Participant characteristics
Table 1 shows the participants’ characteristics. A total of 1139 PLHIV were enrolled; their average age was 43.91 years (SD = 13.78; range: 14–84 years); and 74.36% were male. A total of 52.59% of participants had elementary or middle school education; 49.69% of the participants were employed or students, and the others were unemployed, retired, or otherwise. Further, 52.68% of participants were separated, divorced, or unmarried; 39.33% of participants had a monthly income of 0–1500 yuan. Regarding their lifestyle habits, 76.73% of the participants never smoked and 68.13% never drank alcohol. The average time since HIV diagnosis was 7.19 years (SD = 5.84). 56.01% of participants were without depression, and 43.99% were with mild to severe depressions.
Table 1 Descriptive statistics of the sociodemographic variables (N=1139).
Correlation analysis and regressions of perceived social support, stigma, adherence self-efficacy, and depression
Table 2 shows partial correlations among key variables after controlling for all covariates. All key variables were significantly associated. Perceived social support and adherence self-efficacy were negatively associated with stigma (r = -0.162, P < 0.001; r = -0.123, P < 0.001) and depression (r = -0.126, P < 0.001; r = -0.296, P < 0.001). Perceived social support was negatively associated with adherence self-efficacy (r = 0.138, P < 0.001). Stigma was positively associated with depression (r = 0.237, P < 0.001).
Table 2
Partial correlations between social support, stigma, adherence self-efficacy and depression.
|
Mean (SD)
|
Perceived social support
|
Stigma
|
Self-efficacy
|
Depression
|
Perceived social support
|
59.09(12.20)
|
-
|
|
|
|
Stigma
|
111.20(14.95)
|
-0.162***
|
-
|
|
|
Self-efficacy
|
110.53(14.60)
|
0.138***
|
-0.123***
|
-
|
|
Depression
|
13.46(10.35)
|
-0.126***
|
0.237***
|
-0.296***
|
-
|
Note: All correlations were controlled for age, gender, education, employment, marital status, income, smoking, drinking, and time since HIV diagnosis.*** P < 0.001 |
Table 3 shows the results of the regression analysis of key variables after controlling for the covariables. Perceived social support negatively predicted stigma (B = -0.210, P < 0.001). Perceived social support (B = 0.155, P < 0.001) and stigma (B = -0.102, P < 0.001) significantly predicted self-efficacy. Perceived social support (B = -0.049, P < 0.05), stigma (B = 0.128, P < 0.001), and self-efficacy (B = -0.175, P < 0.001) independently predicted depression.
Table 3
Regression analysis of the variables in the chain intermediary model.
Outcome variable
|
Predictor variable
|
B
|
SE
|
t
|
P
|
LLCI
|
ULCI
|
Stigma
(Model 1)
|
Perceived social support (path a1)
|
-0.210
|
0.038
|
-5.528
|
< 0.001
|
-0.284
|
-0.135
|
Self-efficacy
(Model 2)
|
Perceived social support (path a2)
|
0.155
|
0.038
|
4.1081
|
< 0.001
|
0.080
|
0.229
|
|
Stigma (path d21)
|
-0.102
|
0.029
|
-3.481
|
< 0.001
|
-0.160
|
-0.045
|
Depression
(Model 3)
|
Perceived social support (path c’)
|
-0.049
|
0.024
|
-2.042
|
< 0.05
|
-0.096
|
-0.002
|
|
Stigma (path b1)
|
0.128
|
0.019
|
6.903
|
< 0.001
|
0.091
|
0.164
|
|
Adherence self-efficacy (path b2)
|
-0.175
|
0.019
|
-9.346
|
< 0.001
|
-0.212
|
-0.138
|
Note: All regression analyses were controlled for age, gender, education, employment, marital status, income, smoking, drinking, and time since HIV diagnosis.
Model 1 summary: R = 0.255; R-sq = 0.065; F = 7.8739; P < 0.001.
Model 2 summary: R = 0.225; R-sq = 0.051; F = 5.461; P < 0.001.
Model 3 summary: R = 0.504; R-sq = 0.254; F = 31.986; P < 0.001.
The Mediating roles of Stigma and Adherence Self-efficacy in the Effect of Perceived Social Support on Depression
Table 4 and Fig. 1 show the results of the mediating effects of stigma and adherence self-efficacy and the path of the model. The results show that the mediation effect (B = -0.058, 95% CI [-0.078, -0.039]), direct effect (B = -0.049, 95% CI [-0.096, -0.002]), indirect effect 1 (B = -0.027, 95% CI [-0.043, -0.014]), indirect effect 2 (B = -0.027, 95% CI [-0.043, -0.013]), and indirect effect 3 (B = -0.004, 95% CI [-0.007, -0.010]) were all significant.
Table 4
Analysis of the mediating effect of stigma and adherence self-efficacy.
|
Effect size
|
SE
|
Boot LLCI
|
Boot ULCI
|
Relative mediation effect
|
Direct effect
|
-0.049
|
0.024
|
-0.096
|
-0.002
|
|
Indirect effect 1
|
-0.027
|
0.007
|
-0.043
|
-0.014
|
19.18%
|
Indirect effect 2
|
-0.027
|
0.008
|
-0.043
|
-0.013
|
19.33%
|
Indirect effect 3
|
-0.004
|
0.002
|
-0.007
|
-0.010
|
2.65%
|
Total mediation effect
|
-0.058
|
0.010
|
-0.078
|
-0.039
|
41.52%
|
Indirect effect 1: Perceived social support → Stigma → Depression.
Indirect effect 2: Perceived social support → Adherence self-efficacy → Depression.
Indirect Effect 3: Perceived Social Support → Stigma → Adherence self-efficacy → Depression.
The Moderating roles of Stigma and Adherence Self-Efficacy in the Effect of Perceived Social Support on Depression
Figure 2 shows the paths of the three moderation effect models. Table 5 shows the moderating roles of stigma and adherence self-efficacy. The results demonstrated that stigma (B = -0.003, P < 0.05), adherence self-efficacy (B = 0.003, P < 0.05), and the joint effect of stigma (B = -0.003, P < 0.05) and adherence self-efficacy (B = 0.004, P < 0.05) had moderating effects on the association between perceived social support and depression.
Table 5
Analysis of the moderating effect of stigma and adherence self-efficacy
|
Model 1
B (SE)
|
Model 2
B (SE)
|
Model 3
B (SE)
|
Perceived social support
|
-0.074 (0.025) **
|
-0.074 (0.024) **
|
-0.047 (0.024) *
|
Stigma
|
0.152 (0.019) ***
|
-
|
0.137 (0.019) ***
|
Adherence self-efficacy
|
-
|
-0.180 (0.019) ***
|
-0.165 (0.019) ***
|
Stigma * perceived social support
|
-0.003 (0.001) *
|
-
|
-0.003 (0.001) *
|
Adherence self-efficacy * perceived social support
|
-
|
0.003 (0.002) *
|
0.004 (0.002) *
|
Note: Model 1 presents the moderating effect of stigma on perceived social support on depression. Model 2 presents the moderating effect of perceived social support on depression through self-efficacy. Model 3 presents the moderating effect of perceived social support on depression through stigma and adherence self-efficacy.
Model 1 summary: R = 0.446; R-sq = 0.199; F = 23.331; P < 0.001.
Model 2 summary: R = 0.475; R-sq = 0.225; F = 27.293; P < 0.001.
Model 3 summary: R = 0.511; R-sq = 0.262; F = 28.427; P < 0.001.
*P < 0.05, **P < 0.01, ***P < 0.001.
Figure 3 shows that high stigma levels increased the association between perceived social support and depression. The difference in depression scores between the high-stigma group and the low-stigma group became smaller as perceived social support scores increased. Low adherence self-efficacy increased the association between perceived social support and depression. Adherence self-efficacy moderated the association between perceived social support and depression, and as social support increased, the difference in depression scores between the “high adherence self-efficacy” group and the “low adherence self-efficacy” group became smaller. The association between social support and depression was moderated by the “high stigma and low self-efficacy” group as well as the “low stigma and high self-efficacy” group, but the results were similar for the “both high” group and “both low” group. As the perceived social support scores increased, the difference in depression scores between the “high stigma and low self-efficacy” group and the “low stigma and high self-efficacy” group became smaller. In the case of the “high stigma and low self-efficacy” group, depression scores decreased as perceived social support scores increased, whereas in the case of the “low stigma and high self-efficacy” group, depression scores increased as perceived social support increased. The “both high” and “both low” groups had moderating effects compared to the “high stigma and low self-efficacy” group, and both had lower depression scores than the “high stigma and low self-efficacy” group, with depression decreasing as perceived social support increased.