Gender Differences in the Relationship Between Social Support and Quality of Life Among People Living with HIV During the COVID-19 Pandemic

The introduction of the SARS-CoV-2 virus into the United States in December 2019 and the ensuing rapid spread, morbidity, and mortality from COVID-19 drastically impacted all aspects of life, including social connections. In March of 2020, the Centers for Disease Control and Prevention (CDC) recommended social distancing (i.e., staying at least 6 feet away from others and reducing close contact) to minimize SARS-CoV-2 transmission.¹ Strict social distancing regulations introduced in Washington State and elsewhere were associated with mental distress. For example, a CDC study conducted in the United States in June 2020 found that younger adults and racial and ethnic minorities reported high levels of adverse mental health conditions, substance use, and suicidal ideation.² An online survey in 2020 completed by a diverse population of 2020 individuals from around the world with a wide range of age, gender, income, education level, and occupation found that 60% of participants who experienced self-isolation, defined as staying away from other people to prevent the spread of the virus, reported a deterioration of their mental health after lockdown measures were enforced.³ Notably, those who were self-isolating at the time of responding to the survey scored higher on the Patient Health Questionnaire-9 (PHQ-9), indicating more severe depressive symptoms, than those who were not self-isolating.³

Social support is an important buffer between stressors and deteriorating mental health and can come from family, friends, and significant others. The American Psychological Association defines social support as “the provision of assistance or comfort to others, typically to help them cope with biological, psychological, and social stressors.”⁴ There are many different types of social support, including emotional support (i.e., outward expressions of empathy and love), instrumental support (i.e., tangible aid and service), informational support (i.e., advice and suggestions), and appraisal support (i.e., evaluative comments that could lead to self-improvement).⁵ All types of social support were potentially impacted by the isolation brought forward by the COVID-19 pandemic.³

Prior literature indicates that adequate social support is important for people living with HIV (PWH) and positively related to quality of life using a range of quality-of-life assessments, including those focused on behaviors, health, and subjective well-being.⁶ Since PWH are at an increased risk for depression and anxiety,⁷ adequate social support is even more important for this population’s health and well-being. Indeed, PWH with mental health disorders have been found to have lower quality of life than those without these conditions.⁸ PWH faced many challenges during the pandemic, with impacts on access to care, employment, income, housing, food insecurity, and increased stress, which many faced in situations of relative social isolation and loneliness.⁹ In a quantitative study of PWH conducted in Western Washington during the COVID-19 pandemic, we found that COVID-19-related stress was associated with higher depression and anxiety symptoms, but that greater social support was associated with lower depression symptoms.¹⁰ In a qualitative study based on interviews with a subset of participants in the quantitative study, we found that accessing social support was an important way that PWH coped with pandemic-related stressors.¹¹ Prior research has shown that women living with HIV report lower quality of life than men living with HIV.^12,13 In addition, as the number of PWH age 50 and older has increased, quality of life among older PWH has become an important focus of research and interventions.¹⁴

The goal of this study was to examine the relationship between social support and health-related quality of life (HRQoL) (i.e., that related to the impact of health on a person’s ability to live a fulfilling life) during the COVID-19 pandemic among PWH in Western Washington. In addition, we aimed to determine whether age and gender moderated the relationship between social support and quality of life.

Table 1 presents demographic characteristics of the 397 participants, overall and by gender. Overall, participant age ranged from 18 to 76 (mean 46) years, and 14.9% identified as female, 3.5% non-binary, and 81.6% as male. While 64.5% of participants were White, 13.1% were Black, 8.6% Latinx, 6.8% Asian, 4.0% Native American, and 2.0% Hawaiian/Pacific Islander. The mean social support score was 60.4 (standard deviation [SD] 18.3, range 12–84), and mean HRQoL score was 77.7 (SD 19.2, range 17–100). Female, non-binary, and male participants differed with respect to age, race/ethnicity, and HRQoL, with males being older, more often White, and having higher HRQoL scores. Supplementary Table S1 presents pairwise correlation coefficients for social support, HRQoL, GAD-7, and PHQ-8 scores, which were all significantly correlated (p < 0.0001). Supplementary Table S2 presents social support and quality of life domain scores and moderate-to-severe anxiety and depression prevalence, overall and by gender. There were significant differences in reported pain or discomfort, with female participants reporting more pain than non-binary or male participants (mean score 1.8 vs. 1.4 and 1.5, p = 0.0015).

Table 2 presents the results of regression analyses, with the first two columns showing results of models without interaction terms. In bivariable linear regression, social support, gender, depressive symptoms, anxiety symptoms, and crowded living situation were all significantly associated with HRQoL. In multivariable analysis with adjustment for age, gender, depressive symptoms, anxiety symptoms, and crowded living situation, higher social support score remained associated with higher HRQoL score (adjusted β = 1.05, 95% confidence interval [95% CI] 0.002, 2.10 per increment in social support score). The multivariable analyses also demonstrated that being male (adjusted β = 7.18, 95% CI 2.99, 11.37) was associated with significantly higher HRQoL scores than being female, while being non-binary (adjusted β 7.55, 95% CI −1.48, 16.59) was not. Being older (adjusted β = −0.17, 95% CI −0.30, −0.04 per year), more depressive symptoms (adjusted β = −1.11, 95% CI −1.55, −0.67 per increment in PHQ-8 score), and more anxiety symptoms (adjusted β = −1.02, 95% CI −1.45, −0.60 per increment in GAD-7 score) were associated with lower HRQoL scores.

When interaction terms were added for modeling, age did not moderate the association between social support and HRQoL in the model with only the social-support-by-age interaction (p = 0.114) or in a model with both interaction terms (p = 0.148, not presented). In the multivariable model without the social-support-by-age interaction term, adjusting for confounding variables, gender moderated the association between social support and HRQoL (see column 3, Table 2). The association between social support and HRQoL (adjusted β = −0.28, 95% CI: −0.50, −0.06) was significantly moderated by gender, with a weaker association observed among men compared with women. Non-binary/genderqueer individuals had HRQoL and slopes for the relationship between social support and HRQoL that did not differ significantly from those of females, with estimates closer to those of males than females. Figure 1 illustrates the marginal plot of the relationship between social support and HRQoL, stratified by gender. The plot demonstrates that while HRQoL increases with higher levels of social support across all groups, the slope is considerably steeper for women compared with male and non-binary participants. Predicted HRQoL score at the lowest possible social support score (1) was 55.6 for females, 76.7 for males, and 74.0 for non-binary people. At the highest possible social support score (12), the predicted HRQoL score for females (78.6), was very similar to predicted scores for males and non-binary individuals (79.4 and 81.2, respectively). Higher social support remained associated with higher HRQoL score (adjusted β = 3.84, 95% CI 1.43, 6.25 per increment in social support score). In addition, being older (adjusted β = −0.18, 95% CI −0.31, −0.05 per year), more depressive symptoms (adjusted β = −1.07, 95% CI −1.51, −0.63 per increment in PHQ-8 score), and more anxiety symptoms (adjusted β = −1.06, 95% CI −1.49, −0.64 per increment in GAD-7 score) were associated with lower HRQoL scores.

In a sensitivity analysis in which all individuals were classified as having either male or female sex at birth, results were very similar other than the omission of the non-binary/genderqueer group (Supplementary Table S1). The adjusted r-squared was almost identical (0.4176 for the model with gender and 0.4189 for the model with sex at birth), and the interaction between sex at birth and social support was significant (adjusted β −0.29, 95% CI −0.52, −0.06 for males compared with females).

This study revealed a positive association between social support and HRQoL among PWH during the COVID pandemic, with higher levels of perceived social support associated with higher self-rated HRQoL. This relationship persisted after adjustment for depressive symptoms and anxiety, which themselves were strong negative predictors of self-rated HRQoL. Importantly, gender moderates the relationship between social support and HRQoL. Specifically, at low levels of social support, female participants reported lower HRQoL compared with male and non-binary individuals, suggesting that women may be more vulnerable to reductions in social support. While age was an independent predictor of lower HRQoL, it did not moderate the association between social support and HRQoL.

Several studies of PWH have shown that social support directly relates to quality of life for individuals, in line with this study’s finding that HRQoL increased by 0.32 points for every 1-point increase in social support score in our final adjusted model. For example, a study of 120 PWH from Toronto, Canada reported a positive association between emotional social support and quality of life assessed by the Fanning Quality of Life Scale, a health-related measure.⁶ Also consistent with our current findings, in a qualitative analysis of in-depth interviews with a subsample of participants from the current study, individuals reported feelings of loneliness or lack of social support that resulted in a negative impact on quality of life.¹¹ While social support is generally important for maintaining HRQoL, its importance may increase during stressful events such as the COVID-19 pandemic.

While we found that social support was an independent predictor of HRQoL, our results also demonstrate that mental health (i.e., PHQ-8 and GAD-7 scores) was a strong predictor of HRQoL for our study participants. Abundant literature from before the COVID pandemic supports associations between mental health and quality of life. For example, a prepandemic study of 1241 individuals seeking care in 28 primary care centers in Spain found a strong relationship between depressive symptoms and lower quality of life, assessed using the World Health Organization Quality-of-Life Short Form (WHO QOL-BREF).²¹ A separate study of 2864 adults receiving care for HIV in the United States found that PWH with a probable mood disorder diagnosis had significantly lower quality of life scores (assessed using 28 items covering physical, mental, and social domains of life) than those without the same symptoms.⁸ The pandemic likely exacerbated this relationship. An editorial written early in the pandemic (April 2020) warned of the mental health consequences of COVID-19 and the measures required to prevent its spread, predicting a drastic increase in anxiety, depression, and loneliness.²² In our published analysis of data from the current study among PWH in Western Washington, we found that COVID-related stress remained positively associated with depression and anxiety symptoms during the pandemic, even after adjustment for prepandemic symptoms¹⁰; our HRQoL measure was not included in that study.

While some studies have found that age impacts the relationship between social support and quality of life, we did not find any moderation by age. However, higher age did predict a lower HRQoL. In a study of older adults living in England, Zaninotto et al. found that quality of life was poorer for older than younger respondents and associated with both decreased number of friends and low positive social support.²³ In contrast, we found that gender was both an important predictor of HRQoL and a moderator of the relationship between social support and HRQoL. Specifically, we found that women were more sensitive to the relationship between social support and HRQoL than men or nonbinary individuals. There are several reasons why this might be so. In a study of young adults across Norway, Johansen et al. found an association between low social support and mental distress that was significant for young women only. They hypothesized that this difference may be related to gender-specific stressors that develop from the socially defined roles of women and men.²⁴ Of interest, a literature review on gender differences in social support and physical health found that women report receiving more emotional support than men and are more likely than men to be both support providers and recipients.²⁵

Our findings are also consistent with prior studies that found that women are more likely to report lower quality of life than men, both generally and among PWH. For example, in a study conducted among the general population in England, PWH had lower HRQoL compared with the general population, and being female was associated with lower HRQoL than being male.²⁶ A systematic review of 12 studies regarding the impact of COVID-19 on the HRQoL of patients also found that women had lower HRQoL scores than men.²⁷ Of note, one of the ways our findings add to the body of work on gender and quality of life is by the inclusion of non-binary participants, who are frequently left out of research studies. All 14 non-binary participants in this study were assigned male sex at birth, which aligns with the fact that their results tracked closer to those of males than of females in this study.

The findings of this study have implications for interventions to mitigate the effects of future stressful events. If a new pandemic or any other significantly stressful life event were to occur, our results suggest that social support should be prioritized as an approach to buffer negative impacts on quality of life. A study of 69,066 participants who responded to the COVID-19 Participant Experience (COPE) Survey in 2022 found that social support, including emotional/informational support, positive social interactions, and tangible support, was associated with significantly reduced odds of depression,²⁸ which can be a cause for reduced quality of life. Similar to our study, women were more impacted than men or non-binary individuals, and the effects of gender (and age, in the COPE study) were attenuated at higher levels of social support.²⁸

Our study has several strengths. First, we were able to enroll a diverse sample of 397 PWH for online participation during a period when in-person research was not permitted due to COVID-19 social distancing. Second, the sample size was adequate to evaluate interaction terms for the continuous predictor and outcome we used. Third, responses were collected online using REDCap and not in a face-to-face interview, reducing the chance of self-presentation bias in responses. Our study also has its limitations. First, the generalizability of the results is limited as the sample was restricted to individuals who were enrolled in care at a participating clinic, were included in the UW HIV patient registry, and were willing to participate in the online survey. Participants also had to have internet access, which led to the exclusion of some otherwise eligible individuals without these resources. Second, while the proportion of women in our sample was similar to that in the UW HIV patient registry in 2020 (14.9% vs. 14.3%, respectively [unpublished data]), the number of women in our study was low. Third, some variables that may have been helpful in hindsight, such as household composition and access to online social support, were not collected. Fourth, there was the potential for recall bias in the self-reported responses from participants. Finally, our study design was cross-sectional, and we are not able to infer causal relationships based on these data.

In conclusion, we found a positive association between social support and HRQoL among PWH in Western Washington State during the COVID pandemic. This relationship was independent of depressive symptoms and anxiety and was moderated by gender. While older age predicted lower HRQoL, age did not moderate the relationship between social support and HRQoL. Individuals, especially females, and organizations providing services to PWH should prioritize social support as an important means to promote quality of life, especially during public health emergencies.

The authors would like to thank the UW HIV patient registry team for assistance with recruitment and also express sincere gratitude to their study participants for their contributions.