Abstract
Objectives:
Telemedicine became the primary mode of delivering care during the COVID-19 pandemic. We describe the impact of telemedicine on access to care for people with HIV (PWH) by comparing the proportion of PWH engaged in care prior to and during the COVID-19 pandemic.
Design and Methods:
We conducted an observational analysis of patients enrolled in the Johns Hopkins HIV Clinical Cohort, a single-center cohort of patients at an urban HIV subspecialty clinic affiliated with an academic center. Due to the COVID-19 pandemic, the clinic transitioned from in-person to mostly telemedicine visits. We compared patients receiving care in two time periods. The pre-pandemic period included 2,010 people with ≥1 visit scheduled between September 1st 2019 and March 15th 2020. The pandemic period included 1,929 people with ≥1 visit scheduled between March 16th 2020 and September 30th 2020. We determined the proportion of patients completing ≥1 of their scheduled visits during each period.
Results:
Visit completion increased significantly from 88% pre-pandemic to 91% during the pandemic (p=0.008). Visit completion improved significantly for patients age 20–39 (82% to 92%, p<0.001), women (86% to 93%, p<0.001), Black patients (88% to 91%, p=0.002) and patients with detectable viremia (77% to 85%, p=0.06) during the pandemic. Only 29% of people that completed ≥1 telemedicine visit during the pandemic did so as a video (versus telephone) visit.
Conclusions:
During the pandemic when care was widely delivered via telemedicine, visit completion improved among groups with lower pre-pandemic engagement, but most were limited to telephone visits.
Keywords: Telemedicine, Human Immunodeficiency Virus, COVID-19 Pandemic, Patient Engagement, Continuum of Care
Introduction
Optimizing engagement in care is essential to ending the HIV epidemic.1 With access to antiretroviral therapy and consequent durable viral suppression, prognosis among persons with HIV (PWH) dramatically improves2 and HIV transmission risk is reduced.3 As such, the majority of new infections arise from those inadequately engaged in care.4
The rapid scale-up of telemedicine during the COVID-19 pandemic5,6 provides a unique opportunity to investigate telemedicine’s potential as a tool for engagement. Prior to the pandemic, the use of video visits for care delivery was limited mainly to rural areas7–9, prison systems,10 Veterans Affairs systems,9 and niches where chronically stable patients opted into telemedicine.11 There are limited data on how most patients engage in routine HIV care when telemedicine is the default option for clinic visits. Furthermore, it is unknown whether telemedicine can reach groups that have been previously identified as being at risk for missing visits, including patients with history of substance use disorder, mental health disorders, limited social support, or those living in areas of increased unemployment or poverty.12
In this analysis, we describe visit completion during a period of telemedicine adoption early in the pandemic and compare this to visit completion prior to the pandemic when care was delivered exclusively in-person. We describe patient characteristics associated with visit completion in each period and test whether these associations changed with the introduction of telemedicine. Finally, among patients who completed at least one telemedicine visit, we identified demographic and clinical factors associated with probability of completing a video visit, as opposed to telephone visit(s).
Methods
Study sample.
The John G. Bartlett Specialty Practice is a large subspecialty clinic affiliated with the Johns Hopkins Hospital, caring predominantly for PWH or Hepatitis C in East Baltimore. PWH enrolled in continuity care at the clinic who consent to share their data are enrolled into the Johns Hopkins HIV Clinical Cohort (JHHCC).13 The JHHCC extracts data from the electronic medical record on patient characteristics including self-reported age, sex at birth, race, ethnicity, HIV acquisition risk factors, and clinical information including clinical visits, hospitalizations, labs, prescribed medications, and clinical diagnoses including substance use and mental health disorders.
On March 16th, 2020 the clinic transitioned to almost entirely telemedicine visits in response to the onset of the COVID-19 pandemic.14 At the time of the transition, clinic staff attempted to contact all scheduled patients to reschedule their visits to telemedicine encounters. Telemedicine encounters were conducted between patients and providers as either synchronous audio-video encounters or audio-only telephone encounters. Clinicians were provided written instruction on using the electronic platform and were directed to use the telemedicine encounters to recreate in-person visits to the best of their ability. Providers were located either at home or on-site in individual clinic rooms, and were instructed to confirm that patients had the privacy to speak freely at the start of each encounter and understood the privacy risks inherent to a remote visit. All patients were encouraged to schedule a video visit by default, but if they declined or were unable to connect during their visit, the encounter was converted to a telephone call. During the studied pandemic period, a limited number of in-person visits were permitted for acute concerns or initiation of antiretroviral therapy.
This analysis included all cohort participants with at least one scheduled visit between September 1st 2019 and September 30th 2020, a thirteen-month period centered around the transition to telemedicine. Participants were sorted into two groups: a pre-pandemic group and a pandemic group. The pre-pandemic group included participants with at least one visit scheduled from September 1st 2019 to March 15th 2020. The pandemic group included participants with at least one visit scheduled from March 16th 2020 to September 30th 2020. Participants were eligible for inclusion in both groups if they had visits scheduled in both periods.
Outcome.
The primary outcome for patients in each period was the completion of at least one scheduled visit during the period of interest, irrespective of visit type (in-person or telemedicine). Pre-pandemic, all visits were in-person. During the pandemic, visits were primarily telemedicine (video or telephone), although a small number of in-person visits were also included.
Covariates.
Covariates included age, sex at birth, race, ethnicity, and self-reported HIV risk factors (not mutually exclusive): injection drug use (IDU), high-risk heterosexual intercourse, or men who had sex with men (MSM). Age was defined as a categorical variable using 3 intervals: 20–39 years, 40–60 years, and >60 years. We also included recent substance use, duration in care, and viral suppression. Recent substance use was defined based on the most recent medical record abstraction, conducted every 6 months by trained medical chart abstractors to identify medical record evidence of active substance use including physicians’ notes, toxicology screens, and referrals to treatment. Duration in care was measured as years elapsed from a patient’s first visit to the clinic until the start of the pandemic period (March 16th 2020). Viral suppression was defined as having a viral load ≤200 copies/ml at the most recent lab in the year prior to September 1st 2019 for the pre-pandemic group and the year prior to March 16th 2020 for the pandemic group.
Statistical analysis.
We present the proportion of participants that completed ≥1 visit, both pre-pandemic and during the pandemic. We examined risk factors for visit completion during each period using log-binomial regression models fit with generalized estimating equations to account for correlated outcomes due to the same participants possibly contributing records to both periods.15 We report stratified, unadjusted risk ratios with 95% confidence intervals for each covariate above. We used a model with a single term for time period to examine overall differences in visit completion between the two periods for each subgroup. To identify whether risk ratios for the associations between participant characteristics and visit completion differed between the two periods we report p-values for an interaction term between each characteristic and time period. For interaction terms, given the smaller group sizes we defined statistical significance as p<0.1.16,17 We did not adjust our estimates because we thought crude associations present a more realistic description of the associations in this descriptive analysis.18
Secondary analysis.
To contextualize our findings, a secondary analysis was conducted to specifically assess video visit completion. Our study sample was participants who completed ≥1 telemedicine visits from March 16 – September 30, 2020 to identify factors associated with completion of ≥1 video visit. We adjusted for age, race, HIV risk factor and recent substance use, and report adjusted risk ratios from a log-binomial model.
Results
Patient Characteristics:
During the pre-pandemic period (September 1st 2019 to March 15th 2020) 2,010 participants had at least one visit scheduled. During the pandemic period (March 16th 2020 to September 30th 2020) 1,929 participants had at least one visit scheduled. There were 1,834 participants with a visit scheduled in both periods. Given the substantial overlap, the two groups were very similar in distribution of age, gender, race, ethnicity, HIV risk factors and rates of viral suppression (Table 1). The median (interquartile range) numbers of scheduled visits per person during the pre-pandemic and pandemic periods were 2 (2, 4) and 2 (1, 4) respectively.
Table 1:
Demographic and Clinical Characteristics of Participants With A Scheduled Visit in the 6.5 Months Pre- and During- the COVID-19 Pandemic, Johns Hopkins HIV Clinical Cohort
| Pre-Pandemic Sep 1, 2019 – Mar 15, 2020 (n = 2,010) |
Pandemic Mar 16, 2020 – Sep 30, 2020 (n = 1,929) |
|
|---|---|---|
| Age Category | ||
| 20–39 | 282 (14%) | 273 (14%) |
| 40–59 | 992 (49%) | 949 (49%) |
| 60+ | 736 (37%) | 707 (37%) |
| Male | 1,247 (63%) | 1,213 (63%) |
| Race | ||
| White | 379 (19%) | 369 (19%) |
| Black | 1,552 (77%) | 1,482 (77%) |
| Other | 79 (4%) | 78 (4%) |
| Hispanic | 52 (3%) | 53 (3%) |
| HIV Risk Factor | ||
| Men who had Sex with Men | 634 (32%) | 612 (32%) |
| Intravenous Drug Use | 477 (24%) | 441 (23%) |
| High-risk Heterosexual Contact | 1,070 (53%) | 1,031 (53%) |
| Viral Suppression 1 | ||
| Not Virally Suppressed | 146 (7%) | 116 (6%) |
| Virally Suppressed | 1,713 (85%) | 1,665 (86%) |
| No viral load in the past year | 151 (8%) | 148 (8%) |
| Duration of Care at JHHCC | ||
| < 1 year | 74 (4%) | 79 (4%) |
| 1–5 years | 325 (16%) | 302 (16%) |
| 6–10 years | 349 (17%) | 324 (17%) |
| 10+ years | 1,262 (63%) | 1,224 (63%) |
| History of Depression | 824 (41%) | 775 (40%) |
| Recent Heroin Use 2 | 79 (4%) | 72 (4%) |
| Recent Cocaine Use 2 | 134 (7%) | 121 (6%) |
| Recent Hazardous Alcohol Use 2 | 166 (8%) | 152 (8%) |
| Recent Smoking 2 | 672 (33%) | 624 (32%) |
Virally suppressed at most recent lab within the year prior to each study period
As recorded on medical chart review
Pre-Pandemic Period:
Pre-pandemic, 88% of scheduled participants completed at least one visit (Table 2). During that same period, participants age > 60 were 1.14 (95% Confidence Interval (CI): 1.07,1.20) times as likely as those aged 20–39 to complete ≥1 visit (93% vs. 82%). Men were 1.05 (95% CI: 1.01,1.08) times as likely as women to complete ≥1 visit (90% vs. 86%). Participants who were virally suppressed were 1.17 (95% CI: 1.07, 1.28) times as likely as those who were not to complete ≥1 visit (90% vs. 77%).
Table 2:
Proportion of Participants Completing ≥1 Visit During the 6.5-Month Periods before and during the pandemic stratified by patient characteristics and unadjusted risk ratios for visit completion, Johns Hopkins HIV Clinical Cohort
| 6.5 Mo. Pre-Pandemic | 6.5 Mo. During Pandemic | p-value1 | |||
|---|---|---|---|---|---|
| All Participants | 88% (1773/2010) | 91% (1753/1929) | 0.0082 | ||
| Age Category | |||||
| 20–39 | 82% | - | 92% | - | - |
| 40–59 | 87% | 1.06 (1.00, 1.13) | 90% | 0.98 (0.94, 1.02) | 0.032 |
| 60+ | 93% | 1.14 (1.07. 1.20) | 92% | 1.00 (0.96, 1.05) | 0.001 |
| Female | 86% | - | 93% | - | - |
| Male | 90% | 1.05 (1.01, 1.08) | 90% | 0.97 (0.94, 1.00) | 0.001 |
| Race | |||||
| White | 90% | - | 88% | - | - |
| Black | 88% | 0.98 (0.94, 1.02) | 91% | 1.03 (0.99, 1.08) | 0.07 |
| Other | 89% | 0.99 (0.91, 1.08) | 94% | 1.06 (0.99, 1.13) | 0.22 |
| Not Hispanic | 88% | - | 91% | - | |
| Hispanic | 87% | 0.98 (0.88, 1.09) | 94% | 1.04 (0.97, 1.11) | 0.35 |
| HIV Risk Factor | |||||
| Not MSM | 87% | - | 92% | - | |
| MSM | 90% | 1.03 (1.00, 1.07) | 89% | 0.97 (0.94, 1.01) | 0.009 |
| Not IDU | 89% | - | 91% | - | |
| IDU | 86% | 0.97 (0.93, 1.01) | 91% | 1.00 (0.97, 1.04) | 0.19 |
| No High-risk Heterosexual Contact | 89% | - | 90% | - | |
| High-risk Heterosexual Contact | 88% | 0.98 (0.95, 1.01) | 92% | 1.02 (0.99, 1.05) | 0.11 |
| Viral Suppression | |||||
| Not Virally Suppressed | 77% | - | 85% | - | |
| Virally Suppressed | 90% | 1.17 (1.07, 1.28) | 92% | 1.07 (0.99, 1.16) | 0.14 |
| No viral load in the past year | 82% | 1.07 (0.95, 1.20) | 87% | 1.02 (0.93, 1.13) | 0.59 |
| Duration of Care at JHHCC | |||||
| < 1 year | 84% | - | 89% | - | |
| 1–5 years | 85% | 1.02 (0.91, 1.13) | 90% | 1.01 (0.93, 1.11) | 0.95 |
| 6–10 years | 83% | 0.99 (0.89, 1.11) | 90% | 1.01 (0.93, 1.11) | 0.80 |
| 10+ years | 91% | 1.08 (0.98, 1.19) | 92% | 1.03 (0.95, 1.12) | 0.50 |
| No History of Depression | 89% | - | 91% | - | |
| History of Depression | 87% | 0.97 (0.94, 1.01) | 91% | 0.99 (0.97, 1.02) | 0.89 |
| No Recent Heroin Use | 89% | - | 91% | - | |
| Recent Heroin Use | 80% | 0.90 (0.80, 1.01) | 79% | 0.87 (0.77, 0.98) | 0.62 |
| No Recent Cocaine Use | 89% | - | 91% | - | |
| Recent Cocaine Use | 84% | 0.94 (0.87, 1.02) | 87% | 0.95 (0.89, 1.02) | 0.89 |
| No Recent Hazardous ETOH Use | 88% | - | 91% | - | |
| Recent Hazardous ETOH Use | 88% | 1.00 (0.94, 1.06) | 92% | 1.01 (0.96, 1.06) | 0.82 |
| No Recent Smoking | 90% | - | 91% | - | |
| Recent Smoking | 85% | 0.95 (0.92, 0.99) | 90% | 0.98 (0.95, 1.02) | 0.16 |
p-value for interaction between patient characteristic and time period
p-value for the unadjusted effect of the pandemic on visit completion
Pandemic Period:
Among participants scheduled during the pandemic period, 84% had ≥1 scheduled telemedicine visit, while 16% were scheduled exclusively for in-person visits. During this period, 91% of all scheduled participants completed at least one visit, irrespective of visit type (in-person vs. telemedicine). Of those with telemedicine visits scheduled, 99% completed ≥1 telemedicine visit. Of those with only in-person visits scheduled during the pandemic, 48% completed ≥1 visit. In contrast to pre-pandemic, during the pandemic no participant characteristics were statistically significantly associated with visit completion.
Differences between the Pre-Pandemic and Pandemic Periods:
The increase in visit completion from 88% pre-pandemic to 91% during the pandemic was statistically significant (p=0.008). In the pandemic period, there were no differences in visit completion across age groups. This was primarily due to visit completion improvement among younger age groups compared to pre-pandemic. Similarly, disparities in visit completion associated with sex at birth and race were also diminished; visit completion for women increased from 86% pre-pandemic to 93% during the pandemic (p<0.001) and among Black patients it increased from 88% pre-pandemic to 91% during the pandemic (p=0.002). Participants who had detectable viremia had an increase in visit completion from 77% pre-pandemic to 85% during the pandemic (p=0.06). In the appendix we present adjusted risk ratios, which are largely similar.
Factors Associated with Video Visits:
Among 1,600 participants who completed at least one telemedicine visit, only 468 (29%) completed ≥1 visit using a video visit as opposed to a telephone visit (Table 3). We adjusted for age, sex at birth, HIV risk factor and recent heroin or cocaine use in our subgroup analyses. Participants age ≥60 years were 0.60 [95% CI: 0.49, 0.75] times as likely as those aged 20–39 to complete a video visit (23% vs. 42%). Male participants were 0.76 [95% CI: 0.62, 0.94] times as likely as female participants to complete a video visit. Video visit completion among Black participants was 0.60 [95% CI: 0.42, 0.87] times as likely as among white participants (25% vs. 46%). Participants with IDU as a risk factor for HIV acquisition were 0.62 [95% CI: 0.47, 0.81] times as likely as other participants to complete a video visit (15% vs. 34%). Those with recent heroin or cocaine use were 0.53 [95% CI: 0.32, 0.89] times and those with recent smoking were 0.70 [95% CI: 0.58, 0.85] times as likely as those without to complete a video visit.
Table 3:
Associations between participant characteristics and completion of ≥1 Video Visit, among those who Completed ≥1 Telemedicine visit during the COVID-19 pandemic, Johns Hopkins HIV Clinical Cohort, March 15, 2020 - September 30, 2020
| N (row percent) | ||||
|---|---|---|---|---|
| Phone Visit Only | At least 1 video visit | Unadjusted Risk Ratio for Video | Adjusted Risk Ratioa for Video | |
| Total | 1312 (71%) | 468 (29%) | - | - |
| Age Category | ||||
| 20–39 | 137 (60%) | 91 (40%) | - | - |
| 40–59 | 533 (69%) | 237 (31%) | 0.77 (0.64, 0.93) | 0.82 (0.67, 1.00) |
| 60+ | 462 (77%) | 172 (23%) | 0.58 (0.47, 0.72) | 0.60 (0.49, 0.75) |
| Female | 449 (73%) | 166 (27%) | - | - |
| Male | 683 (69%) | 302 (31%) | 1.14 (0.97, 1.33) | 0.76 (0.62, 0.94) |
| Race | ||||
| White | 159 (54%) | 136 (46%) | - | - |
| Black | 927 (75%) | 311 (25%) | 0.54 (0.47, 0.64) | 0.62 (0.52, 0.73) |
| Other | 46 (69%) | 21 (31%) | 0.68 (0.47, 0.99) | 0.60 (0.42, 0.87) |
| Not Hispanic | 1,100 (71%) | 454 (29%) | - | - |
| Hispanic | 32 (70%) | 14 (30%) | 1.04 (0.69, 1.62) | 0.85 (0.43, 1.71) |
| HIV Risk Factor | ||||
| Not MSM | 848 (77%) | 257 (23%) | - | - |
| MSM | 284 (57%) | 211 (43%) | 1.83 (1.58, 2.13) | 1.49 (1.27, 1.74) |
| Not IDU | 821 (66%) | 414 (34%) | - | - |
| IDU | 311 (85%) | 54 (15%) | 0.44 (0.34, 0.57) | 0.62 (0.47, 0.81) |
| Not Hetero | 475 (64%) | 262 (36%) | - | - |
| Heterosexual Contact | 657 (76%) | 206 (24%) | 0.67 (0.58, 0.78) | 1.01 (0.81, 1.24) |
| No Depression | 674 (71%) | 275 (29%) | - | - |
| History of Depression | 458 (70%) | 193 (30%) | 1.02 (0.88, 1.19) | 1.12 (0.96, 1.30) |
| Viral Suppression | ||||
| Not Virally Suppressed | 64 (74%) | 23 (26%) | - | |
| Virally Suppressed | 990 (71%) | 408 (29%) | 1.10 (0.77, 1.58) | 1.05 (0.74, 1.49) |
| No lab in the past year | 78 (68%) | 37 (32%) | 1.22 (0.78, 1.89) | 1.12 (0.73, 1.70) |
| Duration of Care at JHHCC | ||||
| 1st year | 42 (68%) | 20 (32%) | - | - |
| 1–5 years | 163 (65%) | 89 (35%) | 1.09 (0.74, 1.63) | 1.26 (0.86, 1.85) |
| 6–10 years | 190 (73%) | 71 (27%) | 0.84 (0.56, 1.27) | 1.09 (0.73, 1.64) |
| 10+ years | 737 (72%) | 288 (28%) | 0.87 (0.60, 1.27) | 1.23 (0.84, 1.81) |
| No Recent Cocaine or Heroin Use | 1,032 (69%) | 455 (31%) | - | - |
| Recent Cocaine or Heroin Use | 100 (86%) | 3 (12%) | 0.38 (0.22, 0.63) | 0.53 (0.32, 0.89) |
| No Recent Hazardous Alcohol Use | 1,046 (71%) | 432 (29%) | - | - |
| Recent Hazardous Alcohol Use | 86 (70%) | 36 (30%) | 1.00 (0.76, 1.34) | 1.11 (0.84, 1.47) |
| No Recent Smoking | 726 (67%) | 364 (33%) | - | - |
| Recent Smoking | 406 (80%) | 104 (20%) | 0.61 (0.50, 0.74) | 0.70 (0.58, 0.85) |
Adjusted for age, race, MSM risk factor, and recent substance use.
Discussion
During the first 6.5 months of the COVID-19 pandemic when care was predominantly delivered via telemedicine, overall visit completion was higher than during the pre-pandemic comparison period. This improvement was concentrated among populations with lower pre-pandemic visit completion: younger patients, women, patients whose most recent viral load was not suppressed, patients who had recently established care, or non-MSM patients. While many patients engaged with telemedicine during the pandemic, 71% of their telemedicine visits were telephone visits. Patients who were older, male, Black, or had a history of substance use disorder were most likely to have a telephone rather than video visit. The effect of telephone visits compared to video visits or in-person visits on quality of care is unknown; future studies are needed to investigate the differential impact of these engagement modalities.
Groups with lower probabilities of in-person visit completion pre-pandemic saw the greatest improvement in visit completion with telemedicine during the pandemic. While older participants maintained high levels of visit completion in both study periods, younger participants significantly improved visit completion during the pandemic period. This may reflect high technological literacy among younger patients,19,20 but the appeal of telemedicine compared to in-person visits among this group warrants more study. Similarly, women saw an improvement in visit completion during the pandemic, while men’s visit completion rate was unchanged. This may be due to telemedicine’s ability to mitigate barriers to care that disproportionately impact female patients such as transportation and the burden of caregiving.21,22 Those who were not virologically suppressed also saw improved visit completion during the pandemic, while those who were suppressed maintained the same high pre-pandemic visit completion. The improved visit completion in various subgroups suggests that telemedicine may overcome barriers to care faced by populations historically at risk for missing visits.
Although telemedicine was associated with higher visit completion, a significant proportion of participants used telephone rather than video for their visits. Telephone-only visits were more common in patients who were older, male, had substance use disorder or were Black. The higher use of telephone over video visits among our participants is likely multifactorial, including inconsistent access to a high-speed internet connection or a video-capable device (smartphone, tablet, or computer).5,23–25 This “digital divide” may partly be a function of socioeconomic status,26,27 which has an outsize effect on PWH, a population disproportionately affected by social determinants of health.5,23,28 Our telephone use findings mirror demographics across the digital divide, with older patients and racial minorities at higher risk for limited internet and computer access.5,29,30 In our study, older individuals had higher use of telephone over video visits which may be explained by the lower rates of technological literacy19,20 and disparities in computer and internet access seen in this group29–31 Our findings of lower video visit use among Black participants are consistent with emerging data on this topic,29,31,32 which may also be due to the disparate impact of the digital divide.33 This mirrors a wide array of racial disparities seen across the healthcare system, rooted in the interlocking systems of structural racism. This may include but is not limited to sequelae of historically segregated housing policies, employment discrimination, wage disparities, and other factors that create and compound socioeconomic disparities.34 This in turn likely translates to reduced access to the technology and infrastructure needed for video visits.35 Participants with substance use disorder were also less likely to use video visits, which warrants further study given the increasing interest in telemedicine as a modality for providing substance use treatment, both before and during the pandemic.36 The effect of telephone compared to video visits is not yet fully understood and may negatively impact the quality of care for these patients, because video visits provide for a more personal encounter and allow some degree of visual examination.5 The findings also have meaningful implications for access to care. Prior to the pandemic, telephone visits were not consistently reimbursed at the same rate as video visits.5 If we revert to this payment model, this may disincentivize clinics from offering telephone visits thereby limiting telemedicine access to some groups and exacerbating disparities in care.
The groups scheduled for a visit in both study periods have a large number of overlapping patients. This mitigates the probability that our results are biased by unmeasured, time-fixed confounders (since patients are largely acting as comparisons for themselves). Historically, telemedicine has been implemented selectively in populations where it was anticipated to improve access to care, such as rural settings where the distance to care was a factor, correctional facilities where specialist care was limited, or patients who opt-in due to technological literacy and access.7,11,37–39 This pandemic is the first time that telemedicine has been adopted this widely, creating a unique opportunity to understand if its effects on engagement vary across subgroups. The findings can be used to inform the implementation of telemedicine beyond the pandemic, by recognizing what populations are at risk for disparate telemedicine access and tailoring interventions to mitigate those effects or continuing telephone visit reimbursement at levels comparable to video visits.
We note several caveats and limitations in this analysis. Given the high baseline visit completion, the maximum possible observable risk ratios for this outcome are capped at 1.14. This would correspond to visit completion of 100%, relative to 88% pre-pandemic. Thus, despite the modest magnitude of the presented ratios, we interpret them as representing meaningful gains among the 12% of patients who were missing visits pre-pandemic. It is important to note however that the transition to telemedicine occurred due to the pandemic, thus we are unable separate the effects of conversion to telemedicine from the effects of the pandemic. This should frame the interpretation of the data, and raises several important points. Telemedicine visit completion rates during this early period of the pandemic may not reflect rates during other periods of time when there is not an abrupt conversion of in-person visits to telemedicine visits. Our data were also collected from a single-site, in an urban setting at an academic center in a comparatively resource rich setting. Any generalizations of our findings made beyond this context need to be done so cautiously, given the heterogeneity of patient populations and means of implementation of telemedicine across the world. Additionally, any future implementation of telemedicine beyond the pandemic would likely be complementary to, rather than a replacement of, in-person care. Importantly, this work only analyzes visit completion and does not evaluate the quality of care provided via telephone versus video versus in-person visits. Future studies will be needed to determine if engagement through telemedicine translates into viral suppression, in the same way in-person engagement does,40 which would have implications for disease mortality2 and infection transmission.3 Questions around the impact of telemedicine apply not only to viral suppression, but to the management of comorbid conditions, prevention measures such as screening and counseling, and overall quality of care. This includes the effect of telemedicine on privacy, which exceeds the scope of this study but remains a critical consideration given the stigma associated with HIV. Despite these potential limitations, telemedicine visits are still a form of engagement, which is valuable for patients who would otherwise be largely disengaged when only in-person care is available.5 Our findings demonstrate that telemedicine has the potential to augment access to care for these patients, by providing an additional avenue for engagement. Moving forward, if telemedicine is incorporated into a mixed model of care combining remote and in-person visits to maximize engagement it can be used a means to reach a wider segment of our patient population.
Conclusion
Telemedicine uptake beyond the pandemic will likely persist at higher than pre-pandemic levels, so understanding telemedicine’s effect on engagement is critical to continued implementation. In this study, telemedicine improved visit completion among PWH, and had the greatest impact on groups who historically have not fully engaged with in-person care. However, many patients were limited to telephone rather than video visits, particularly among certain subgroups. The effect of either video or telephone visits on the quality of care compared to in-person visits is not yet known, and warrants further study but it is plausible that at least some engagement via telemedicine (even telephone) improves care for patients who would otherwise be fully disengaged from in-person care. Disengaged patients account for most new HIV infections,1 so successfully engaging them in care is crucial to ending the HIV epidemic. While telemedicine is unlikely to replace in-person care, capitalizing on the wider uptake of telemedicine offers a promising approach to improve engagement for PWH. Successful implementation will require ongoing study of equitable approaches to telemedicine delivery and its impact on clinical outcomes in different populations.
Supplementary Material
Acknowledgements & Funding:
All authors have contributed significantly to this work and have approved of the manuscript as submitted. This work was supported by grants from the National Institutes of Health [K24 AA027483, K01 AA028193, K08 MH113094, U01 DA036935 and P30 AI094189].
References
- 1.Shah M, Perry A, Risher K, et al. Effect of the US National HIV/AIDS Strategy targets for improved HIV care engagement: A modelling study. Lancet HIV. 2016;3(3):e140–e146. doi: 10.1016/S2352-3018(16)00007-2 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Poorolajal J, Hooshmand E, Mahjub H, Esmailnasab N, Jenabi E. Survival rate of AIDS disease and mortality in HIV-infected patients: a meta-analysis. Public Health. 2016;139:3–12. doi: 10.1016/j.puhe.2016.05.004 [DOI] [PubMed] [Google Scholar]
- 3.Eisinger RW, Dieffenbach CW, Fauci AS. HIV viral load and transmissibility of HIV infection undetectable equals untransmittable. JAMA - J Am Med Assoc. 2019;321(5):451–452. doi: 10.1001/jama.2018.21167 [DOI] [PubMed] [Google Scholar]
- 4.Li Z, Purcell DW, Sansom SL, Hayes D, Hall HI. Vital Signs: HIV Transmission Along the Continuum of Care — United States, 2016. MMWR Morb Mortal Wkly Rep. 2019;68(11):267–272. doi: 10.15585/mmwr.mm6811e1 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Wood BR, Young JD, Abdel-Massih RC, et al. Advancing Digital Health Equity: A Policy Paper of the Infectious Diseases Society of America and the HIV Medicine Association. Clin Infect Dis. 2021;72(6):913–919. doi: 10.1093/cid/ciaa1525 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Dandachi D, Freytag J, Giordano TP, Dang BN. It is Time to Include Telehealth in Our Measure of Patient Retention in HIV Care. AIDS Behav. 2020;1:1. doi: 10.1007/s10461-020-02880-8 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Jain KM, Bhat P, Maulsby C, et al. Extending access to care across the rural US south: Preliminary results from the Alabama eHealth programme. J Telemed Telecare. 2019;25(5):301–309. doi: 10.1177/1357633X18755227 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Saifu HN, Asch SM, Goetz MB, et al. Evaluation of human immunodeficiency virus and hepatitis C telemedicine clinics. Am J Manag Care. 2012;18(4):207–212. Accessed July 3, 2020. https://pubmed.ncbi.nlm.nih.gov/22554009/ [PubMed] [Google Scholar]
- 9.Ohl M, Dillon D, Moeckli J, et al. Mixed-methods evaluation of a telehealth collaborative care program for persons with HIV infection in a rural setting. J Gen Intern Med. 2013;28(9):1165–1173. doi: 10.1007/s11606-013-2385-5 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Young JD, Patel M, Badowski M, et al. Improved virologic suppression with HIV subspecialty care in a large prison system using telemedicine: An observational study with historical controls. Clin Infect Dis. 2014;59(1):123–126. doi: 10.1093/cid/ciu222 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.León A, Cáceres C, Fernández E, et al. A new multidisciplinary home care telemedicine system to monitor stable chronic human immunodeficiency virus-infected patients: A randomized study. PLoS One. 2011;6(1). doi: 10.1371/journal.pone.0014515 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Tarantino N, Brown LK, Whiteley L, Fernández MI, Nichols SL, Harper G. Correlates of missed clinic visits among youth living with HIV. AIDS Care - Psychol Socio-Medical Asp AIDS/HIV. 2018;30(8):982–989. doi: 10.1080/09540121.2018.1437252 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Moore RD. Understanding the clinical and economic outcomes of HIV therapy: The Johns Hopkins HIV clinical practice cohort. J Acquir Immune Defic Syndr Hum Retrovirology. 1998;17(SUPPL. 1):S38–41. doi: 10.1097/00042560-199801001-00011 [DOI] [PubMed] [Google Scholar]
- 14.Nitkin K Milestones in a Year of COVID-19 at Johns Hopkins Medicine. Published March 8, 2021. Accessed May 14, 2021. https://www.hopkinsmedicine.org/news/articles/covid-19-timeline-for-johns-hopkins-medicine
- 15.Hanley JA. Statistical Analysis of Correlated Data Using Generalized Estimating Equations: An Orientation. Am J Epidemiol. 2003;157(4):364–375. doi: 10.1093/aje/kwf215 [DOI] [PubMed] [Google Scholar]
- 16.Durand CP. Does Raising Type 1 Error Rate Improve Power to Detect Interactions in Linear Regression Models? A Simulation Study. PLoS One. 2013;8(8):e71079. doi: 10.1371/journal.pone.0071079 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Marshall SW. Power for tests of interaction: Effect of raising the Type I error rate. Epidemiol Perspect Innov. 2007;4(1):1–7. doi: 10.1186/1742-5573-4-4 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Kaufman JS. Statistics, adjusted statistics, and maladjusted statistics. Am J Law Med. 2017;43(2–3):193–208. doi: 10.1177/0098858817723659 [DOI] [PubMed] [Google Scholar]
- 19.Estacio EV, Whittle R, Protheroe J. The digital divide: Examining socio-demographic factors associated with health literacy, access and use of internet to seek health information. J Health Psychol. 2019;24(12):1668–1675. doi: 10.1177/1359105317695429 [DOI] [PubMed] [Google Scholar]
- 20.Moore AN, Rothpletz AM, Preminger JE. The effect of chronological age on the acceptance of internet-based hearing health care. Am J Audiol. 2015;24(3):280–283. doi: 10.1044/2015_AJA-14-0082 [DOI] [PubMed] [Google Scholar]
- 21.Caregiving in America | The National Alliance for Caregiving. Accessed May 7, 2021. https://www.caregiving.org/research/caregivingusa/
- 22.Addressing Gender Disparities in Transportation - Gender Policy Report. Accessed May 7, 2021. https://genderpolicyreport.umn.edu/addressing-gender-disparities-in-transportation/
- 23.Lesko CR, Bengtson AM. HIV and COVID-19: Intersecting Epidemics with Many Unknowns. Am J Epidemiol. 2021;190(1):10–16. doi: 10.1093/aje/kwaa158 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Demographics of Mobile Device Ownership and Adoption in the United States | Pew Research Center. Accessed April 16, 2021. https://www.pewresearch.org/internet/fact-sheet/mobile/ [Google Scholar]
- 25.Velasquez D, Mehrotra A. Ensuring The Growth Of Telehealth During COVID-19 Does Not Exacerbate Disparities In Care | Health Affairs. Health Affairs Blog. Accessed May 15, 2021. https://www.healthaffairs.org/do/10.1377/hblog20200505.591306/full/ [Google Scholar]
- 26.Wong C, Gange SJ, Moore RD, et al. Multimorbidity among Persons Living with Human Immunodeficiency Virus in the United States. Clin Infect Dis. 2018;66(8):1230–1238. doi: 10.1093/cid/cix998 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Fang ML, Canham SL, Battersby L, Sixsmith J, Wada M, Sixsmith A. Exploring Privilege in the Digital Divide: Implications for Theory, Policy, and Practice. Gerontologist. 2019;59(1):E1–E15. doi: 10.1093/geront/gny037 [DOI] [PubMed] [Google Scholar]
- 28.Pellowski JA, Kalichman SC, Matthews KA, Adler N. A pandemic of the poor: Social disadvantage and the U.S. HIV epidemic. Am Psychol. 2013;68(4):197–209. doi: 10.1037/a0032694 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.Lam K, Lu AD, Shi Y, Covinsky KE. Assessing Telemedicine Unreadiness among Older Adults in the United States during the COVID-19 Pandemic. JAMA Intern Med. 2020;180(10):1389–1391. doi: 10.1001/jamainternmed.2020.2671 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.Roberts ET, Mehrotra A. Assessment of Disparities in Digital Access among Medicare Beneficiaries and Implications for Telemedicine. JAMA Intern Med. 2020;180(10):1386–1389. doi: 10.1001/jamainternmed.2020.2666 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31.Census Bureau U Computer and Internet Use in the United States: 2016. Accessed May 15, 2021. www.census.gov/acs
- 32.Rodriguez JA, Betancourt JR, Sequist TD, Ganguli I. Differences in the use of telephone and video telemedicine visits during the COVID-19 pandemic. Am J Manag Care. 2021;27(1):21–26. doi: 10.37765/ajmc.2021.88573 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33.African Americans and Technology Use | Pew Research Center. Accessed May 8, 2021. https://www.pewresearch.org/internet/2014/01/06/african-americans-and-technology-use/ [Google Scholar]
- 34.Bailey ZD, Feldman JM, Bassett MT. How Structural Racism Works — Racist Policies as a Root Cause of U.S. Racial Health Inequities. Malina D, ed. N Engl J Med. 2020;384(8):768–773. doi: 10.1056/nejmms2025396 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35.Lower-income Americans still lag in tech adoption | Pew Research Center. Accessed April 16, 2021. https://www.pewresearch.org/fact-tank/2019/05/07/digital-divide-persists-even-as-lower-income-americans-make-gains-in-tech-adoption/ [Google Scholar]
- 36.Oesterle TS, Kolla B, Risma CJ, et al. Substance Use Disorders and Telehealth in the COVID-19 Pandemic Era: A New Outlook. Mayo Clin Proc. 2020;95(12):2709–2718. doi: 10.1016/j.mayocp.2020.10.011 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 37.Rogers BG, Coats CS, Adams E, et al. Development of Telemedicine Infrastructure at an LGBTQ+ Clinic to Support HIV Prevention and Care in Response to COVID-19, Providence, RI. AIDS Behav. Published online 2020. doi: 10.1007/s10461-020-02895-1 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 38.Coombes CE, Gregory ME. The Current and Future Use of Telemedicine in Infectious Diseases Practice. Curr Infect Dis Rep. 2019;21(11). doi: 10.1007/s11908-019-0697-2 [DOI] [PubMed] [Google Scholar]
- 39.Gras G Use of telemedicine in the management of infectious diseases. Med Mal Infect. 2018;48(4):231–237. doi: 10.1016/j.medmal.2018.01.005 [DOI] [PubMed] [Google Scholar]
- 40.Mugavero MJ, Westfall AO, Zinski A, et al. Measuring retention in HIV care: The elusive gold standard. In: Journal of Acquired Immune Deficiency Syndromes. Vol 61. J Acquir Immune Defic Syndr; 2012:574–580. doi: 10.1097/QAI.0b013e318273762f [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.