Abstract
Background
Existing studies have documented the negative impact of the COVID-19 pandemic on the utilization of health services among people living with HIV (PLWH), including reduced access to medical consultations, antiretroviral therapy (ART) uptake, and viral load monitoring. However, few studies have investigated the pandemic’s effect on clinical outcomes—such as viral suppression rates—particularly within U.S. settings. Using a nationwide data source, this study aims to examine the impact of the COVID-19 pandemic on healthcare visits and viral suppression outcomes of PLWH.
Methods
This retrospective cohort study used data from All of Us (AoU) Research Program, a nationwide cohort designed to be broadly representative of the U.S. population, particularly groups historically underrepresented in biomedical research. Data from more than 340 sites, deposited into AoU and harmonized into a release set spanning January 2018 through December 2021, were used for this study. Using CD4 count or viral load measures as proxies for health care visits, we employed generalized linear mixed regression interrupted time series (ITS) models with a log link to assess the impact of the COVID-19 pandemic on the number of healthcare visits and viral suppressions, comparing data two years before and after the onset of the pandemic. Among 7,667 PLWH identified, 1,772 who had at least one CD4 count or viral load measure were included in the healthcare visit analysis, and 1,382 who had at least one viral load measure were included in the viral suppression analysis.
Results
For the outcomes of interest, the total number of healthcare visits dropped from 7,902 (average of 4.90 per person) before pandemic to 5,252 (average of 3.70 per person) after the pandemic. Similarly, the total number of viral suppression events declined from 2,361 (average of 1.98 per person) before pandemic to 1,893 (average of 1.88 per person) after pandemic. Compared to the pre-COVID-19 period, the number of healthcare visits among PLWH significantly decreased (adjusted odds ratio [aOR] = 0.80, 95%CI: 0.75–0.86) whereas the number of viral suppression events did not change significantly (aOR = 1.10, 95%CI: 0.97–1.25) after the pandemic. Older individuals had lower odds of healthcare visits, while those with Medicaid insurance and individuals with more comorbidities were more likely to have healthcare visits. Regarding viral suppression, older individuals, Hispanic or Latino participants, and those who were married or living with partners were more likely to achieve viral suppression, whereas individuals with a greater number of comorbidities were less likely to achieve viral suppression.
Conclusions
The COVID-19 pandemic disrupted healthcare visits and contributed to suboptimal viral suppression among underrepresented HIV populations. Targeted interventions are needed to ensure continuous engagement in HIV care during unprecedented health crises.
Keywords: HIV/AIDS, Service interruption, Viral suppression, COVID-19
Introduction
Since the World Health Organization declared COVID-19 a pandemic, governments have urgently implemented non-pharmaceutical interventions such as lockdowns, household confinements, workplace restructuring, and cancellations of mass gatherings [1]. These measures caused widespread disruptions to healthcare services, including HIV care, due to the increased demand on overall health system capacity and the diversion of health care personnel to COVID-19 related activities [2]. The impact of the COVID-19 pandemic on healthcare utilization and clinical outcomes among people living with HIV (PLWH) has been reported in many countries, including the US, Italy, China, and sub-Saharan African countries [3–9]. While a growing body of global literature has documented interruptions in HIV services during the pandemic [10], evidence specific to PLWH in US settings remains limited.
Although the impact of adverse COVID-19 outcomes among PLWH has been well documented, research remains insufficient to fully capture the broader implications of COVID-19 on healthcare utilization in this population. Sustained retention in care is essential for PLWH to achieve viral suppression and improve physical outcomes [11, 12]. During the COVID-19 pandemic, worse HIV care engagement outcomes were observed in both modelling and observational studies as a result of the service interruptions [7, 10, 13]. According to a systematic review [14], most studies reported a decline in ART supply and uptake amid the pandemic (ranging from 5% to 26%), which compromise the ART adherence. In the US, while the use of telehealth has been encouraged by the Health Resources and Services Administration (HRSA) Ryan White Program during the COVID-19 pandemic, this option remains difficult for many PLWH due to limited access to technology [15, 16]. In US studies, while decreased in-person clinical visits, viral suppression or viral load testing were observed [17–19], one study conducted in Veterans observed undisrupted ART access and sustained viral suppression rate potentially due to increased virtual visits [17]. Another US study in Washington DC, assessed the impact of the pandemic on HIV care delivery by examining HIV clinic-level services. This study found that clinics experienced an increase in temporary clinic closures, reduction in clinic hours, increased telehealth utilization, adoption of multi-month dispensation of ART, and alternative medication delivery methods [20]. Existing studies on the pandemic’s impact on HIV care often rely on cross-sectional data and non-representative samples (e.g., male veterans, single-site cohorts), which limits generalizability, especially for women, minorities, and rural populations [21]. In addition, regional and system-level differences have produced inconsistent findings regarding care engagement and viral suppression. Longitudinal analyses of HIV care patterns before and during the pandemic are therefore needed to better understand its long-term effects on underrepresented PLWH.
Racial and ethnic minorities have been disproportionately affected by HIV. These disparities worsened during the COVID-19 pandemic, which also disproportionately impacted these groups. The compounded effects of both epidemics underscore persistent structural inequities and the need for targeted public health efforts. Given the sparce evidence of the healthcare service interruptions for PLWH, this study aims to examine the impact of the COVID-19 pandemic on healthcare utilizations and viral suppression for underrepresented PLWH, using the national data harmonized into the All of Us (AoU) research program.
Methods
Data sources
We used data in the AoU research platform for analysis. The AoU research program is a national NIH effort which currently tries to enroll all US residents who are 18 years of age or older from a network of more than 340 recruitment sites across the US. Specifically, the AoU Research Program aims to recruit eligible participants across the US covering more than 80% individuals (~ 1 million) that have been and continue to be underrepresented in biomedical research [22]. Specifically, this program defines individuals with inadequate access to medical care, low household income, low education attainment, and racial or sexual and gender minorities as ‘underrepresented population’ [23]. AoU started the participants recruitment in May 2018 and it harmonizes both electronic health record (EHR) data and survey data from a diverse regions across the US [22]. The description of the recruitment methods and scientific rationale for AoU have been published elsewhere [22]. All data and analyses were conducted using R Studio within a secure, cloud-based platform. This platform enables users to access and analyze a centralized version of AoU data. The AoU data are available in three tiers: the Public Tier, which includes only aggregate data; the Registered Tier, which contains individual-level EHR, wearable device data, and survey responses; and the Controlled Tier, which includes all Registered Tier data plus genomic data. We exclusively used Controlled Tier data from release version 7, which includes information for all participants enrolled in the program between May 18, 2018 to June 1, 2022 [22]. This data extraction did not involve any data merging process.
Study participants and outcome measures
The study population was PLWH who were aged > 18 years old and had at least one CD4 count or viral load measure before and after COVID-19 pandemic. We first identified PLWH using computational phenotyping, which has published elsewhere [24]. Then we restricted to PLWH who had at least one healthcare encounter or had at least one CD4 count or viral load measure between March 2018 (given that some contributing sites may have provided retrospective data collected prior to the program’s launch) to the most recent record in the platform (i.e., December 2021). In accordance with the literature [25], outcomes and predictors were measured longitudinally in each 6-month interval starting from March 2018. In each time window, we defined the occurrence of viral suppression (VL < 50 copies/ml) (i.e., outcome) and all the time-dependent predictors (COVID-19 outbreak and infection). Then the count of viral suppression records across different time windows was used as the outcome. Similar time alignment was used for the number of healthcare visits. Healthcare visits were defined as either CD4 count or viral load testing records during a 6-month time interval. The count of healthcare visits was used as another outcome.
Exposure measures and covariates
To examine the impact of the COVID-19 pandemic on healthcare visits and viral suppression outcome, we defined both COVID-19 pandemic indicator and COVID-19 infection status. We used March 1, 2020 as the cutoff time point to indicate the beginning of the outbreak and differentiated the time window into two periods (i.e., before [March 2018-March 2020] or after the pandemic [March 2020-August 2021]). The difference in the frequency of health care visit and viral suppression in the two time periods were calculated and compared. COVID-19 infection was defined by both EHR (ICD-10 codes: B97.21, B97.29, U07.1, U09.9, J12.82, M35.81) and survey data [26]. If one individual has conflicts of the responses of COVID-19 infection status in the two data sources, we used the data in EHR.
Covariates included time-independent demographics, such as sex, age, race, ethnicity group, marital status, sexual orientation, education, income level, insurance status. We also included the time-independent comorbidities (e.g., cardiovascular disease, chronic pulmonary disease, diabetes), which were defined by ICD codes with any diagnosis on or before March 2018. The number of comorbidities was counted and used for the analysis, which was categorized into three groups (i.e., 0, 1, ≥2).
Statistical analysis
We summarized the distributions of demographic and clinical characteristics of the study cohort using means and standard deviations for continuous variables, and counts with percentages for categorical variables. To examine the longitudinal impact of the COVID-19 pandemic on healthcare visits and viral suppression (VS) events, we employed generalized linear mixed effects interrupted time series (ITS) models with a log link (i.e. Poisson mixed effects ITS models). Each participant contributed repeated measures across consecutive 6-month intervals from March 2018 to December 2021. The general form of the model was:
 |
where 
is the expected count of events for individual i at time interval t, 
is the subject-specific random intercept, time is a continuous variable indicating the number of 6-month periods since March 2018, representing the underlying pre-intervention trend, covid_outbreak is the COVID-19 pandemic indicator, and post_time was coded as 0 for pre-pandemic intervals and as the number of 6-month intervals since March 2020 for post-pandemic periods, representing the change in slope after the pandemic onset. We fitted models using the glmer function in the lme4 package (https://cran.r-project.org/web/packages/lme4/citation.html) with the bobyqa optimizer. To test whether the data is zero-inflated, we compared observed vs. expected zero counts under the full Poisson model to evaluate the possibility of zero inflation. For the outcome “number of viral suppression (VS) events,” we observed 1,727 zero values, which is fewer than the expected number of zeros (2362.4) estimated by the model. This suggests that zero inflation was not present, and a standard Poisson model was appropriate. Model diagnostics and goodness-of-fit measures, including log-likelihood, Akaike Information Criterion (AIC), and Bayesian Information Criterion (BIC), were examined and documented. The AIC and BIC of the first outcome, i.e., number of healthcare visits, are 22825.6 and 23057.0 respectively. The AIC and BIC of the second outcome, i.e., number of viral suppression events, are 11272.4 and 11488.6 respectively. Model assumptions for Poisson mixed effects ITS were deemed appropriate for our data.
Regarding data cleaning, viral load records with non-numeric values or missing units were excluded from the outcome measure. Additionally, any COVID-19 diagnoses dated prior to 2020 were disregarded, as the virus had not yet emerged and such records likely reflect data entry errors or misclassifications. For data missingness issue, demographic variables such as sex, race/ethnicity, and annual household income with missing values were categorized as ‘Unknown’ for analysis. The proportion of missing data in these variables was minimal (4.35%−5.19%) except for the income level (23.53%) and is unlikely to have impacted on the overall results. All analyses were performed in R (version 4.5.0) within the secure, cloud-based AoU Researcher Workbench platform.
Results
Characteristics of study cohort
Among 7,667 PLWH identified, 1,772 of them who had at least one CD4 count/viral load measure were used for the healthcare visit analysis and 1,382 who had at least one viral load measure was used for the viral suppression analysis. Among 1,772 participants, 50.9% were aged 50–64 years old, 64.2% were male, 52.7% were Blacks, 21.3% were Hispanic or Latino, 47.9% were never married, 42.6% were gay or bisexual, 59.9% had an annual income less than 35,000 USD, 90.4% attained high school or above, and 52.1% used Medicaid as their primary insurance. The demographic distribution of the subcohort for viral suppression is similar with the main cohort for healthcare visit outcome. Among 1,382 participants, 50.1% were aged 50–64 years old, 63.4% were male, 54.6% were Blacks, 22.0% were Hispanic or Latino, 49.3% were never married, 41.1% were gay or bisexual, 61.7% had an annual income less than 35,000 USD, 90.5% attained high school or above, and 53.6% used Medicaid as their primary insurance. (Table 1) For the two outcomes of interest, the total count of healthcare visits dropped from 7902 (average of 4.90 per person) before pandemic to 5252 (average of 3.70 per person) after pandemic, while the total count of viral suppression events dropped from 2361 (average of 1.98 per person) before pandemic to 1893 (average of 1.88 per person) after pandemic.
Table 1.
Baseline characteristics of all PWH
| Characteristic | Sample for healthcare visits N (%) N = 1,772 |
Sample for viral suppression N (%) N = 1,382 |
|---|---|---|
| Sex | ||
| Female | 565 (31.9) | 450 (32.6) |
| Male | 1,137 (64.2) | 876 (63.4) |
| Gender non-binary or other | 70 (4.0) | 56 (4.1) |
| Age (years, mean, SD) | 55 (12.1) | 54 (12.2) |
| Age Group (years old) | ||
| 50–64 | 902 (50.9) | 693 (50.1) |
| 65+ | 341 (19.2) | 234 (16.9) |
| < 50 | 529 (29.9) | 455 (32.9) |
| Race Group | ||
| Asian/Other/Unknown | 474 (26.8) | 372 (26.9) |
| Black or African American | 934 (52.7) | 755 (54.6) |
| White | 364 (20.5) | 255 (18.5) |
| Ethnicity Group | ||
| Hispanic or Latino | 377 (21.3) | 304 (22.0) |
| Not Hispanic or Latino | 1,311 (74.0) | 1,020 (73.8) |
| Unknown | 84 (4.7) | 58 (4.2) |
| Marital Status | ||
| Divorced/Separated/Widowed | 443 (25.0) | 345 (25.0) |
| Married/Living With Partner | 389 (22.0) | 284 (20.6) |
| Never Married | 848 (47.9) | 681 (49.3) |
| Unknown | 92 (5.2) | 72 (5.2) |
| Sex Orientation* | ||
| Prefer Not To Answer or Skip | 107 (6.0) | 75 (5.4) |
| Bisexual | 148 (8.4) | 123 (8.9) |
| Gay | 605 (34.2) | 445 (32.2) |
| Straight | 860 (48.6) | 698 (50.5) |
| Education | ||
| High school degree or more | 1,602 (90.4) | 1,250 (90.5) |
| Less than a high school degree | 86 (4.9) | 68 (4.9) |
| Unknown | 84 (4.7) | 64 (4.6) |
| Income | ||
| Greater than 35,000 US dollars | 293 (16.5) | 190 (13.8) |
| Less than 35,000 US dollars | 1,062 (59.9) | 853 (61.7) |
| Unknown | 417 (23.5) | 339 (24.5) |
| Medicaid | ||
| No | 848 (47.9) | 641 (46.4) |
| Yes | 924 (52.1) | 741 (53.6) |
| Medicare | ||
| No | 1,332 (75.2) | 1,045 (75.6) |
| Yes | 440 (24.8) | 337 (24.4) |
| Employer or Union | ||
| No | 1,587 (89.6) | 1,254 (90.7) |
| Yes | 185 (10.4) | 128 (9.3) |
| Other Health Plans | ||
| No | 1,583 (89.3) | 1,244 (90.0) |
| Yes | 189 (10.7) | 138 (10.0) |
| Ever diagnosed with COVID-19 | ||
| No | 1,517 (85.6) | 1,177 (85.2) |
| Yes | 255 (14.4) | 205 (14.8) |
| Disability status | ||
| no | 1,554 (87.7) | 1,212 (87.7) |
| yes | 218 (12.3) | 170 (12.3) |
| Number of Comorbidities | ||
| 0 | 886 (50.0) | 710 (51.4) |
| 1 | 436 (24.6) | 337 (24.4) |
| ≥ 2 | 450 (25.4) | 335 (24.2) |
*”Lesbian” or “None” category was not reported given the smaller cell size
Factors associated with healthcare visits frequency
Results from interrupted time series model revealed that COVID-19 pandemic has a negative impact on the number of clinical visits (adjusted odds ratio [aOR] = 0.80, 95%CI: 0.75, 0.86). Older individuals had a lower odds of healthcare visits (50–64 yrs vs. < 50 yrs: aOR = 0.91, 95% CI: 0.86, 0.96; and 65 + yrs vs. < 50 yrs: aOR = 0.88, 95%CI: 0.82, 0.94), while individuals who were Asian/other/unknown (aOR = 1.17, 95% CI: 1.06, 1.30), who had Medicaid insurance (aOR = 1.12, 95%CI: 1.06, 1.17) and with more comorbidities (aOR = 1.07, 95%CI: 1.02–1.13) were more likely to have healthcare visits. (Table 2)
Table 2.
Factors associated with number of healthcare visits among PWH, results from interrupted time series results
| Effect | OR (95% C.I.) | p-value |
|---|---|---|
| Time | 0.99 (0.97, 1.01) | 0.47 |
| Post time | 0.99 (0.96, 1.02) | 0.57 |
| Covid-19 Outbreak | ||
| Yes | 0.80 (0.75, 0.86) | < 0.01 |
| No | Reference | |
| Covid Infection | ||
| Yes | 0.98 (0.90, 1.08) | 0.73 |
| No | Reference | |
| Age | ||
| 50–64 | 0.91 (0.87, 0.96) | < 0.01 |
| 65+ | 0.88 (0.82, 0.94) | < 0.01 |
| 50- | Reference | |
| Race | ||
| Asian/Other/Unknown | 1.17 (1.06, 1.30) | < 0.01 |
| Black Or African American | 1.05 (0.98, 1.12) | 0.15 |
| White | Reference | |
| Gender | ||
| Male | 1.01 (0.95, 1.06) | 0.79 |
| Not Man Only, Not Woman Only, Prefer Not To Answer, Or Skipped | 0.99 (0.88, 1.11) | 0.82 |
| Female | Reference | |
| Ethnicity | ||
| Hispanic Or Latino | 1.03 (0.94, 1.13) | 0.56 |
| Unknown | 0.94 (0.82, 1.08) | 0.39 |
| Not Hispanic Or Latino | Reference | |
| Marital Status | ||
| Marital Status Divorced/Separated/Widowed Vs. Never Married | 1.01 (0.95, 1.06) | 0.85 |
| Marital Status Married/Living With Partner Vs. Never Married | 0.97 (0.92, 1.02) | 0.26 |
| Marital Status Unknow Vs. Never Married | 0.96 (0.87, 1.07) | 0.44 |
| Never Married | Reference | |
| Sex Orientation | ||
| Prefer Not To Answer, Or Skipped | 0.96 (0.87, 1.06) | 0.38 |
| Bisexual | 1.02 (0.94, 1.11) | 0.58 |
| Gay | 0.97 (0.91, 1.03) | 0.27 |
| Lesbian | 1.06 (0.85, 1.33) | 0.61 |
| None | 0.97 (0.84, 1.12) | 0.66 |
| Straight | Reference | |
| Education | ||
| High School Degree Or More | 1.04 (0.94, 1.15) | 0.46 |
| Unknown | 1.02 (0.88, 1.18) | 0.79 |
| Less Than A High School Degree | Reference | |
| Income | ||
| Greater Than 35,000 Us Dollars | 1.01 (0.94, 1.09) | 0.74 |
| Unknown | 1.04 (0.99, 1.10) | 0.11 |
| Less Than 35,000 Us Dollars | Reference | |
| Disability | ||
| Yes | 0.95 (0.89, 1.01) | 0.12 |
| No | Reference | |
| Medicaid | ||
| Yes | 1.12 (1.06, 1.17) | < 0.01 |
| No | Reference | |
| Medicare | ||
| Yes | 0.98 (0.93, 1.04) | 0.58 |
| No | Reference | |
| Employer Or Union | ||
| Yes | 0.95 (0.87, 1.03) | 0.23 |
| No | Reference | |
| Other Health Plans | ||
| Yes | 0.92 (0.85, 1.00) | 0.04 |
| No | Reference | |
| Number Of Comorbidities | ||
| 1 | 1.02 (0.97, 1.07) | 0.49 |
| 2+ | 1.07 (1.02, 1.13) | 0.01 |
| 0 | Reference | |
Factors associated with viral suppression status
Results from interrupted time series model did not find an association between COVID-19 pandemic and number of viral suppression events (aOR = 1.10, 95%CI: 0.97–1.25). Regarding socio-demographic characteristics, individuals who were older (50–64 yrs vs. < 50 yrs: aOR = 1.27, 95%CI: 1.15, 1.40 and 65 + yrs vs. < 50: aOR = 1.41, 95%CI: 1.22, 1.62), Hispanic or Latino (aOR = 1.32, 95%CI: 1.09–1.61), and married/living with partners (aOR = 1.14, 95%CI: 1.03–1.28) were more likely to achieve viral suppression, while individuals who had more comorbidities (aOR = 0.89, 95%CI: 0.80–0.99) were less likely to achieve viral suppression. (Table 3)
Table 3.
Factors associated with number of viral suppression among PWH, results from interrupted time series results
| Effect | OR (95% C.I.) | p-value |
|---|---|---|
| Time | 1.01 (0.98, 1.05) | 0.50 |
| Post time | 0.98 (0.93, 1.03) | 0.43 |
| Covid Outbreak | ||
| Yes | 1.10 (0.97, 1.25) | 0.15 |
| No | Reference | |
| Covid Infection | ||
| Yes | 1.06 (0.90, 1.24) | 0.48 |
| No | Reference | |
| Age | ||
| 50–64 | 1.27 (1.15, 1.40) | < 0.01 |
| 65+ | 1.41 (1.22, 1.62) | < 0.01 |
| 50- | Reference | |
| Race | ||
| Asian/Other/Unknown | 0.84 (0.69, 1.03) | 0.09 |
| Black Or African American | 0.90 (0.79, 1.03) | 0.12 |
| White | Reference | |
| Gender | ||
| Male | 0.94 (0.85, 1.04) | 0.26 |
| Not Man Only, Not Woman Only, Prefer Not To Answer, Or Skipped | 0.87 (0.69, 1.09) | 0.23 |
| Female | Reference | |
| Ethnicity | ||
| Hispanic Or Latino | 1.32 (1.09, 1.61) | < 0.01 |
| Unknown | 1.18 (0.89, 1.57) | 0.26 |
| Not Hispanic Or Latino | Reference | |
| Marital Status | ||
| Divorced/Separated/Widowed | 1.03 (0.93, 1.15) | 0.55 |
| Married/Living With Partner | 1.14 (1.03, 1.28) | 0.02 |
| Unknown | 1.11 (0.90, 1.36) | 0.35 |
| Never Married | Reference | |
| Sex Orientation | ||
| Prefer Not To Answer, Or Skipped | 0.92 (0.75, 1.12) | 0.40 |
| Bisexual | 0.83 (0.71, 0.98) | 0.03 |
| Gay | 1.07 (0.95, 1.21) | 0.26 |
| Lesbian | 0.96 (0.59, 1.58) | 0.88 |
| None | 1.03 (0.79, 1.33) | 0.84 |
| Straight | Reference | |
| Education | ||
| High School Degree Or More | 1.02 (0.84, 1.24) | 0.83 |
| Unknown | 0.95 (0.72, 1.26) | 0.75 |
| Less Than A High School Degree | Reference | |
| Income | ||
| Greater Than 35,000 Us Dollars | 0.91 (0.78, 1.07) | 0.26 |
| Unknown | 0.96 (0.87, 1.06) | 0.43 |
| Less Than 35,000 Us Dollars | Reference | |
| Disability | ||
| Yes | 1.06 (0.93, 1.2) | 0.38 |
| No | Reference | |
| Medicaid | ||
| Yes | 0.94 (0.86, 1.03) | 0.20 |
| No | Reference | |
| Medicare | ||
| Yes | 1.06 (0.95, 1.18) | 0.29 |
| No | Reference | |
| Employer Or Union | ||
| Yes | 1.18 (0.99, 1.40) | 0.07 |
| No | Reference | |
| Other Health Plans | ||
| Yes | 1.11 (0.95, 1.3) | 0.17 |
| No | Reference | |
| Number Of Comorbidities | ||
| 1 | 1.01 (0.91, 1.12) | 0.81 |
| 2+ | 0.89 (0.80, 0.99) | 0.03 |
| 0 | Reference | |
Discussion
Using the nationwide EHR dataset, this study investigated the impact of the pandemic on healthcare encounters and viral suppression among underrepresented PLWH. Our findings provided incremental real-world evidence regarding the impact of COVID-19 pandemic on HIV healthcare outcomes among vulnerable PLWH. The reduced number of healthcare visits has revealed the overall negative impact of the pandemic on HIV care outcomes. These results are corroborated with most global evidence including some southern US studies [10, 20, 27]. The sustained viral suppression rate observed both in our study and one VA study might be attributed to the significant increase in virtual visits (~ 90%) and extended refill length for ARVs for PLWH—facilitated by VA-specific guidance and support in the form of additional training and tablets that were provided for virtual visits [17]. In contrast, other healthcare systems may not have implemented such comprehensive support strategies to promote telehealth service. The differentiated service delivery likely contributes to ARV coverage and maintained viral suppression throughout the pandemic. The non-pharmaceutical interventions for COVID-19, such as social distancing, stay-at-home orders, might form the structural barriers and indirectly affect the patients’ access to care alongside the personal choice to avoid accessing healthcare [28–31]. Meanwhile, the closure of HIV clinics or postponing of non-emergent, elective medical services could cause the service disruptions [32, 33]. The strategic plan to deal with different levels of service disruptions and efforts to mitigate the negative impact of the pandemic on HIV care varied between US and other European countries. For instance, a review conducted in the European countries suggested that HIV clinics rarely closed or experienced ART shortages since the beginning of the pandemic and majority of them were running at normal capacity [34]. In contrast, a study conducted in the Southern US indicated that most HIV clinics (~ 80%) faced with complete or partial disruptions during the early phase of the pandemic [35]. Telehealth was recommended by the US HRSA Ryan White Programs, which ensures the provision of safe patient care and minimize infectious risk of PLWH [36]. Despite this, the telehealth access remains challenging for many PLWH and some patients felt less involved during the telemedicine visit [35, 37, 38].
Individuals with Medicaid insurance were less likely to miss medical encounters but not necessarily more likely to achieve VS. A US study previously reported that individuals with Medicaid or no insurance were relatively less likely to be retained/suppressed compared to their respective counterparts [39]. Our study did not find such a relationship. Since the majority of the subjects in our study are underrepresented PLWH, those who are not enrolled in the Medicaid insurance program might enrolled in other supplementary health services, such as the AIDS Drug Assistance Program (ADAP) funded through Part B of Ryan White HIV/AIDS program [40]. These supplementary services provision (e.g., ADAP) often have a positive impact on HIV care outcomes [41], which might explain why we did not see poorer outcomes for VS in our study.
Our study found that older adults with HIV had fewer healthcare visits but maintained viral suppression rate compared with younger population. On the one hand, the older population might go through more structural barriers, such as transportation barriers during the pandemic. While telehealth services expanded during COVID-19 pandemic period, older patients often encountered barriers such as limited digital literacy, lack of access to necessary technology, and a preference for in-person consultations. These factors contributed to disruptions in care continuity and increased feelings of isolation among this demographic. In contrast, younger individuals with HIV were generally more adept at utilizing telehealth platforms, leading to fewer interruptions in their care [42]. On the other hand, older PLWH often demonstrates higher levels of ART adherence than younger PLWH, which strongly correlates with viral suppression [43]. Despite of the pandemic, older population still maintained high viral suppression rates, which is consistent with other studies [44]. Older adults with HIV are often long-term survivors, they might have accumulated experience and familiarity with medication regimens [45], which contribute to sustained viral control, even amid healthcare disruptions with fewer visits [45].
Although racial disparities in HIV outcomes have historically existed, our analysis found no significant differences in healthcare visits or viral suppression rates between Black or African American and White PLWH during the COVID-19 pandemic. Interestingly, Hispanic or Latino PLWH were more Likely to achieve viral suppression. One VA study found that ARV coverage and viral load testing percentage was similar between Black, White and Hispanic for both 2019 and 2020[17], but viral suppression rate was not measured, which prohibit us from do a head-to-head comparison. The lack of difference in healthcare visits and viral suppression rates between Black and White PLWH may be attributed to the temporary rapid telehealth expansion. In general, many HIV clinics have historically high no-show rates pre-pandemic [46] and the increasing use of phone/telehealth encounters may have actually been beneficial in increasing visit adherence. This may be particularly true for Black population who often face challenges in obtaining transportation, child care, or time off from work [47] or individuals for whom the experience coming to clinic is stigmatizing and reminds them of their HIV status [48]. Additionally, economic hardships and job losses during the lockdown may have newly affected insurance coverage or healthcare engagement among some White populations. In this context, the observed narrowing of disparities could reflect the benefit of telehealth expansion and a convergence due to worsening outcomes among some White PLWH. Thus, the observed attenuation of racial disparities in care outcomes may obscure the nuanced and differential impacts of the pandemic across racial and ethnic groups.
Several limitations of this study are worth noting. First, due to unavailable data, we could not examine the impact of the COVID-19 pandemic on ART initiation. Second, the time frame of this study is limited to the available data at the time of analysis. Future studies should consider a longer follow up period to observe and confirm the long-term effect of the pandemic on HIV care continuum outcomes.
Conclusions
Despite these limitations, our findings, using a nationwide database demonstrated the negative impact of the COVID-19 pandemic on the disruption of healthcare visits among underrepresented PLWH. Being unable to maintain HIV care engagement during the COVID-19 pandemic is potentially a significant hindrance to achieving the “90-90-90” goal and possibly derailing progress of the Ending the HIV Epidemic Initiative. Notably, younger PLWH or PLWH with more comorbidities were less likely to maintain viral suppression. These results underscore the need for tailored interventions to mitigate care disruptions. For younger PLWH, incorporating telemedicine into routine practice in an effective manner might mitigate negative impacts to healthcare systems in the future. The non-significant differences of health outcomes between black and white populations could reflect the benefit of telehealth expansion and a convergence due to worsening outcomes among some White PLWH. As we experience unprecedented financial stressors and necessary changes to our healthcare delivery system, we must adapt our care and services to continue to engage PLWH to avoid poor outcomes amongst our more vulnerable patients.
Acknowledgements
The research reported in this publication was supported by the National Institute of Allergy and Infectious Diseases of the National Institutes of Health under Award Number R01AI164947-02S1 and R21AI170159-01A1. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health or USC. Funders had no role in the design of the study, collection, analysis and interpretation of the data.
Abbreviations
- PLWH
People living with HIV
- AoU
All of Us
- aOR
Adjusted odds ratio
- MSM
Men who have sex with men
- EHR
Electronic health record
- ICD
International Classification Disease
- VS
Viral suppression
- HRSA
Health Resources and Services Administration
- ADAP
AIDS Drug Assistance Program
Authors’ contributions
XY and JZ conceptualized the study design. XY wrote the first draft and critical revision of the manuscript. RC and BH led efforts on All of Us HIV markers harmonization, as well as critically reviewed the manuscript. RC wrote data preparation code and SQL, R code for data analysis, which was reviewed and verified by JZ. XY prepared tables and figures with input from BH. SW provided clinical input regarding patient identification and validation. BO, SW, JZ, and XL reviewed and edited the manuscript. Authorship was determined using ICMJE recommendations. The corresponding author (and some other authors) had full access to all the data in the study. All authors had final responsibility for the decision to submit for publication.
Funding
The research reported in this publication was supported by the National Institute of Allergy and Infectious Diseases of the National Institutes of Health under Award Number R01AI164947-02S1. Dr. Xueying Yang’s effort was supported by the National Institute of Allergy and Infectious Diseases of the National Institutes of Health under Award Number (R21AI170159). The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health. Both NIAID and NIH had no role in the design of the study, collection, analysis and interpretation of the data.
Data availability
Access to the Researcher Workbench and data is free and available at https://www.researchallofus.org/. All researchers must be authorized and approved via a 3-step process that includes registration, completion of ethics training and attestation to a data use agreement.
Declarations
Ethics approval and consent to participate
Research using All of Us scientific resources are conducted in accordance with the Federal Policy for the Protection of Human Subjects (The Common Rule, 45 CFR 46), which is founded upon the ethical principles of the Belmont Report. Data use is governed by additional safeguards through the Resource Access Board (RAB), which reviews research projects and ensures compliance with the Data User Code of Conduct, including guarding against stigmatizing or discriminatory outcomes.Due to the use of secondary data, our study was approved by the Institutional Review Board at the University of South Carolina (Pro00124806) as a ‘non-human subject research designation’. Therefore, the patient informed consent is not applicable.
Consent for publication
Not applicable.
Competing interests
The authors declare no competing interests.
Footnotes
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
References
- 1.Walker PGT, Whittaker C, Watson OJ, et al. The impact of COVID-19 and strategies for mitigation and suppression in low- and middle-income countries. Science. 2020(6502). 10.1126/science.abc0035. [DOI] [PMC free article] [PubMed]
- 2.Amura CR, Thorne J, Bean M, et al. Evolution of HIV health care workforce needs in the U.S. Mountain West during the COVID-19 pandemic: a mixed method study. J Assoc Nurses AIDS Care. 2024;01(Mar-Apr):78–90. 10.1097/jnc.0000000000000448. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Mshenga MM, Khalid FJ, Haji SH, Ali TB, Mohamed KA, Damian DJ. The hidden effects of COVID-19 on HIV services in Zanzibar: country report. AIDS Res Ther. 2023;20(1):72. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Finlayson-Trick E, Tam C, Wang L, et al. Access to care and impact on HIV treatment interruptions during the COVID‐19 pandemic among people living with HIV in British Columbia. HIV Med. 2024;25(9):1007–18. [DOI] [PubMed] [Google Scholar]
- 5.Khaing M, Lwin S, Paw N, et al. Service interruption in HIV care amid COVID-19 pandemic in Myanmar: results from analysis of routine program data 2018–2022. Journal of the International Association of Providers of AIDS Care (JIAPAC). 2024;23:23259582241299466. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Sanchez TH, Zlotorzynska M, Rai M, Baral SD. Characterizing the impact of COVID-19 on men who have sex with men across the United States in april, 2020. AIDS Behav. 2020;24(7):2024–32. 10.1007/s10461-020-02894-2. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Jewell BL, Mudimu E, Stover J, et al. Potential effects of disruption to HIV programmes in sub-Saharan Africa caused by COVID-19: results from multiple mathematical models. Lancet HIV Sep. 2020;7(9):e629–40. 10.1016/S2352-3018(20)30211-3. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Quiros-Roldan E, Magro P, Carriero C, et al. Consequences of the COVID-19 pandemic on the continuum of care in a cohort of people living with HIV followed in a single center of Northern Italy. AIDS Res Ther. 2020;17(1):1–8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Sun Y, Li H, Luo G, et al. Antiretroviral treatment interruption among people living with HIV during COVID-19 outbreak in China: a nationwide cross‐sectional study. J Int AIDS Soc. 2020. 10.1002/jia2.25637. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.SeyedAlinaghi S, Mirzapour P, Pashaei Z, et al. The impacts of COVID-19 pandemic on service delivery and treatment outcomes in people living with HIV: a systematic review. AIDS Res Ther. 2023;20(1):4. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Bradley H, Hall HI, Wolitski RJ, et al. Vital signs: HIV diagnosis, care, and treatment among persons living with HIV—United states, 2011. MMWR Morb Mortal Wkly Rep. 2014;63(47):1113–7. [PMC free article] [PubMed] [Google Scholar]
- 12.Ulett KB, Willig JH, Lin H-Y, et al. The therapeutic implications of timely linkage and early retention in HIV care. AIDS Patient Care STDS. 2009;23(1):41–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Chakrabarti R, Agasty D, Majumdar A, et al. Syndemic effect of COVID-19 outbreak on HIV care delivery around the globe: a systematic review using narrative synthesis. Health Promot Perspect. 2023;13(4):243. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Ukaegbu E, Maulenkul T, Sarria-Santamera A. Impact of COVID-19 on utilization of healthcare services among people living with HIV (PLHIV): a systematic review. Medicina (Kaunas). 2025;61(1):111. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.(COVID-19) CD. Frequently Asked Questions. HIV/AIDS Bureau. Accessed Sept 28. 2020. https://hab.hrsa.gov/coronavirus/frequently-asked-questions
- 16.Budak JZ, Scott JD, Dhanireddy S, Wood BR. The impact of COVID-19 on HIV care provided via telemedicine—past, present, and future. Curr HIV AIDS Rep. 2021;18:98–104. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.McGinnis KA, Skanderson M, Justice AC, et al. HIV care using differentiated service delivery during the COVID-19 pandemic: a nationwide cohort study in the US department of veterans affairs. J Int AIDS Soc. 2021;24:e25810. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Monroe AK, Xiao J, Greenberg AE, et al. Risk of severe COVID-19 disease and the pandemic’s impact on service utilization among a longitudinal cohort of persons with HIV-Washington, DC. AIDS Behav. 2022;26(10):3289–99. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Norwood J, Kheshti A, Shepherd BE, et al. The impact of COVID-19 on the HIV care continuum in a large urban Southern clinic. AIDS Behav. 2022;26(8):2825–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Barish N, Barth S, Monroe AK, Greenberg AE, Castel AD. Site assessment survey to assess the impact of the COVID-19 pandemic on HIV clinic site services and strategies for mitigation in Washington, DC. BMC Health Serv Res. 2023;23(1):1130. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Matsumoto S, Nagai M, Luong DAD, et al. Evaluation of SARS-CoV-2 antibodies and the impact of COVID-19 on the HIV care continuum, economic security, risky health behaviors, and mental health among HIV-infected individuals in Vietnam. AIDS Behav. 2021. 10.1007/s10461-021-03464-w. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Investigators AoURP. The all of Us research program. N Engl J Med. 2019;381(7):668–76. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Mapes BM, Foster CS, Kusnoor SV, et al. Diversity and inclusion for the all of Us research program: a scoping review. PLoS ONE. 2020;15(7):e0234962. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Yang X, Zhang J, Cai R, et al. Computational phenotyping with the all of Us research program: identifying underrepresented people with HIV or at risk of HIV. JAMIA Open. 2023;6(3):ooad071. 10.1093/jamiaopen/ooad071. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Arnold EM, Swendeman D, Harris D, et al. The stepped care intervention to suppress viral load in youth living with HIV: protocol for a randomized controlled trial. JMIR Res Protoc. 2019;(2):e10791. 10.2196/10791. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Zhang J, Yang X, Weissman S, Li X, Olatosi B. Protocol for developing a personalised prediction model for viral suppression among under-represented populations in the context of the COVID-19 pandemic. BMJ Open. 2023;(5):e070869. 10.1136/bmjopen-2022-070869. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Li Z, Qiao S, Ning H, et al. Place visitation data reveals the geographic and racial disparities of COVID-19 impact on HIV service utilization in the deep South. AIDS Behav. 2023(S1). 10.1007/s10461-023-04163-4. [DOI] [PubMed]
- 28.Karjadi TH, Maria S, Yunihastuti E, Widhani A, Kurniati N, Imran D. Knowledge, attitude, behavior, and socioeconomic conditions of people living with HIV in Indonesia during the COVID-19 pandemic: a cross-sectional study. HIV/AIDS. 2021. 10.2147/HIV.S333469. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.Campbell LS, Masquillier C, Knight L, et al. Stay-at-home: the impact of the COVID-19 lockdown on household functioning and ART adherence for people living with HIV in three sub-districts of Cape Town, South Africa. AIDS Behav. 2022;26(6):1905–22. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.Beima-Sofie K, Ortblad KF, Swanson F, Graham SM, Stekler JD, Simoni JM. HIV service provision for priority populations during the COVID-19 pandemic in Seattle, WA. AIDS Behav. 2020;24(10):2760–3. 10.1007/s10461-020-02902-5. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31.Ridgway JP, Schmitt J, Friedman E, et al. HIV care continuum and COVID-19 outcomes among people living with HIV during the COVID-19 pandemic, chicago, IL. AIDS Behav Oct. 2020;24(10):2770–2. 10.1007/s10461-020-02905-2. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32.Yang X, Zeng C, Tam CC, et al. HIV service interruptions during the COVID-19 pandemic in China: the role of COVID-19 challenges and institutional response from healthcare professional’s perspective. AIDS Behav. 2022;26(4):1270–8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33.Qiao S, Li Z, Weissman S, et al. Disparity in HIV service interruption in the outbreak of covid-19 in South Carolina. AIDS Behav. 2020;(1):9. 10.1007/s10461-020-03013-x. [DOI] [PMC free article] [PubMed]
- 34.Kowalska JD, Skrzat-Klapaczyńska A, Bursa D, Europe? et al. (1878–3511 (Electronic)).
- 35.Qiao S, Li Z, Weissman S, et al. Disparity in HIV service interruption in the outbreak of COVID-19 in South Carolina. AIDS Behav. 2021;25(1):49–57. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36.Siewe Fodjo JN, Faria de Moura Villela E, Van Hees S, Vanholder P, Reyntiens P, Colebunders R. Follow-up survey of the impact of COVID-19 on people living with HIV during the second semester of the pandemic. Int J Environ Res Public Health. 2021;18(9):4635. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 37.Ballivian J, Alcaide ML, Cecchini D, Jones DL, Abbamonte JM, Cassetti I. Impact of COVID-19-related stress and lockdown on mental health among people living with HIV in Argentina. J Acquir Immune Defic Syndr. 2020;1(Dec):475–82. 10.1097/qai.0000000000002493. [DOI] [PubMed] [Google Scholar]
- 38.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. 2020;24(10):2743–7. 10.1007/s10461-020-02895-1. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 39.Yehia BR, Stephens-Shields AJ, Fleishman JA, et al. The HIV care continuum: changes over time in retention in care and viral suppression. PLoS ONE. 2015;10(6):e0129376. 10.1371/journal.pone.0129376. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 40.Kay ES, Batey DS, Mugavero MJ. The Ryan white HIV/AIDS program: supplementary service provision Post-Affordable care act. AIDS Patient Care STDS Jul. 2018;32(7):265–71. 10.1089/apc.2018.0032. [DOI] [PubMed] [Google Scholar]
- 41.Diepstra KL, Rhodes AG, Bono RS, Patel S, Yerkes LE, Kimmel AD. Comprehensive Ryan White assistance and human immunodeficiency virus clinical outcomes: retention in care and viral suppression in a Medicaid nonexpansion state. Clin Infect Dis. 2017;15(4):619–25. 10.1093/cid/cix380. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 42.Reyes E, Silvis J, Gandhi M, Shi Y, Greene M. Telehealth access and experiences of older adults with HIV during the COVID-19 pandemic: lessons for the future. J Am Geriatr Soc. 2024;72(9):2816–24. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 43.Beer L, Skarbinski J. Adherence to antiretroviral therapy among HIV-infected adults in the united States. AIDS Educ Prev. 2014;26(6):521–37. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 44.Lewis T-AJ, Kaiser ME, Goldshteyn N, Sepkowitz D, Briggs WM, Sepkowitz DV. A retrospective analysis of the disruptions in the HIV continuum of care during the COVID-19 pandemic: lessons from a Clinic-Based study. Cureus. 2024;16(2). [DOI] [PMC free article] [PubMed]
- 45.Ghidei L, Simone MJ, Salow MJ, et al. Aging, antiretrovirals, and adherence: a meta analysis of adherence among older HIV-infected individuals. Drugs Aging. 2013;30:809–19. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 46.Adams JA, Whiteman K, McGraw S. Reducing missed appointments for patients with HIV: an evidence-based approach. J Nurs Care Qual. 2020;35(2):165–70. [DOI] [PubMed] [Google Scholar]
- 47.Yehia BR, Stewart L, Momplaisir F, et al. Barriers and facilitators to patient retention in HIV care. BMC Infect Dis. 2015;15:1–10. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 48.Vanable PA, Carey MP, Blair DC, Littlewood RA. Impact of HIV-related stigma on health behaviors and psychological adjustment among HIV-positive men and women. AIDS Behav. 2006;10:473–82. [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.
Data Availability Statement
Access to the Researcher Workbench and data is free and available at https://www.researchallofus.org/. All researchers must be authorized and approved via a 3-step process that includes registration, completion of ethics training and attestation to a data use agreement.