Also available as PDF (figures and tables appear in the PDF).
Background: This cross-sectional study examined the effects of face-to-face (real) and virtual social support networks on the mental health status of Japanese HIV-positive males. Method: A nationwide online cross-sectional survey was conducted by HIV Futures Japan project from July 2013 to February 2014: 1,095 responses were obtained, and we analyzed those of 879 not females. Two subscales of the Hospital Anxiety and Depression Scale (HADS) were measured in depressive and anxiety tendencies. Results: The mean (SD) scores for depressive and anxiety tendencies were 8.1 (4.2) and 8.6 (4.8), respectively (the range was 0-21). Hierarchical regression analysis showed that virtual support networks did not display direct effects on depressive and anxiety tendency. “No one”, “only one” and “two to three” in the real-support network showed significant relations compared to “20 or more” (B=2.76, 1.86, 1.46, respectively. 95% CIs [1.35, 4.17], [0.28, 3.44], [0.10, 2.82], respectively) to the depressive tendency. With anxiety tendency, “no one”, “only one” and “two to three” in the real-support network showed a significant relation compared to “20 or more” (B=2.42, 2.04, 1.61, respectively. 95% CIs [0.83, 4.00], [0.27, 3.81], [0.08, 3.14], respectively). Conclusions: It is necessary to develop face-to-face social support systems to promote mental health among people living with HIV/AIDS.
Keywords: Social support; Mental health; Virtual-support network; HIV/AIDS; Japan; HIV/AIDS, Japan
Since 2007, around 1,500 new HIV/AIDS patients have been reported annually in Japan. According to the National Institute of Infectious Disease, the number of HIV/AIDS patients in the country was 23,000 in 2013.
People living with HIV (PLHIV) have to suffer internalized stigma.1 This stigma causes stress, affects their mental health,2 and presents a long-standing problem for care of PL- HIV. Gay PLHIV are subject to social segregation, rejection in their relationships, and social withdrawal.3 The World Health Organization (WHO) quality of life (QoL) group has underlined the importance of social inclusion for PLHIV.4 Positive social support for PLHIV is reflected in a good mental health status,5 6 7 medication adherence,6 and appropriate risk behavior.7
In recent years, personal connections via the Internet have grown alongside the development of social networking services. Social support has increasingly adopted the form of virtual communities and electronic support groups.8 Social support in online communities currently has a similar level of acceptance to that in face-to-face situations.9
The aim of the present study was to examine how faceto-face (real) and virtual social support networks for HIV/AIDS are related to mental health status in Japanese HIV-positive males.
In this cross-sectional study, we used data from a national online survey conducted by HIV Futures Japan from July 2013 to February 2014. The HIV Futures Japan national survey was performed as part of the HIV Futures Japan project, which is intended to help realize healthy, independent lifestyles and create a liveable social community for HIV-positive individuals.10 In Japan, there have been many surveys that measured medicinal and curative aspects. Moreover, there have not developed adequately HIV sero-positive registration system in Japan. However, that research did not include aspects of members of the general population living with HIV/AIDS. Therefore, a multidimensional and multidisciplinary survey for HIV-positive individuals was needed.
The questionnaire in this survey was composed of the following large categories: (1) about yourself (socio-demographic items), (2) health status, (3) medicine, (4) sexual health, (5) substance use and addiction, (6) having children, (7) social relations, (8) mental health, (9) social welfare/health management and (10) others. The survey posted online based on self-reported HIV infection, and subjects were recruited in 2 ways. First, we set up online links and banners to promote the survey site. We began by placing links on general sites directed at HIV-positive groups, sites for HIV-positive groups, social networking sites for HIV-positive participants, and so on. Second, we distributed the fliers at medical facilities, HIV-related events, and meetings of HIV-prevention groups and support groups.
After excluding incomplete answers, we obtained responses from 1095 individuals. We determined 913 responses to be valid, and we subjected these to analysis. The survey participants were resident in 46 of Japan’s 47 prefectures (the exception was Tottori).
Of the valid responses received, 879 subjects remained for analysis after we excluded females. Descriptive statistics of subjects were shown in Table 1. The mean age of the participants was 38.2 years (standard deviation [SD], 8.1 years); the average number of years since testing positive for HIV was 5.9 years (SD, 4.9 years).
Virtual- and real-support network for HIV/AIDS: In the survey, we inquired whether participants were able to talk about HIV/AIDS online (virtual support) and whether they were able to speak about it directly (real support). We defined each 6 categories: (1) no one; (2) only 1; (3) 2-3; (4) 4-9; (5) 10-19; and (6) 20 or more.
Mental health status: We assessed mental health status by means of the HADS.11 This scale comprises 2 subscales-depression and anxiety. Each subscale is made up of 7 items assessed on a fivepoint Likert scale. Cronbach’s alpha coefficients are 0.75 for the depression subscale and 0.85 for the anxiety subscale.
Demographics and time since HIV infection: In the survey, we inquired about age, sexual orientation (gay, bisexual, heterosexual, other) and number of years since HIV infection.
We used the depression and anxiety subscales as continuous variables for depressive and anxiety tendencies. We performed a multiple regression analysis estimated by restricted maximumlikelihood method, in which the dependent variables were depressive tendency and anxiety tendency; the independent variables were age, sexual orientation, number of years since HIV infection, virtual-support network, and real-support network. In the hierarchical regression model (model 1), the independent variables were age, sexual orientation, number of years since HIV infection and the virtual-support network. In model 2, the real-support network was added to model 1. All statistical analysis were performed using IBMSPSS, version 19.0 (IBM Corporation).
The mean (SD) scores for depressive and anxiety tendencies were 8.1 (4.2) and 8.6 (4.8), respectively (the range was 0-21).
Sexual orientation gay 717 (81.6) bisexual 50 (5.7) heterosexual 95 (10.8) other 17 (1.9) Virtual-support network on HIV/AIDS no one 441 (50.2) only one 40 (4.6) two to three 182 (20.7) four to nine 85 (9.7) 10–19 44 (5.0) 20 or more 82 (9.3) Real-support network on HIV/AIDS no one 199 (22.6) only one 93 (10.6) two to three 309 (35.2) four to nine 143 (16.3) 10–19 65 (7.4) 20 or more 66 (7.5) Age 38.2 (8.1)
Number of years since HIV infection 6.9 (4.9) aF-test in one-way ANOVA bmean (SD) cPearson's product-moment correlation coefficient
Table 1. Descriptive statistics and destribution of depression and anxiety score (n=879).
Table 2 presents the results of hierarchical regression analysis. In terms of depressive tendency, “no one” and “two to three” in the virtual-support network showed significant relations in model 1, (B=1.65, 95% CI [0.63, 2.67], B=1.29, 95% CI [0.16, 2.42]), though there was no significant relation in model 2. “No one”, “only one” and “two to three” in the real-support network showed significant relations (B=2.76, 1.86, 1.46, respectively. 95% CIs [1.35, 4.17], [0.28, 3.44], [0.10, 2.82], respectively) to the depressive tendency in model 2.
With anxiety tendency, however, the virtual-support network evidenced no significant relation in models 1 and 2. “No one”, “only one” and “two to three” in the real-support network showed a significant relation (B=2.42, 2.04, 1.61, respectively. 95% CIs [0.83, 4.00], [0.27, 3.81], [0.08, 3.14], respectively) to depressive tendency in model 2.
We found that the mechanisms related to the social support network for HIV/AIDS differed with respect to depression and anxiety. In the case of depression, virtual social support was indirectly related to depression mediation by the real-support network. However, the virtual-support network showed no relation
.896 .201 8.1 (4.3) 8.5 (4.8) 8.5 (4.3) 8.9 (4.6) 7.9 (3.7) 8.9 (4.4) 8.1 (4.4) 10.0 (6.4) .048 .239 8.4 (4.4) 8.6 (4.8) 8.0 (4.0) 8.9 (4.8) 8.1 (4.2) 8.9 (4.6) 8.0 (4.0) 9.0 (4.8) 7.4 (3.4) 8.0 (4.5) 6.9 (4.2) 7.4 (4.8) <.001 .001 9.3 (4.3) 9.3 (4.6) 8.3 (4.7) 9.0 (4.9) 8.0 (3.9) 8.7 (4.6) 7.8 (3.9) 8.4 (4.8) 7.2 (4.5) 7.5 (5.2) 6.5 (4.3) 6.6 (4.8) b c c -0.015 -0.187 b c c -0.045 -0.133
relation with respect to anxiety.
This study shows that virtual support did not have a direct effect on mental health. In contrast to online support groups, Setoyama et al9 identified a high degree of emotional insight as part of the peer support function in face-to-face support groups. The authors believed this to be possibly the result of participants being able to establish a closer relationship with other group members in face-to-face support. As elsewhere in the world,12 Japanese PLHIV are subject to severe mental health problems. In this research conducted by HIV Futures Japan, we found high proportion of high-risk depressive (28.7%) and anxiety (33.1%) disorders which are higher than those in the general Japanese population.
Also, a social phenomenon called ‘AIDS panic’ occurred in 1990’s Japanese society. Research which was conducted to find quality of life of medically induced PLHIV in Japan reported a lot of felt stigmas were in Japanese PLHIV at that time, and these were related to social isolation and negative mental health.13 In this HIV Futures Japan survey, we found almost Japanese PLHIV still remain in having felt stigma. For example, 35% of participants answered “Since I am HIV+, I hardly interact with others”. It is therefore necessary to provide
Intercept 6.32 (3.58 9.06) <.001 5.25 (2.41 8.09) <.001 12.46 (9.41 15.52) <.001 11.45 (8.26 14.64) <.001
Virtual-support network on HIV/AIDS no one 1.65 (0.63 2.67) .002 0.52 (-0.68 1.73) only one 1.24 (-0.39 2.88) .136 0.20 (-1.58 1.97) two to three 1.29 (0.16 2.42) .025 0.41 (-0.90 1.72) four to nine 1.24 (-0.06 2.54) .062 0.57 (-0.86 1.99) 10–19 0.61 (-0.95 2.17) .442 0.12 (-1.55 1.79) 20 or more ref. ref.
Real-support network on HIV/AIDS no one 2.76 (1.35 4.17) <.001 2.42 (0.83 4.00) .003 only one 1.86 (0.28 3.44) two to three 1.46 (0.10 2.82) four to nine 1.28 (-0.14 2.70) 10–19 0.77 (-0.76 2.30) 20 or more ref.
-2 Restricted Log Likelihood 4957.56 4913.74 All models were adjusted for age, number of years since HIV infection and sexual orientation. CI: confidential interval afixed effect model estimated by restricted maximum-likelihood method
Table 2. Hierarchical multiple regression analysis about effects of virtual and real support networks on mental health.
This study has the following implications. It is necessary to establish face-to-face support systems to promote mental health among PLHIV. To reduce depressive tendencies, it is important that virtual support systems for HIV/AIDS be designed to facilitate the creation of face-to-face support networks.
This study has a number of limitations. First, we relied on online survey data, which limited participants to those with access to the Internet, such as individuals with smart phones or personal computers, and that may have led to selection bias. It would be appropriate to confirm the generalizability of our findings by comparing our online survey results with those acquired by other means, such as from mailed surveys. Second, this was a cross-sectional study. It is therefore necessary to clarify the relationships among the studied factors using a vertical study design through follow-up surveys. Finally, we did not examine the various kinds of social support, such as instrumental and emotional: such aspects should be measured and analyzed in future research.
Advice and comments given by HIV Futures Japan reference group members has been a great help in this study. The authors would like to express the deepest appreciation to them. This work was supported by Japan Society for the Promotion of Science KAKENHI under grant number 24330158 and 26670936.
.395 0.79 (-0.34 1.93) .171 -0.35 (-1.70 1.01) .616 .828 0.53 (-1.29 2.35) .568 -0.62 (-2.62 1.37) .539 .540 1.01 (-0.24 2.27) .114 0.06 (-1.41 1.54) .934 .437 1.14 (-0.32 2.59) .125 0.39 (-1.22 1.99) .635 .889 0.24 (-1.49 1.98) .782 -0.29 (-2.17 1.59) .764 ref. ref.
.021 2.04 (0.27 3.81) .024 .035 1.61 (0.08 3.14) .039 .077 1.29 (-0.30 2.89) .113 .322 0.51 (-1.20 2.23) .558 ref.
5140.82 5109.56
The authors declare that they have no conflicts of interest.
This study was approved by Open University of Japan Research Ethics Committee (No. 2013-5) and Osaka National Hospital Institutional Review Board (IRB) (No. 13-037).