0.94 | 0.88–1.01 |
| 75+ | 1.07 | 1.05–1.09 | 0.80 | 0.75–0.86 |
| Male gender (Ref: female) | 1.04 | 1.02–1.05 | 1.06 | 0.99–1.14 |
| Rurality (Ref: urban) | ||||
| Rural | 0.99 | 0.98–1.00 | 0.87 | 0.83–0.90 |
| Highly rural | 0.99 | 0.97–1.01 | 0.77 | 0.68–0.87 |
| Race (Ref: White) | ||||
| American Indian/Alaska Native | 0.98 | 0.95–1.02 | 0.87 | 0.73–1.04 |
| Asian | 1.00 | 0.96–1.05 | 0.98 | 0.83–1.16 |
| Black or African American | 1.01 | 1.00–1.02 | 1.03 | 0.99–1.07 |
| Native Hawaiian/Pacific Islander | 0.97 | 0.94–1.01 | 0.99 | 0.88–1.10 |
| Unknown | 0.98 | 0.96–1.00 | 1.04 | 0.97–1.11 |
| Ethnicity (Ref: not Hispanic/Latino) | ||||
| Hispanic or Latino | 1.11 | 1.09–1.13 | 0.88 | 0.84–0.93 |
| Unknown | 1.01 | 0.99–1.04 | 1.02 | 0.94–1.10 |
| Enrollment priority (Ref: no specific enrollment priority) | ||||
| Low income | 1.01 | 0.99–1.02 | 0.94 | 0.89–0.98 |
| Low/moderate disability | 1.01 | 1.00–1.03 | 1.01 | 0.96–1.06 |
| High disability | 1.01 | 1.00–1.02 | 1.02 | 0.98–1.07 |
| Homeless (Ref: not homeless) | 0.93 | 0.91–0.95 | 0.85 | 0.80–0.92 |
| Drive time (Ref: short) | ||||
| Medium | 1.04 | 1.03–1.04 | 1.08 | 1.05–1.12 |
| Long | 0.99 | 0.98–1.00 | 0.89 | 0.85–0.93 |
| Chronic conditions (N) (Ref: 0–3) | ||||
| 4–7 | 1.06 | 1.04–1.07 | 1.06 | 1.01–1.12 |
| 8–11 | 1.08 | 1.07–1.10 | 1.19 | 1.12–1.26 |
| 12+ | 1.07 | 1.05–1.08 | 1.28 | 1.19–1.36 |
| Primary care visits (N) (Ref: 0–4) | ||||
| 5–8 | 1.00 | 10.99–1.01 | 1.06 | 1.03–1.10 |
| 9+ | 0.98 | 0.97–0.99 | 1.16 | 1.12–1.20 |
| Emergency use (Ref: no use) | 1.01 | 1.00–1.02 | 1.02 | 0.99–1.05 |
| Mental healthcare use (Ref: no use) | 0.98 | 0.97–0.98 | 0.96 | 0.94–1.00 |
| Observations | 1 669 514 | 1 674 556 | ||
Virtual care comprises phone and video visits.
Models stratified by prepandemic versus during pandemic and adjusted for VISN.
Ref = reference category.
Figure 2.
Adjusted hazard ratios for VA cardiology virtual care and video care use. Labels represent adjusted hazard ratios (and 95% confidence intervals) for a VA Cardiology virtual care (phone or video care, in circle-shaped points) and video care (in diamond-shaped points) visit between January 1, 2019 through March 10th, 2021.
Older patients were more likely than younger patients to receive virtual care overall (hazard ratio [HR] for ages 50–64 1.15, 95% confidence interval [CI] 1.13–1.18; for ages 65–74 1.16, 95% CI 1.13–1.18; for ages 75 and older 1.07, 95% CI 1.05–1.09), but the oldest group of patients was less likely than younger patients to receive video care (HR for ages 75 and older 0.80, 95% CI 0.75–0.86), indicating higher use of telephone care. No differences were seen by gender or race for either virtual care overall or video care; however, Hispanic or Latino patients were more likely than non-Hispanic or Latino Veterans to receive any virtual care (adjusted HR 1.11, 95% CI 1.09–1.13), but less likely to receive video care (adjusted HR 0.88, 95% CI 0.84–0.93).
There were no differences in likelihood of receiving any virtual care by rurality of residence, but rural and highly rural residents were less likely to receive video care compared to urban dwellers (adjusted HRs 0.87, 95% CI 0.83–0.90 and 0.77, 95% CI 0.68–0.87, respectively). Those living at a moderate distance were more likely than those living a short drive from the facility to receive video care, whereas those living a longer drive away were less likely to use video care (HRs 1.08, 95% CI 1.05–1.12 and 0.89, 95% CI 0.85–0.93, respectively). Low-income patients were as likely as those without special VA enrollment priority to receive any virtual care, but less likely to receive video care (HR 0.94, 95% CI 0.89–0.98); those experiencing homelessness were less likely to receive either any virtual care (HR 0.93, 95% CI 0.91–0.95) or video care (0.85, 95% CI 0.80–0.92, respectively). No differences were seen in likelihood of receiving virtual or video care for patients with disability.
Patients with more chronic medical conditions were more likely than those with three or fewer chronic medical conditions to receive either any virtual care or video care specifically, with a stronger effect for video care (HR for video care for those with 4–7 chronic conditions 1.06, 95% CI 1.01–1.12; for those with 8–11 chronic conditions 1.19, 95% CI 1.12–1.26; for those with 12 or more chronic conditions 1.28, 95% CI 1.19–1.36).
Prior use of primary care services was associated with a higher likelihood of receiving video care (adjusted HR for those with 5–8 primary care visits in the year prior to the study period 1.06, 95% CI 1.03–1.10; adjusted HR for those with 9 or more visits 1.16, 95% HR 1.12–1.20). However, prior use of emergency/urgent care or mental health services did not affect likelihood of use of virtual or video care.
In a sensitivity analysis, the Breslow and Efron methods of handling ties produced similar results (Supplementary Table S6).
DISCUSSION
In this study of virtual cardiology care use in VA before and during the COVID-19 pandemic, we find that a rapid conversion to majority virtual care in the early days of the pandemic was accompanied by significant disparities in video visit use that affected older adults and patients who lived in rural areas, had low income, and who were experiencing homelessness, whereas overall virtual care use was more evenly distributed. Patients with more chronic conditions, conversely, were more likely to use video visits and virtual care overall.
Transition to virtual care
The transition from in-person to virtual care in VA for care overall, and specifically in primary care, has been previously documented;5,6 as virtual care was established in primary and mental health care prior to the COVID-19 pandemic, this shift may have been more straightforward for these specialties than for subspecialties in VA, where virtual care use prior to March 2020 was rare. Here we document that at least in cardiology specialty care, and consistent with patterns in other ambulatory cardiology health care systems,36,37 clinics converted the majority of visits to virtual, predominately telephone, within weeks of the pandemic onset. A level change in the total number of cardiology encounters accompanied this transition, remaining at 62.6–76.9% of the average monthly visits preceding the pandemic (Figure 1). This stepwise change to a new plateau is notable and corresponds to a significant proportion of patients (38.1%) without any cardiology visits after the pandemic’s onset. The reasons for this drop-off in visits require further exploration; these patients tended to be younger and have fewer comorbidities, suggesting some may have had a lower need for specialty cardiology care (Supplementary Table S5), but others may have been reluctant to seek cardiology care during the pandemic despite ongoing care needs. Extending this investigation further into the pandemic will help differentiate between these effects and explore the characteristics of patients who lost their connections to care during this transition to virtual care.
Sociodemographic disparities
Our analysis finds that many subpopulations of patients with higher levels of need (those of older age, in more rural areas, with lower income, and experiencing homelessness) were less likely to receive cardiology-specific care by video, though equally likely to receive cardiology care by phone, even after adjusting for number of chronic conditions. Earlier findings examining the expansion of virtual care among all active care VA patients6 found similarly lower use of video visits for older, homeless, and rural Veterans. However, in this study of cardiology patients, we found Veterans with lower income were less likely to access video care, whereas a null association was reported previously.6 Distinct care patterns in cardiology may explain these differences: cardiology patients are on average older and have more medical comorbidities than VA patients overall, and while our analysis adjusts for these characteristics, cardiologists may have a lower threshold than generalists to perform a physical exam and/or imaging or other interventions requiring face-to-face visits for a patient with a given level of medical complexity. Nonetheless, that these more vulnerable groups are less likely to have used video cardiology care in this period raises concern that VA may be falling short of its goal to ensure Veterans most in need of VA care can access it.4 In addition to the above, these disparities may be attributable to patient preferences, barriers to usability or to technology or Internet access, or as yet undefined factors. In particular, individuals with lower socioeconomic status have fewer connections to broadband Internet38 and to smartphones or other devices necessary for video calls,39 though the VA is attempting to intervene at this point with its initiative distributing tablets to high-need Veterans with access barriers.12
Our finding that Hispanic/Latino individuals were less likely to use video care also reinforces concerns that virtual care might increase racial and ethnic disparities in health care access and quality of care.1,40,41 While there could exist interaction between race/ethnicity and other sociodemographic characteristics, as well as between rurality and these sociodemographic characteristics, these disparities persisted after controlling for some of these factors with VA enrollment priority group and VA VISN, which incorporates geography.
Given these observed differences across groups in rates of video versus telephone visits, it is critical to understand the relative value of video visits. Existing research comparing video visits to other visit modalities is limited but suggests that video has the potential to create better bonding between participants and a more personal experience.42,43 If true, the trust that these connections engender could be disproportionately important for historically disenfranchized populations that may be more likely to harbor well-founded mistrust in medical institutions.44–46 Video visits also allow for the collection of nonverbal communication cues and other clinically relevant information.42 The differences identified are therefore concerning and present an opportunity for further exploration and intervention. However, it is equally possible that for at least some patients and clinical scenarios, phone visits may represent a less expensive, more accessible modality than video or in-person visits for specialty care services. Future research comparing utilization and clinical outcomes across visit modalities will be essential to understand whether the differences identified in this work truly represent disparities rather than reflecting different needs or preferences.
Limitations
This study has several limitations. We have focused on the likelihood of virtual care encounters, and have not assessed measures of quality of care. In terms of generalizability, this analysis was conducted among established VA cardiology patients, and as such results may not be applicable to VA patients newly establishing cardiology care after the onset of the COVID-19 pandemic or, more broadly, the general United States population; in particular, women represent a small minority of VA cardiology patients. Incomplete information in our sample also limits the conclusions to be drawn from our results; specifically, VA enrollment priority group is likely an imperfect surrogate for income or overall social risk, and rural dwelling status is only partially associated with limited technology or connectivity and the virtual care barrier this represents.47,48 In addition, while we have classified visit modality as accurately as possible, we were unable to capture instances where video encounters were converted to phone visits or vice versa. However, technology barriers more often cause video visits to be converted to phone visits than the converse, particularly among some of the vulnerable groups of patients found to use less video care in this study.49 Hence, to the extent visit modality was misclassified in our study, we would expect disparities to be even more pronounced in reality. We have not linked visit data in this study to clinicians and so cannot draw definitive conclusions about virtual care use among clinician cadres, though our restricted set of visit “stop codes” and inclusion of only visits with Evaluation and Management CPT codes largely limits our data to visits conducted by licensed independent practitioners (ie, physicians and advanced practice providers). Lastly, we focus on the “demand” side of virtual care use and do not consider the “supply” side. Policy and intervention design must account for high variation in resources and provision of virtual care services across sites in addition to patient characteristics.
Directions for future research
This analysis was designed to identify variation in virtual care use across patient subpopulations; future qualitative and quantitative research will build on these observations to explore possible reasons or mechanisms for the disparities identified, including any potential differences in frequency or quality associated with the modality of care among various cardiac diagnoses. Follow-up studies will link encounters to clinicians to assess patterns of virtual care use among different clinician cadres. Noting the large minority of our cohort (38.1%) without pandemic-era follow-up visits, we will investigate whether clinical outcomes (eg, heart failure hospitalizations) differed between patients with and without follow-up after March 10, 2020. Future work will also consider “supply” side predictors of virtual care provision. In addition, it is critical to move from describing variation to developing action plans to support patients at risk of falling on the wrong side of the digital divide. Existing programs, such as VA’s initiative distributing tablets to high-need Veterans with access barriers,12 have been popular with patients and have demonstrated improved access for mental health care.9,10 Strengthening knowledge of this program and others designed to address deficient broadband access within subspecialty clinics would be a logical starting point to ensure the most vulnerable are not left behind as virtual care becomes, and perhaps remains, a mainstay of subspecialty care. Finally, it is important to assess whether these findings apply to commercially or publicly insured populations. If reimbursement differences for telephone and video visits among these payors promote the use of video visits, and certain groups have less access to the necessary technology for video visits, access to care could suffer.
CONCLUSION
In this analysis of virtual cardiology care in VA before and during the COVID-19 pandemic, certain vulnerable subpopulations of VA cardiology patients were less likely to receive video cardiology care. This suggests that the pandemic worsened the digital divide for cardiology care to the extent that video visits add value over phone visits. Older patients and those living in more rural locations or experiencing homelessness might benefit from targeted interventions to address this digital divide and fulfill the promise of virtual care for access. If, as many believe, virtual care will continue to comprise a large proportion of patient care,50,51 it is all the more critical to ensure that its benefits accrue equitably.
Supplementary Material
ACKNOWLEDGMENTS
The authors thank the Virtual Access QUERI team for supporting this evaluation, including Cindie Slightam, MPH and Camila Chaudhary, MPH, from VA Palo Alto Health Care System, for providing project management support. Preliminary results of this work were presented at the national meeting for the Society of General Internal Medicine, April 20, 2021, and at the AcademyHealth Annual Research Meeting, June 16, 2021. Views expressed are those of the authors and the contents do not represent the views of the US Department of Veterans Affairs or the United States Government.
CONFLICT OF INTEREST STATEMENT
None declared.
Contributor Information
Rebecca L Tisdale, Health Services Research and Development, Center for Innovation to Implementation (Ci2i), VA Palo Alto Health Care System, Menlo Park, CA, USA; Department of Health Policy, Stanford University School of Medicine, Stanford, CA, USA.
Jacqueline Ferguson, Health Services Research and Development, Center for Innovation to Implementation (Ci2i), VA Palo Alto Health Care System, Menlo Park, CA, USA.
James Van Campen, Health Services Research and Development, Center for Innovation to Implementation (Ci2i), VA Palo Alto Health Care System, Menlo Park, CA, USA; Division of Primary Care and Population Health, Department of Medicine, Stanford University School of Medicine, Stanford, CA, USA.
Liberty Greene, Health Services Research and Development, Center for Innovation to Implementation (Ci2i), VA Palo Alto Health Care System, Menlo Park, CA, USA; Division of Primary Care and Population Health, Department of Medicine, Stanford University School of Medicine, Stanford, CA, USA.
Alexander T Sandhu, Division of Cardiovascular Medicine and the Cardiovascular Institute, Department of Medicine, Stanford University School of Medicine, Stanford, CA, USA.
Paul A Heidenreich, Division of Cardiovascular Medicine and the Cardiovascular Institute, Department of Medicine, Stanford University School of Medicine, Stanford, CA, USA; Department of Medicine, VA Palo Alto Health Care System, Palo Alto, CA, USA.
Donna M Zulman, Health Services Research and Development, Center for Innovation to Implementation (Ci2i), VA Palo Alto Health Care System, Menlo Park, CA, USA; Division of Primary Care and Population Health, Department of Medicine, Stanford University School of Medicine, Stanford, CA, USA.
FUNDING
A US Department of Veterans Affairs Quality Enhancement Research Initiative PEI 18-205 (Zulman, PI), by the Veterans Administration (VA) Office of Academic Affairs Advanced Fellowship in Health Services Research (RT), and by the VA Big Data-Scientist Training Enhancement Program (JF); the National Heart, Lung, and Blood Institute 1K23HL151672-01 (ATS). The funders had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.
AUTHOR CONTRIBUTIONS
All authors meet criteria for authorship per ICMJE requirements and made the following contributions: Conception and design: RLT, JF, LG, ATS, PAH, and DMZ. Acquisition, analysis, and interpretation of data: RLT, JF, JVC, LG, and DMZ. Drafting original draft: RLT. Critical revision: RLT, JF, JVC, LG, ATS, PAH, and DMZ. Accountability for accuracy and integrity of work: RLT, JF, JVC, LG, ATS, PAH, and DMZ.
SUPPLEMENTARY MATERIAL
Supplementary material is available at JAMIA Open online.
DATA AVAILABILITY
The data underlying this article cannot be shared publicly due to Veterans’ Health Affairs policy on sharing patient data, that is, for the privacy of individuals that participated in the study.
References
- 1. Lau J, Knudsen J, Jackson H, et al. Staying connected in the COVID-19 pandemic: telehealth at the largest safety-net system in the United States: a description of NYC health + hospitals telehealth response to the COVID-19 pandemic. Health Aff (Millwood) 2020; 39 (8): 1437–42. [DOI] [PubMed] [Google Scholar]
- 2. Wosik J, Fudim M, Cameron B, et al. Telehealth transformation: COVID-19 and the rise of virtual care. J Am Med Inform Assoc 2020; 27 (6): 957–62. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.World Health Organization. WHO Director-General’s opening remarks at the media briefing on COVID-19 - 11 March 2020 [Internet]. [cited April 20, 2021]. https://www.who.int/director-general/speeches/detail/who-director-general-s-opening-remarks-at-the-media-briefing-on-covid-19–-11-march-2020. Accessed December 8, 2022.
- 4. Heyworth L, Kirsh S, Zulman D, Ferguson JM, Kizer KW.. Expanding access through virtual care: the VA’s early experience with Covid-19. NEJM Catal 2020; 11. [Google Scholar]
- 5. Reddy A, Gunnink E, Deeds SA, et al. A rapid mobilization of ‘virtual’ primary care services in response to COVID-19 at Veterans Health Administration. Healthcare (Amst) 2020; 8 (4): 100464. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6. Ferguson JM, Jacobs J, Yefimova M, Greene L, Heyworth L, Zulman DM.. Virtual care expansion in the Veterans Health Administration during the COVID-19 pandemic: clinical services and patient characteristics associated with utilization. J Am Med Inform Assoc 2021; 28 (3): 453–62. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7. Der-Martirosian C, Chu K, Dobalian A.. Use of telehealth to improve access to care at the United States department of veterans affairs during the 2017 Atlantic Hurricane Season. Disaster Med Public Health Prep 2020; 1–5. [DOI] [PubMed] [Google Scholar]
- 8. Der-Martirosian C, Heyworth L, Chu K, Mudoh Y, Dobalian A.. Patient characteristics of VA telehealth users during hurricane Harvey. J Prim Care Community Health 2020; 11: 2150132720931715. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9. Slightam C, Gregory AJ, Hu J, et al. Patient perceptions of video visits using veterans affairs telehealth tablets: survey study. J Med Internet Res 2020; 22 (4): e15682. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10. Jacobs JC, Blonigen DM, Kimerling R, et al. Increasing mental health care access, continuity, and efficiency for veterans through telehealth with video tablets. Psychiatr Serv 2019; 70 (11): 976–82. [DOI] [PubMed] [Google Scholar]
- 11. Jacobs JC, Hu J, Slightam C, Gregory A, Zulman DM.. Virtual savings: patient-reported time and money savings from a VA national telehealth tablet initiative. Telemed J E Health 2020; 26 (9): 1178–83. [DOI] [PubMed] [Google Scholar]
- 12. Zulman DM, Wong EP, Slightam C, et al. Making connections: nationwide implementation of video telehealth tablets to address access barriers in veterans. JAMIA Open 2019; 2 (3): 323–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13. Funderburk CD, Batulis NS, Zelones JT, et al. Innovations in the plastic surgery care pathway: using telemedicine for clinical efficiency and patient satisfaction. Plast Reconstr Surg 2019; 144 (2): 507–16. [DOI] [PubMed] [Google Scholar]
- 14. Martinez KA, Rood M, Jhangiani N, et al. Patterns of use and correlates of patient satisfaction with a large nationwide direct to consumer telemedicine service. J Gen Intern Med 2018; 33 (10): 1768–73. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15. Nord G, Rising KL, Band RA, Carr BG, Hollander JE.. On-demand synchronous audio video telemedicine visits are cost effective. Am J Emerg Med 2019; 37 (5): 890–4. [DOI] [PubMed] [Google Scholar]
- 16. Rogers EM. The digital divide. Convergence 2001; 7 (4): 96–111. [Google Scholar]
- 17. Chang BL, Bakken S, Brown SS, et al. Bridging the digital divide: reaching vulnerable populations. J Am Med Inform Assoc 2004; 11 (6): 448–57. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18. Seifert A. The digital exclusion of older adults during the COVID-19 pandemic. J Gerontol Soc Work 2020; 63 (6-7): 674–6. [DOI] [PubMed] [Google Scholar]
- 19. Ramsetty A, Adams C.. Impact of the digital divide in the age of COVID-19. J Am Med Inform Assoc 2020; 27 (7): 1147–8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20. Eberly LA, Khatana SAM, Nathan AS, et al. Telemedicine outpatient cardiovascular care during the COVID-19 pandemic: bridging or opening the digital divide? Circulation 2020; 142 (5): 510–2. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21. Abnousi F, Kang G, Giacomini J, et al. A novel noninvasive method for remote heart failure monitoring: the EuleriAn video Magnification apPLications In heart Failure studY (AMPLIFY). NPJ Digit Med 2019; 2 (1): 80. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22. Sandhu AT, Goldhaber-Fiebert JD, Owens DK, Turakhia MP, Kaiser DW, Heidenreich PA.. Cost-effectiveness of implantable pulmonary artery pressure monitoring in chronic heart failure. JACC Heart Fail 2016; 4 (5): 368–75. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23. Gorodeski EZ, Goyal P, Cox ZL, et al. Virtual visits for care of patients with heart failure in the era of COVID-19: a statement from the Heart Failure Society of America. J Card Fail 2020; 26 (6): 448–56. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.US Department of Veterans Affairs. 172VA10P2: VHA Corporate Data Warehouse – VA. 79 FR 4377.
- 25. Yu W, Ravelo A, Wagner TH, et al. Prevalence and costs of chronic conditions in the VA health care system. Med Care Res Rev 2003; 60 (3 Suppl): 146S–67S. [DOI] [PubMed] [Google Scholar]
- 26. Fishman P, Von Korff M, Lozano P, Hecht J.. Chronic care costs in managed care. Health Aff (Millwood) 1997; 16 (3): 239–47. [DOI] [PubMed] [Google Scholar]
- 27. Ray GT, Collin F, Lieu T, et al. The cost of health conditions in a health maintenance organization. Med Care Res Rev 2000; 57 (1): 92–109. [DOI] [PubMed] [Google Scholar]
- 28. Yoon J, Chee CP, Su P, Almenoff P, Zulman DM, Wagner TH.. Persistence of high health care costs among VA patients. Health Serv Res 2018; 53 (5): 3898–916. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.US Census Bureau. Urban and Rural [Internet]. 2020. [cited 2021 Feb 26]. https://www.census.gov/programs-surveys/geography/guidance/geo-areas/urban-rural.html. Accessed December 8, 2022.
- 30. Wang ZJ, Dhanireddy P, Prince C, Larsen M, Schimpf M, Pearman G.. 2019. Survey of Veteran Enrollees’ Health and Use of Health Care. 2020; 207.
- 31. Breslow N. Covariance analysis of censored survival data. Biometrics 1974; 30 (1): 89–99. [PubMed] [Google Scholar]
- 32. Efron B. The efficiency of Cox’s likelihood function for censored data. J Am Stat Assoc 1977; 72 (359): 557–65. [Google Scholar]
- 33. Hertz-Picciotto I, Rockhill B.. Validity and efficiency of approximation methods for tied survival times in Cox regression. Biometrics 1997; 53 (3): 1151–6. [PubMed] [Google Scholar]
- 34. Cleves M, Gould W, Gould WW, Gutierrez R, Marchenko Y.. An Introduction to Survival Analysis Using Stata, 2nd ed.College Station, TX: Stata Press; 2008: 398 p. [Google Scholar]
- 35. Strobe I. The Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) statement: guidelines for reporting observational studies. [cited 2021 Apr 20]; https://core.ac.uk/reader/33050540?utm_source=linkout. Accessed December 8, 2022.
- 36. Sanderson KE, Spithoff KD, Corovic M, Langdon KM, Schwalm JD.. An evaluation of cardiology virtual care during the COVID-19 pandemic. CJC Open 2021; 3 (10): 1294–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 37. Drake C, Lian T, Cameron B, Medynskaya K, Bosworth HB, Shah K.. Understanding telemedicine’s “new normal”: variations in telemedicine use by specialty line and patient demographics. Telemed J E-Health 2022; 28 (1): 51–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 38. Pew Research Center. Internet/Broadband Fact Sheet [Internet]. Pew Research Center: Internet, Science & Tech. [cited 2022 May 4]. https://www.pewresearch.org/internet/fact-sheet/internet-broadband/. Accessed December 8, 2022.
- 39. Na L, Sheu JJ.. Health and information disparities among non-adopters of smartphones. Health Policy Technol 2022; 11 (1): 100600. [Google Scholar]
- 40. Nouri S, Khoong EC, Lyles CR, Karliner L.. Addressing equity in telemedicine for chronic disease management during the Covid-19 pandemic. N Engl J Med Catal 2020; 13: 1–13. [Google Scholar]
- 41. Ortega G, Rodriguez JA, Maurer LR, et al. Telemedicine, COVID-19, and disparities: policy implications. Health Policy Technol 2020; 9 (3): 368–71. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 42. Lindsay JA, Hogan JB, Ecker AH, Day SC, Chen P, Helm A.. The importance of video visits in the time of COVID‐19. J Rural Health 2021; 37 (1): 242–5. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 43. Franzosa E, Gorbenko K, Brody AA, et al. “There is something very personal about seeing someone’s face”: provider perceptions of video visits in home-based primary care during COVID-19. J Appl Gerontol 2021; 40 (11): 1417–24. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 44. Sullivan LS. Trust, risk, and race in American medicine. Hastings Cent Rep 2020; 50 (1): 18–26. [DOI] [PubMed] [Google Scholar]
- 45. Crews DC, Purnell TS.. COVID-19, racism, and racial disparities in kidney disease: galvanizing the kidney community response. J Am Soc Nephrol 2020; 31 (8): 1–3. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 46. Bazargan M, Cobb S, Assari S.. Discrimination and medical mistrust in a racially and ethnically diverse sample of California adults. Ann Fam Med 2021; 19 (1): 4–15. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 47. Klee A, Stacy M, Rosenheck R, Harkness L, Tsai J.. Interest in technology-based therapies hampered by access: a survey of veterans with serious mental illnesses. Psychiatr Rehabil J 2016; 39 (2): 173–9. [DOI] [PubMed] [Google Scholar]
- 48. Miller CJ, McInnes DK, Stolzmann K, Bauer MS.. Interest in use of technology for healthcare among veterans receiving treatment for mental health. Telemed J E Health 2016; 22 (10): 847–54. [DOI] [PubMed] [Google Scholar]
- 49. Gray C, Wray C, Tisdale R, Chaudary C, Slightam C, Zulman D.. Factors influencing how providers assess the appropriateness of video visits: interview study with primary and specialty health care providers. J Med Internet Res 2022; 24 (8): e38826. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 50. Artandi M, Thomas S, Shah NR, Srinivasan M.. Rapid system transformation to more than 75% primary care video visits within three weeks at Stanford: response to public safety crisis during a pandemic. N Engl J Med Catal 2020; 1 (2): 11. [Google Scholar]
- 51. Rose L, Tran LD, Asch SM, Vashi A. Assessment of changes in US Veterans Health Administration care delivery methods during the COVID-19 pandemic. JAMA Netw Open 2021; 4 (10): e2129139. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Data Availability Statement
The data underlying this article cannot be shared publicly due to Veterans’ Health Affairs policy on sharing patient data, that is, for the privacy of individuals that participated in the study.