Impact of the COVID-19 Pandemic on Mental Health and Viral Suppression among Persons Living with HIV in Western Washington

Abstract

The COVID-19 pandemic and social distancing measures elevated stress levels globally, exacerbating mental health challenges for people with HIV (PWH). We examined the effect of COVID-19-related stress on mental health among PWH in western Washington, exploring whether social support and coping self-efficacy were protective. Data on COVID-19-related stress, mental health, social support, and coping self-efficacy were collected using online surveys during the pandemic. Pre-COVID-19 mental health data were available for a subset of participants and were linked with the survey data. In the total sample (N=373), COVID-19-stress was associated with elevated depression (PHQ-8, β=0.21, 95%CI [0.10, 0.32]) and anxiety (GAD-7, β=0.28, 95%CI [0.17, 0.39]). Among the subset of respondents with pre-pandemic mental health data (N=103), COVID-19-related stress was associated with elevated PHQ-8 scores (β=0.35, 95%CI [0.15, 0.56]) and GAD-7 scores (β=0.35, 95%CI [0.16, 0.54]), adjusted for baseline mental health and other confounders. Coping self-efficacy was negatively associated with GAD-7 scores (β=−0.01, 95%CI [−0.01, 0.00]), while social support was negatively associated with PHQ-8 scores (β=−0.06, 95%CI [−0.12, −0.01]). Viral suppression before and during the pandemic did not differ among participants with available data. While COVID-19-related stress predicted elevated depression and anxiety symptoms among PWH, social support and coping self-efficacy were protective.

Introduction

According to the World Health Organization, the COVID-19 pandemic has resulted in approximately 0.6 billion confirmed cases and 6.9 million deaths as of June 14, 2023 (WHO, 2023). To curtail the spread of COVID-19 once the pandemic was recognized, most countries implemented stringent public health measures, such as shelter-in-place orders, travel bans, and social distancing rules (Adhikari et al., 2020). The resultant disruptions in daily routines, financial hardship, and housing instability, together with worries about personal safety and a heightened sense of uncertainty, contributed to an increase in psychological distress at the population level (Hossain et al., 2020; Nicola et al., 2020). A review of 28 studies on the mental health impact of COVID-19 found that subclinical mental health problems were prevalent in countries across the world (Rajkumar, 2020).

As in the general population, COVID-19 has had negative impacts on mental health outcomes among people with HIV (PWH), whose prevalence of mental illness is higher than that of the general population (Cook et al., 2018). Both qualitative and quantitative studies have reported anxiety and depressive symptoms among PWH during the pandemic, due to concerns about acquiring COVID-19, as well as isolation and stress due to social distancing measures (Kuman Tunçel et al., 2020; Parisi et al., 2022). Although data for an increased risk of mortality among PWH from COVID-19 are mixed, PWH were understandably anxious that their condition would place them at higher risk of poor clinical outcomes (Geretti et al., 2020; Waters & Pozniak, 2021). In addition, PWH may have been more vulnerable to the impact of COVID-19 for multiple reasons. For example, PWH faced the challenge of disrupted access to services including medication refills and routine monitoring visits during lockdowns in the early COVID-19 pandemic (Cao et al., 2020; Thoits, 1986). Additionally, structural inequalities among PWH were amplified by the COVID-19 pandemic, as PWH anticipated and experienced increased financial stress and food insecurity compared to HIV-uninfected individuals (Kalichman & El-Krab, 2022). The synergistic effects of multiple epidemics among PWH, including mental health challenges, illicit drug use, and other comorbidities, together with structural inequalities, amplified psychological distress among PWH during the pandemic (Shiau et al., 2020).

Understanding protective factors mitigating mental distress due to COVID-19 is important to plan for future pandemics or public health emergencies, as well as to inform the development of health services to address mental distress, including among PWH. In the face of life challenges such as COVID-19, individuals often utilize their own resources to cope with stress, using both problem-focused and emotion-focused coping skills (Chew et al., 2020). Coping self-efficacy in particular, as an important predictor of effective coping, has been associated with lower levels of depression and better health status among PWH (Prati & Pietrantoni, 2016; Rodkjaer et al., 2014). Social support could be conceptualized as coping assistance that comes in the form of emotional, instrumental, or information support, and may offer protection against the psychological impact of COVID-19 (Cao et al., 2020). Social support may enhance the receiver’s sense of control over the environment and may bolster coping self-efficacy (Thoits, 1986). Perceived social support and receipt of support from important persons within their social networks were known to predict better mental health outcomes among PWH before the COVID-19 pandemic (McDowell & Serovich, 2007; Reich et al., 2010). Therefore, both coping self-efficacy and social support are important in assessing the impact of COVID-19-related stress on mental health among PWH.

Given the widely reported mental health implications of the COVID-19 pandemic and the potential need for future interventions to ameliorate psychological distress, we aimed to capture the impact of the COVID-19 pandemic and social distancing policies on mental health and other aspects of the lives of PWH in western Washington. In addition, we examined whether social support and coping self-efficacy were protective against mental health deterioration. Our aims were to: 1) describe the impact of the COVID-19 pandemic on participants’ mental health and viral load suppression; 2) investigate associations between COVID-19-related stress, individual challenges (e.g., housing instability, job loss) and mental health outcomes; and 3) explore whether social support and coping-self-efficacy served as protective factors among study participants.

Methods

Participants

Study participants were recruited from the main University of Washington (UW) HIV clinic in Seattle and four of its satellite clinics in western Washington (in Federal Way and in Kitsap, Snohomish, and Thurston Counties). Participants were eligible if they were 18 years of age or older, able to speak English, enrolled in HIV care at a UW site, had consented to participate in the UW HIV patient registry, and were willing to complete an online survey. Due to COVID-19 social distancing restrictions in place, in-person recruitment, consenting, and study participation were not permitted during the study period. Recruitment was therefore conducted by telephone, e-mail or mail in accordance with stated preferences for UW registry participation. Survey participants received a small incentive ($20) for participation, which lasted from 20–30 minutes. Enrollment was closed after 400 individuals completed the survey, due to limited funding.

Design and procedures

The western Washington HIV and COVID-19 study consisted of a cross-sectional Computer-Assisted Personal Interview (CAPI) survey taken online, followed by virtually conducted in-depth interviews with a subset of participants. The CAPI survey was created using REDCap (Research Electronic Data Capture) electronic data capture tools hosted at the University of Washington and was open for participation from 9/2020 to 3/2021. CAPI has been used to reduce response bias and underreporting due to perceived social stigma in studies of psychological conditions and substance abuse among stigmatized populations (Brown et al., 2013). During the consent process, survey participants were also asked for permission to link their CAPI data to clinical data in the UW HIV Information System (UWHIS) from the 12 months preceding the Washington State shelter-in-place order in March 2020. Figure 1 presents the study time periods and how they were defined.

Figure 1.

Study data collection timeline

Ethical review

Ethical review and oversight were provided by the University of Washington Human Subjects Division (STUDY00010385). All participants provided electronic informed consent. Persons who declined participation were encouraged to reach out to their clinic provider for any needs. Persons who participated were also encouraged to do this and additionally provided with information on local service organizations providing case management and mental health referrals.

Measures collected in CAPI survey

The CAPI survey included questions assessing mental health, COVID-19 stress, COVID-19-related challenges, social support, and coping self-efficacy.

COVID-19 stress and related challenges

Seven items assessed different types of stress due to the pandemic (e.g., “I’m afraid of getting COVID-19”) and related social distancing measures (e.g., “Social distancing has resulted in increased mental stress”). For these seven items, participants indicated the intensity of the stress using a five-point Likert scale, ranging from 0 (not at all) to 4 (extremely). Responses to the seven items were summed to derive a total COVID-19 stress score. The internal consistency (i.e., Cronbach’s alpha) of this scale in this study was acceptable at 0.71. This scale was created by selecting items from the 36-item COVID Stress Scale to capture several dimensions of COVID-19 stress while minimizing participant burden (Taylor et al., 2020).

In addition to capturing stress, the survey captured experiences with COVID-19 illness, job loss, housing challenges, and HIV-related service access and engagement. Participants answered two dichotomous (yes/no) questions to assess COVID-19 illness (experiencing COVID-like symptoms, testing positive), and one question to evaluate job loss (yes/no). Three questions used 5-point Likert scales (1=significantly decreased; 5=significantly increased) to capture participants’ perceived change in paid work hours, number of jobs held, and number of job opportunities. Housing challenges were assessed by asking participants questions about loss of housing, crowding, or housing instability related to the order to shelter in place. Changes in HIV-related service access were assessed by asking how the patient’s access to pharmacy, ability to get medication refills, and ability to adhere to antiretroviral therapy (ART) changed due to COVID-19.

Mental Health

Perceived mental health challenges during the pandemic were assessed by asking participants to answer the question, “Has your mental health improved, worsened, or not changed during the COVID-19 pandemic?” Responses included 1 (improved), 2 (no change), and 3 (worsened).

Depressive symptoms over the prior two weeks were measured with the Patient Health Questionnaire-8 (PHQ-8), which contains all items in the PHQ-9 except the one assessing suicide and thoughts of self-harm.(Kroenke et al., 2009) Participants indicated how often they experienced depressive symptoms over the past two weeks by choosing among four response options, from 0=not at all to 3=nearly every day. A total score was calculated by summing individual responses, with a higher score indicating more severe depressive symptoms. PHQ-8 scores were analyzed as continuous outcomes; internal consistency of the PHQ-8 in this study was good (Cronbach’s alpha = 0.90).

Anxiety symptoms were measured using the General Anxiety Disorder-7 (GAD-7).(Löwe et al., 2008) Participants indicated how frequently they experienced these symptoms by choosing from four response options, from 0=not at all to 3=nearly every day. A total score was calculated by summing the responses to all items, with a higher score indicating more severe anxiety symptoms. The GAD-7 score was analyzed as a continuous outcome, and its internal consistency in this study was excellent (Cronbach’s alpha = 0.92).

Coping self-efficacy

Self-efficacy for different types of coping was assessed using the 13-item Coping Self-Efficacy (CSE) Scale, which measures three dimensions of coping self-efficacy, including using problem-focused coping (6 items, e.g. break an upsetting problem down into smaller parts), stopping unpleasant emotions and thoughts (4 items, e.g. make unpleasant thoughts go away), and getting support from friends and family (3 items, e.g., get friends to help you with the things you need). A total score was calculated by summing the responses for all items to indicate participants’ overall level of self-efficacy. This scale had been validated, and it had excellent internal consistency in this study (Cronbach’s alpha = 0.92) (Chesney et al., 2006).

Social support

Social support was assessed using the Multidimensional Scale of Perceived Social Support, a 12-item scale of perceived social support from family (e.g., I get the emotional help and support I need from my family), friends (e.g., I can count on my friends when things go wrong), or others (e.g., There is a special person who is around when I am in need).(Zimet et al., 1990) Participants responded using a 7-point Likert scale, ranging from 1 (very strongly disagree) to 7 (very strongly agree). A total score was calculated by summing responses to all items to indicate overall perceived social support from family, friends, and others. This scale has been widely used in a variety of settings, and it demonstrated excellent internal consistency in this study (Cronbach’s alpha = 0.94) (Bruwer et al., 2008).

Adherence

Change in ART adherence during COVID-19 was assessed with one item: How has your adherence to ART changed due to COVID-19? The responses were on a five-point Likert scale ranging from 1=significantly decreased to 5=significantly increased.

Measures abstracted from chart data

UWHIS data were abstracted for the 12 months before COVID-19 social distancing in Washington State (Mar 23, 2019 through March 22, 2020) and the 12 months after the start of COVID-19 social distancing (March 23, 2020 through March 22, 2021) from the electronic health record and other institutional sources. These data included visit dates, medications prescribed, and laboratory test results including viral loads.

Viral load

HIV viral load data were analyzed as a binary outcome, with viral load <200 copies/mL considered suppressed and viral load ≥200 copies/mL considered unsuppressed (Centers for Disease Control and Prevention, 2022). Viral loads were classified as pre-COVID-19 (March 23, 2019 through March 22, 2020) versus during COVID-19 social distancing (March 23, 2020 through March 22, 2021). If no viral load data were found during a given period, viral load was classified as missing.

Mental health measures

UWHIS collects patient-reported information on sexual behavior, substance abuse and mental health from participants in the UW HIV patient registry who voluntarily complete a brief survey at biannual clinic visits. Completion rates for the patient-reported data were very low after the start of COVID-19 social distancing, due to severely limited in-person clinical care visits. Completion of the questionnaire resumed in the summer of 2020 on a limited basis, first with the addition of a remote option for the subset of PWH doing Telehealth rather than in-person clinical visits, with a slow return to clinic-based completion for those PWH attending clinical care visits. Therefore, most UWHIS participants had only pre-COVID-19 patient-reported data available as a baseline indicator of mental health.

We abstracted patient-reported data on mental health measures from the last pre-COVID-19 visit to compare with the post-social distancing data collected via the CAPI survey. Depressive symptoms were captured using PHQ-8, and so were directly comparable to the CAPI PHQ-8 measure. The GAD-7 is not included in UWHIS patient assessments, and so was not available in the pre-COVID-19 period. As a proxy measure of anxiety symptoms in the pre-COVID-19 period, we abstracted responses to a UWHIS question on anxiety attacks (In the last 4 weeks, have you had an anxiety attack?).

Data analysis

Data from all participants who provided consent to link chart data were analyzed using summary descriptive statistics, both for the overall study population and by whether participants had mental health data available from the pre-COVID-19 period. Chi square or Fisher’s exact tests for categorical data and independent t tests were used to compare characteristics of participants with and without pre-COVID-19 mental health data. In the subpopulation with viral load data in the pre-COVID-19 and COVID-19 periods, McNemar’s test was used to compare viral suppression rates across periods.

In the subpopulation with mental health data in the pre-COVID-19 and COVID-19 periods, multivariable linear regression was used to evaluate the impact of COVID-19 stress, coping self-efficacy and social support on depression and anxiety, after adjustment for age, race, and ethnicity, job loss, housing loss, and baseline depression or anxiety. Moderation effects of coping self-efficacy and social support on the association between COVID-19 stress and mental health outcomes were tested by adding an interaction term between COVID-19 stress and either coping self-efficacy or social support into the multivariable regression model. In a sensitivity analysis, the same set of regression models were repeated without controlling for baseline depression or anxiety, using CAPI survey data from the entire study population.

Results

Participants

Out of 397 participants who provided consent and completed the survey, 373 also consented to link their survey data to clinical data, and 103 had data on self-reported mental health symptoms before and during COVID-19 pandemic. As seen in Table 1, the mean age of all participants was 46 years (SD = 12), ranging from 18 to 76. Most participants (71.3%) identified as White, 12.9% were Black, and the remainder Asian, American Indian, Hawaiian/Pacific Islander, multi-racial or other. Overall, 8.3% identified as Hispanic or Latinx. Most participants identified as male (81.5%), while 15.0% identified as female and 3.5% identified as non-binary. The vast majority of the participants (82.7%) identified as sexual minority by orientation, with only 17.3% identifying as heterosexual. About 67.0% of participants had an annual income of less than 40,214 USD, the cut point for the lowest income quartile in King County, Washington as of 2018 (Household Income in King County, 2022).

Table 1.

Participant Characteristics, COVID-19-Related Challenges and COVID-19-Related Stress Reported by CAPI Survey,^a Overall and by Whether Pre-COVID-19 Mental Health Data Were Available or Not

	Overall (N = 373)	Pre-COVID-19 Data Available (N=103)	Pre-COVID-19 Data Not Available (N = 270)	p value
Sociodemographic characteristics
Age				0.320b
Mean (SD)c	46.0 (12.0)	47.0 (11.6)	45.7 (12.2)
Range	18 – 76	18 – 67	21 – 76
Race				0.897d
American Indian	15 (4.0%)	3 (2.9%)	12 (4.4%)
Asian	29 (7.8%)	8 (7.8%)	21 (7.8%)
Black	48 (12.9%)	12 (11.7%)	36 (13.3%)
Hawaiian or Pacific Islander	6 (1.6%)	2 (1.9%)	4 (1.5%)
White	266 (71.3%)	77 (74.8%)	189 (70.0%)
Multi-Race	1 (0.3%)	1 (0.4%)	0 (0.0%)
Prefer not to answer	8 (2.1%)	1 (1.0%)	7 (2.6%)
Ethnicity				0.135d
Non-Hispanic	342 (91.7%)	98 (95.1%)	244 (90.4%)
Hispanic	31 (8.3%)	5 (4.9%)	26 (9.6%)
Sexual Orientation				0.816d
Straight or Heterosexual	65 (17.3%)	17 (16.5%)	48 (17.6%)
Bisexual	37 (9.9%)	10 (9.7%)	27 (9.9%)
Gay or Lesbian	249 (66.4%)	72 (69.9%)	177 (65.1%)
Queer	14 (3.7%)	3 (2.9%)	11 (4.0%)
Something else	8 (2.1%)	1 (1.0%)	7 (2.6%)
Don’t know	2 (0.5%)	0 (0.0%)	2 (0.7%)
Gender Identity				0.341d
Female	56 (15.0%)	15 (14.6%)	41 (15.2%)
Male	304 (81.5%)	82 (79.6%)	222 (82.2%)
Non-Binary	13 (3.5%)	6 (5.8%)	7 (2.6%)
Annual Incomee				0.546d
<40,214	243 (65.1%)	69 (67.0%)	174 (64.4%)
40,215 – 75,049	48 (12.9%)	11 (10.7%)	37 (13.7%)
75,050 – 117,801	24 (6.4%)	9 (8.7%)	15 (5.6%)
>117,801	58 (15.5%)	14 (13.6%)	44 (16.3%)
Education				0.409d
12th grade or less	31 (8.3%)	7 (6.8%)	24 (8.9%)
High school graduate or GEDf	36 (9.7%)	14 (13.6%)	22 (8.1%)
Some college/AA degree/Technical school training	179 (48.0%)	44 (42.7%)	135 (50.0%)
College graduate	84 (22.5%)	26 (25.2%)	58 (21.5%)
Graduate school degree	43 (11.5%)	12 (11.7%)	31 (11.5%)
COVID-19-related challenges
How has your mental health changed due to COVID-19?				0.201d
Worsened	193 (51.7%)	52 (50.5%)	141 (52.2%)
No change	141 (37.8%)	36 (35.0%)	105 (38.9%)
Improved	28 (7.5%)	9 (8.7%)	19 (7.0%)
Prefer not to answer	11 (2.9%)	6 (5.8%)	5 (1.9%)
How has access to your pharmacy changed due to COVID-19?				0.917d
Somewhat Decreased	34 (9.1%)	11 (10.7%)	23 (8.5%)
Significantly Decreased	15 (4.0%)	3 (2.9%)	12 (4.4%)
No Change	283 (75.9%)	79 (76.7%)	204 (75.6%)
Somewhat Increased	7 (1.9%)	2 (1.9%)	5 (1.9%)
Significantly Increased	18 (4.8%)	5 (4.9%)	13 (4.8%)
Missing	16 (4.3%)	3 (2.9%)	13 (4.8%)
How has your ability to get medication refills changed due to COVID-19?				0.845d
Significantly Decreased	9 (2.4%)	3 (2.9%)	6 (2.2%)
Somewhat Decreased	31 (8.3%)	9 (8.7%)	22 (8.1%)
No Change	292 (78.3%)	81 (78.6%)	211 (78.1%)
Somewhat Increased	10 (2.7%)	1 (1.0%)	9 (3.3%)
Significantly Increased	16 (4.3%)	4 (3.9%)	12 (4.4%)
Missing	15 (4.0%)	5 (4.9%)	10 (3.7%)
How has your adherence to ART changed due to COVID-19?				0.488d
Significantly Decreased	10 (2.7%)	2 (1.9%)	8 (3.0%)
Somewhat Decreased	19 (5.1%)	3 (2.9%)	16 (5.9%)
No Change	292 (78.3%)	86 (83.5%)	206 (76.3%)
Somewhat Increased	12 (3.2%)	3 (2.9%)	9 (3.3%)
Significantly Increased	7 (1.9%)	3 (2.9%)	4 (1.5%)
Missing	33 (8.8%)	6 (5.8%)	27 (10.0%)
Have you lost a job?				0.122d
No	265 (71.0%)	81 (78.6%)	184 (68.1%)
Yes	107 (28.7%)	22 (21.4%)	85 (31.5%)
Missing	1 (0.3%)	0 (0.0%)	1 (0.4%)
How has your number of paid work hours changed?				0.395d
Significantly Decreased	96 (25.7%)	21 (20.4%)	75 (27.8%)
Somewhat Decreased	46 (12.3%)	16 (15.5%)	30 (11.1%)
No Change	122 (32.7%)	39 (37.9%)	83 (30.7%)
Somewhat Increased	18 (4.8%)	4 (3.9%)	14 (5.2%)
Significantly Increased	13 (3.5%)	2 (1.9%)	11 (4.1%)
Missing	78 (20.9%)	21 (20.4%)	57 (21.1%)
How has the number of jobs you hold changed?				0.443d
Significantly Decreased	78 (20.9%)	17 (16.5%)	61 (22.6%)
Somewhat Decreased	27 (7.2%)	8 (7.8%)	19 (7.0%)
No Change	170 (45.6%)	53 (51.5%)	117 (43.3%)
Somewhat Increased	12 (3.2%)	3 (2.9%)	9 (3.3%)
Significantly Increased	6 (1.6%)	0 (0.0%)	6 (2.2%)
Missing	80 (21.4%)	22 (21.4%)	58 (21.5%)
How has the number of job opportunities changed?				0.523c
Significantly Decreased	77 (20.6%)	22 (21.4%)	55 (20.4%)
Somewhat Decreased	59 (15.8%)	15 (14.6%)	44 (16.3%)
No Change	137 (36.7%)	37 (35.9%)	100 (37.0%)
Somewhat Increased	22 (5.9%)	10 (9.7%)	12 (4.4%)
Significantly Increased	7 (1.9%)	2 (1.9%)	5 (1.9%)
Missing	71 (19.0%)	17 (16.5%)	54 (20.0%)
Have you lost housing?				0.681d
No	353 (94.6%)	255 (94.4%)	98 (95.1%)
Yes	18 (4.8%)	13 (4.8%)	5 (4.9%)
Missing	2 (0.5%)	2 (0.7%)	0 (0.0%)
Is your current living situation more crowded than before?				0.027d
No	329 (88.2%)	97 (94.2%)	232 (85.9%)
Yes	44 (11.8%)	6 (5.8%)	38 (14.1%)
How afraid are you of losing housing?				0.594d
Not concerned at all	174 (46.6%)	123 (45.6%)	51 (49.5%)
Mildly concerned	64 (17.2%)	43 (15.9%)	21 (20.4%)
Somewhat concerned	38 (10.2%)	29 (10.7%)	9 (8.7%)
Very concerned	53 (14.2%)	40 (14.8%)	13 (12.6%)
Missing	44 (11.8%)	35 (13.0%)	9 (8.7%)
COVID-19-related stress items and total scoreb
COVID-19 impacted my day-to-day life.				0.572
Mean (SD)	2.7 (1.2)	2.7 (1.1)	2.7 (1.2)
I am afraid of getting COVID-19.				0.068
Mean (SD)	2.2 (1.3)	2.4 (1.2)	2.2 (1.3)
I am afraid of spreading COVID-19.				0.649
Mean (SD)	2.1 (1.4)	2.2 (1.3)	2.1 (1.4)
I am afraid of being an asymptomatic carrier.				0.537
Mean (SD)	2.0 (1.3)	2.0 (1.2)	1.9 (1.3)
I fear stigma/ discrimination from other people (e.g., people treating you differently because of your identity, having symptoms, or other factors related to COVID-19).				0.635
Mean (SD)	1.3 (1.4)	1.2 (1.3)	1.3 (1.4)
I feel that I am contributing to the greater good by practicing social distancing. (Reverse scored)				0.681
Mean (SD)	0.8 (1.0)	0.7 (1.0)	0.8 (1.0)
Social distancing has resulted in increased mental stress.				0.697
Mean (SD)	2.0 (1.4)	2.0 (1.3)	2.1 (1.4)
COVID-19 stress total score				0.773
Mean (SD)	13.1 (5.1)	13.3 (4.5)	13.1 (5.3)
Range	2 – 24	3 – 24	2 – 24

^aCAPI: Computer-Assisted Personal Interview

^bIndependent t test

^cStandard deviation

^dPearson’s Chi-squared or Fisher’s exact tests

^eBased on King County, Washington Income Quantiles.²²

^fGED: General Educational Development

Few participants reported having been diagnosed with COVID-19 by the time of their survey participation (14 of 373, or 3.7%). Table 1 presents other impacts of COVID-19 social distancing. The majority of participants (51.7%) reported worsened mental health due to COVID-19, while only 7.5% reported improved mental health. Several participants (13.1%) reported decreased pharmacy access and reduced ability to get medication refills after social distancing was implemented. Overall, 28.7% of participants reported losing their job due to COVID-19 and about a third reported that job opportunities and paid work hours had decreased due to the pandemic. Eighteen participants (4.8%) reported the loss of housing due to COVID-19, while approximately half reported at least some concern about losing their housing.

Participants with and without pre-COVID-19 mental health data had similar sociodemographic characteristics and reported similar levels of COVID-19-related stress and challenges. The only exception was that participants without pre-COVID-19 mental health data were more likely to report crowded housing than those with this data (14.1% vs. 5.8%, Table 1).

Adherence and viral suppression

Among all study participants (N = 373), most (78.3%) reported no change in ART adherence, but 29 (7.8%) participants reported somewhat or significantly decreased adherence during COVID-19 social distancing. A small percentage of participants had missing viral load data either before (16.1%) or during the pandemic (23.6%). Among 256 participants (68.6%) who had viral load test results available both before and during COVID-19, 68.4% of these participants (175 out of 256) were virally suppressed at both timepoints and 5.5% (14 out of 256) were unsuppressed at both timepoints (Table 2). A McNemar test was not significant (p = 1), suggesting no difference in viral suppression rates before and during COVID-19 among participants for whom viral load data were available.

Table 2.

McNemar’s Test for Viral Suppression Before versus During COVID-19 Social Distancing (N = 256)^a

		Viral suppression during COVID-19
		Yes	No	Total	p value
Viral suppression before COVID-19	Yes	175 (68.4%)	33 (12.9%)	208 (81.2%)	1b
	No	34 (13.3%)	14 (5.5%)	48 (18.8%)
	Total	209 (81.6%)	47 (18.4%)	256 (100%)

^aViral load suppression was defined as <200 copies/mL. Row or column percentages are provided for totals, while percentages of the total are provided for cells shaded in gray. Excludes 117 participants for whom viral load data were missing in one or both periods.

^bMcNemars Chi-squared = 0, df = 1.

COVID-19 stress, mental health outcomes, social support and coping self-efficacy

Among the subset of 103 participants with pre-COVID-19 mental health data, half (50.5%) reported worsened mental health after social distancing measures were put in place. Depression severity (total score on PHQ-8) significantly increased (mean = 7.6, SD = 5.8) during COVID-19 social distancing, compared to the year before (mean = 5.5, SD = 5.9, t (102) = 3.94, 95% CI [1.03, 3.11], p = 0.012). The percentage of participants who reported mild to severe depression increased from 41.7% pre-pandemic to 65.0% post-pandemic. This worsening of self-reported depressive symptoms based on PHQ-8 score was consistent with and associated with participants’ overall perception of worsening mental health measured by a single-item question (p=0.02). As for anxiety symptoms, 30% of participants reported having anxiety attacks in the year before social distancing. The mean GAD-7 score after social distancing was 6.6 (SD = 5.7), with 25.2% of participants categorized as having mild anxiety and another 25.2% as having moderate or severe anxiety (Table 3).

Table 3.

Mental Health Measures Before versus During COVID-19 Social Distancing (N = 103)^a

	Before	During	p value
Anxiety attacks reportedb			Not applicable
Yes	30 (29.1%)	Not available
GAD-7 (continuous)b
Mean (SD)	Not available	6.6 (5.7)
Range	Not available	0 – 21
GAD-7 (categorical)b			Not applicable
Minimal anxiety	Not available	51 (49.5%)
Mild anxiety	Not available	26 (25.2%)
Moderate anxiety	Not available	13 (12.6%)
Severe anxiety	Not available	13 (12.6%)
PHQ-8 score			0.012c
Mean (SD)	5.5 (5.8)	7.6 (5.8)
Range	0 – 20	0 – 24
PHQ-8 (categorical)			0.011d
None to minimal	60 (58.3%)	36 (35.0%)
Mild	19 (18.4%)	37 (35.9%)
Moderate	13 (12.6%)	16 (15.5%)
Moderately severe	8 (7.8%)	8 (7.8%)
Severe	3 (2.9%)	6 (5.8%)

^aExcludes 270 participants for whom pre-COVID-19 mental health data were not available.

^bBefore the pandemic, a single patient-reported item measured whether participants’ had anxiety attacks in the past four weeks. During COVID-19 social distancing, the GAD-7 score from the CAPI survey was used to measure anxiety symptoms.

^cPaired t-test.

^dChi-squared test

Table 4 presents results of bivariable and multivariable linear regressions evaluating associations with the two outcomes (i.e., PHQ-8 and GAD-7) during the COVID-19 pandemic among the 103 participants with pre-COVID-19 mental health data. COVID-related stress was positively associated with higher PHQ-8 score (β = 0.351, 95% CI [0.145, 0.558] and GAD-7 score (β=0.348, 95% CI [0.155, 0.540]), controlling for sociodemographic factors (age, gender, race, ethnicity). Social support was negatively associated with PHQ-8 score (β = −0.065, 95% CI [−0.121, −0.010]) and coping self-efficacy was negatively associated with GAD-7 score (β= −0.007, 95% CI [−0.012, −0.002]). The moderating effects of social support and coping self-efficacy on the associations between COVID-19-related stress and both depression and anxiety were not significant (Table 5). Sensitivity analyses conducted in the entire study population without controlling for pre-COVID-19 mental health variables produced similar results, but the impact of COVID-19-related stress on depression and anxiety score was underestimated, as indicated by the smaller regression coefficients in these models (depression: 0.211 vs. 0.351 and anxiety: 0.280 vs. 0.348 in Supplemental Table 1 compared to Table 4). Moderating effects of social support and coping self-efficacy were not significant in sensitivity analyses (Supplemental Table 2).

Table 4.

Bivariable and Multivariable Associations with PHQ-8 Score and GAD-7 Score during COVID-19 Social Distancing as Outcomes (N = 103)

	PHQ-8 Score
Predictors	Bivariable				Multivariablea
	B	95% CI		p	β	95% CI		p
COVID-19 stress score	0.288	0.036	0.539	0.025*	0.351	0.145	0.558	0.001*
Lost housing	5.925	0.733	11.116	0.026*	1.261	−3.701	6.222	0.614
Lost job	2.520	−0.225	5.266	0.072	1.498	−0.972	3.968	0.231
Depression (baseline)	0.579	0.421	0.738	<0.001*	0.464	0.272	0.656	<0.001*
Coping efficacy score	−0.015	−0.019	−0.011	<0.001*	−0.004	−0.010	0.002	0.219
Social support score	−0.105	−0.156	−0.054	<0.001*	−0.065	−0.121	−0.010	0.022*

	GAD-7 Score
Predictors	Bivariable				Multivariablea
	β	95% CI		p	β	95% CI		p
*COVID-19 stress score	0.428	0.191	0.664	<0.001*	0.348	0.155	0.540	0.001*
Lost housing	7.125	2.134	12.115	0.006*	3.144	−1.456	7.743	0.177
Lost job	2.447	−0.229	5.122	0.073	1.101	−1.187	3.389	0.341
Anxiety Attack (baseline)	7.260	5.270	9.2500	<0.001*	3.542	1.363	5.722	<0.001*
Coping efficacy score	−0.015	−0.019	−0.011	<0.001*	−0.007	−0.012	−0.002	0.007*
Social support score	−0.090	−0.142	−0.039	<0.001*	−0.047	−0.098	0.005	0.077

^aEach multivariable model included all listed predictors and adjustment for age, race (white/non-white), and ethnicity (Hispanic/non-Hispanic)

^*p < 0.05

Table 5.

Multivariable Analysis Testing Coping Efficacy and Social Support as Moderators between COVID-19 Stress and Mental Health Outcomes (N = 103)

	Mental Health Outcomesa
Moderator: Coping Efficacy	Depression				Anxiety
	β	95% CI		p	β	95% CI		p
COVID-19 stress score	0.570	−0.147	1.286	0.118	0.829	0.181	1.477	0.013*
Depression or Anxiety (baseline)	0.456	0.259	0.653	<0.001*	3.442	1.332	5.553	0.002*
Coping efficacy score	−0.003	−0.016	0.089	0.582	−0.002	−0.013	9.113	0.753
COVID-19 stress score* Coping efficacy score	−0.000	−0.001	0.005	0.423	−0.001	−0.001	0.001	0.098

Moderator: Social support	Depression				Anxiety
	β	95% CI		p	β	95% CI		p
Depression or Anxiety (baseline)	0.488	0.334	0.641	<0.001*	5.070	3.158	6.981	<0.001*
COVID-19 stress score	0.367	0.181	0.553	<0.001*	0.375	0.189	0.562	<0.001*
Social support score	−0.020	−0.143	0.104	0.751	0.014	−0.106	0.133	0.822
COVID-19 stress score* Social support score	−0.004	−0.012	0.005	0.391	−0.007	−0.015	0.001	0.098

^aEach multivariable model included all listed predictors and adjustment for age, gender, race, ethnicity, housing loss, and job loss.

^*p < 0.05

Discussion

The COVID-19 pandemic has impacted every aspect of people’s lives, particularly early in the pandemic before vaccines were available. This study, conducted in the first year after social distancing was ordered in Washington State, evaluated the impact of the pandemic on mental health outcomes among PWH. We recruited a diverse population, with many participants reporting high levels of COVID-19-related stress, barriers to accessing pharmacies for medication refills, and housing and job challenges. Nevertheless, the majority of the participants in this study maintained medication adherence and viral suppression during the pandemic. In contrast, both depressive and anxiety symptoms were common, and over half of our participants reported worse mental health during the pandemic. We found that high COVID-19-related stress was associated with worse mental health, while social support and individual coping seem to have been protective against this adverse impact.

The COVID-19 rate was 3.7% in this study, similar to the test positivity rate in King County, Washington, which ranged from a low of 1.9% to a high of 9.5% over the course of our study period (The New York Times, 2021). Almost all participants expressed concerns about getting infected or spreading COVID-19. This high level of concern could be due to a heightened perception of risk in our study population due to their HIV infection but may simply be due to the timing of the study, prior to the widespread availability of effective vaccines. Early studies of SARS CoV-2 reported an increased risk of death from COVID-19 among PWH compared to people not living with HIV, after adjusting for smoking, obesity, and ethnicity (Bhaskaran et al., 2021). Moreover, many risk factors for adverse COVID-19 outcomes, such as smoking and comorbid illnesses (e.g., diabetes) are more prevalent among PWH, compared to HIV-uninfected individuals (Guaraldi et al., 2020). While we did not have a comparison group to assess COVID-19-related stress among people without HIV, our findings of high levels of COVID-19-related stress among PWH and negative impacts on mental health are similar to that reported in other studies (Kalichman & El-Krab, 2022).

In the present study, several participants lost jobs or housing, while others reported worrying about the possibility of losing their source of income and being unable to afford housing. In March 2020, Washington state established policy and programs such as the King County Eviction Prevention and Rental Assistance Program to mitigate housing insecurity due to COVID-19 pandemic (Seattle King County Public Health, 2022). Nevertheless, the COVID-19 pandemic exacerbated pre-existing disparities in the social determinants of health among PWH, increasing food and housing insecurity, decreasing access to care, and adding to a burden of mental health symptoms (Kalichman & El-Krab, 2022; Shiau et al., 2020). For example, a study conducted in a rural area of the Southern United States (US) found that 74% of PWH lost their jobs due to the pandemic. Meanwhile, the utilization of housing services in that area increased by 69% shortly after the pandemic started (Sherbuk et al., 2020). Housing instability as a contextual factor, coupled with individual vulnerabilities, could lead to negative psychological consequences and adverse HIV-related health outcomes, such as poor adherence, disengagement from care, and lack of viral suppression (Aidala et al., 2016; Marquez et al., 2019). Crowded living conditions due to housing instability also increased the risks of exposure to both COVID-19 and HIV (Sachdev et al., 2021). Given these health implications, providing housing assistance for PWH, especially in the context of a pandemic, could reduce barriers to HIV care access and relieve mental distress due to housing instability.

Despite reduced ART adherence reported by some of our participants, the vast majority had undetectable viral loads both before (80.5%) and during the pandemic (80.4%), similar to the 84% suppression rates in King County in 2019 (Public Health – Seattle & King County, 2020). Similar findings were reported in a few other US studies where ART adherence was maintained or even increased during the early COVID-19 pandemic. For example, in a group of young PWH in Atlanta, ART adherence increased from 53% to 78% in the first few months of the pandemic (Kalichman et al., 2020). This could be due to increased health consciousness in the context of shelter-in-place in the early pandemic (Gwadz et al., 2021). Results from the qualitative interviews of a subset of our study participants suggested that some were afraid they would be at risk for severe COVID-19 if their CD4 count dropped, possibly motivating increased adherence (Smith et al., 2021). Although only about 13% of participants in our study reported decreased access to their pharmacy and to refills, this number was higher in other studies. A study of gay, bisexual, and other men who have sex with men in 20 countries found that close to half of PWH were unable to obtain ART medication refills, especially in countries such as Mexico, Brazil, and Russia (Rao et al., 2021). This difference could be due to stronger support from the Medicaid and AIDS Drug Assistance Programs in Washington state, which protected the supply chain to pharmacies and mitigated loss of insurance during the pandemic (Armstrong et al., 2020; AIDS AIDS Drug Assistance Program, 2022).

PWH already experience a higher prevalence of mental health conditions such as depression and anxiety, compared to individuals without HIV (Remien et al., 2019). The COVID-19 pandemic and subsequent social distancing measures exacerbated these mental health disparities through isolation, stress, and worry, including among PWH. For example, a study of 49 PWH in New York between April and September 2020 found that 40% of participants reported anxiety and depression symptoms (Pizzirusso et al., 2021). While other studies have reported a high rate of mental health symptoms during COVID-19 pandemic, ours is the first study, to our knowledge, that used pre-COVID-19 mental health data to demonstrate an increase in mental health symptoms over pre-existing levels among PWH. In this study, participants’ self-reported depressive symptoms indicated by the PHQ-8 score increased during COVID-19 social distancing, compared to PHQ-8 scores in the same individuals before the pandemic. Our COVID-19 stress score based on seven questions about COVID-19-related stress was positively associated with PHQ-8 scores during the pandemic. In addition, anxiety symptoms during COVID-19 were also associated with COVID-19-related stress scores among participants in this study, including after adjustment for pre-COVID-19 anxiety attacks. Results of our sensitivity analyses suggest that the impact of COVID-19 stress on depression and anxiety was underestimated if the baseline scores were not controlled for. Given the relationship between stress and mental health symptoms during the pandemic, public health measures such as online screenings and referrals should be included in planning for future pandemics. Self-paced online mental health resources were beneficial in reducing worry, anxiety, depression and improving resilience and overall well-being during the pandemic (Heckendorf et al., 2022; Rackoff et al., 2022). The mental health implications of COVID-19 highlight the urgent need to increase access to patient supports by expanding digital mental health resources and online menta health services (Figueroa & Aguilera, 2020).

In the face of these negative impacts, social support and coping seemed to have had protective effects on depressive symptoms and anxiety among our study participants. While we were not able to demonstrate moderation effects, potentially due to the limited power in this study, we found that participants with higher levels of social support from family, friends, or significant others had lower depressive symptoms scores. Similarly, individuals with higher coping self-efficacy had lower anxiety scores during COVID-19 social distancing. A study of PWH in Argentina found that responding adaptively to adversity, including responding with strategies such as problem solving and reaching out for support, had buffering effects on economic hardship and stress due to COVID-19 (Ballivian et al., 2020). In addition, a study by the same group including 1,554 PWH from both Argentina and the US found that higher satisfaction with social support and higher resilient coping (e.g., “I look for creative ways to alter difficult situations”) were associated with lower levels of depression, measured by CES-D, during the COVID-19 pandemic (Jones et al., 2021). Initiatives to promote social support and resilient coping among PWH could be especially beneficial to the mental health of this population in the context of an ongoing global COVID-19 pandemic and could also be relevant after the pandemic is over.

There are several limitations of this study. First, our study population had limited racial/ethnic diversity, with underrepresentation of Black (13%) and Hispanic (8%) PWH. According to statistics on PWH in Washington State, approximately 57% are White, 17% Black, and 15% Hispanic (Washington State Department of Health, 2020). This study also excluded individuals who did not speak English and thus the results have limited generalizability to non-English speakers. Due to COVID-19 restrictions on research during the study period, participants were recruited and enrolled online. For this reason, our sample is not representative of individuals who were unstably housed or had poor or no internet access. We therefore have likely underestimated the impact of the COVID-19 pandemic on PWH in western Washington. Second, our measures of anxiety were different in the clinical data from the period before COVID-19 social distancing and in our REDCap survey, which presented challenges with drawing conclusions about whether anxiety symptoms had changed after the COVID-19 outbreak. In addition, not all participants completed the optional mental health questions at their clinic visits before the pandemic. However, we used the more common and validated GAD-7 as a measure of anxiety symptoms during the pandemic, and were able to adjust for pre-pandemic anxiety attacks, finding that the COVID-19 stress score remained associated with GAD-7 score during the pandemic, after adjusting for a history of anxiety attacks. Finally, this study focused on the socioeconomic and mental health impact of COVID-19-related stress within the first year of the pandemic. Future studies are needed to examine the long-term impact of the COVID-19 pandemic on PWH.

At a time of global health crisis like the COVID-19 pandemic, it is difficult to accurately capture the degree of change in mental health caused by the crisis unless pre-event data are available. For this reason, very few studies have examined differences in mental health for a given population before and after the start of the pandemic. This study is one of the first studies to examine the impact of COVID-19 on mental health outcomes controlling for baseline mental health, using pre-pandemic clinical data and survey data distributed online during pandemic during a period of social distancing and “shelter-in-place” orders.

Conclusion

The COVID-19 pandemic has had important socioeconomic and mental health impacts on PWH. COVID-19-related stress was associated with increased depression and anxiety symptoms during the pandemic. At the structural level, financial and housing instability contributed to increased stress among PWH. Government programs and community services focused on stable housing and job security during public health emergencies are one mechanism that could reduce mental distress among this key population. At the individual level, social support and the ability to cope mitigated negative mental health outcomes. Related interventions and services, such as case management and counseling may improve mental health outcomes. In the future, the long-term impact of the COVID-19 pandemic on the mental health outcomes of PWH should continue to be monitored, and mental health should be considered in planning for future pandemics.