This paper is in the following e-collection/theme issue:

Web-based and Mobile Health Interventions (3297) Medicine 2.0: Social Media, Open, Participatory, Collaborative Medicine (1641) HIV/AIDS/STI Prevention and Care (873) Depression and Mood Disorders; Suicide Prevention (1363)

Published on in Vol 21, No 11 (2019): November

Preprints (earlier versions) of this paper are available at https://preprints.jmir.org/preprint/14729, first published .
Mechanisms and Effects of a WeChat-Based Intervention on Suicide Among People Living With HIV and Depression: Path Model Analysis of a Randomized Controlled Trial

Mechanisms and Effects of a WeChat-Based Intervention on Suicide Among People Living With HIV and Depression: Path Model Analysis of a Randomized Controlled Trial

Mechanisms and Effects of a WeChat-Based Intervention on Suicide Among People Living With HIV and Depression: Path Model Analysis of a Randomized Controlled Trial

Authors of this article:

Yiran Li1 Author Orcid Image ;   Yan Guo1, 2, 3 Author Orcid Image ;   Y Alicia Hong4 Author Orcid Image ;   Mengting Zhu1 Author Orcid Image ;   Chengbo Zeng5, 6 Author Orcid Image ;   Jiaying Qiao1 Author Orcid Image ;   Zhimeng Xu1 Author Orcid Image ;   Hanxi Zhang7 Author Orcid Image ;   Yu Zeng1 Author Orcid Image ;   Weiping Cai8 Author Orcid Image ;   Linghua Li8 Author Orcid Image ;   Cong Liu8 Author Orcid Image
Article Authors Cited by (18) Tweetations (2) Metrics

    Original Paper

    1Department of Epidemiology and Biostatistics, School of Public Health, Sun Yat-sen University, Guangzhou, China

    2Sun Yat-sen Center for Migrant Health Policy, Guangzhou, China

    3Sun Yat-sen Center for Global Health, Institute of State Governance, Guangzhou, China

    4Department of Health Administration and Policy, College of Health and Human Services, George Mason University, Fairfax, VA, United States

    5South Carolina SmartState Center for Healthcare Quality, Arnold School of Public Health, University of South Carolina, Columbia, SC, United States

    6Department of Health Promotion, Education, and Behavior, Arnold School of Public Health, University of South Carolina, Columbia, SC, United States

    7National Center of AIDS/STD Control and Prevention, Chinese Center for Disease Control and Prevention, Beijing, China

    8Department of Infectious Diseases, Eight People’s Hospital, Guangzhou, China

    *these authors contributed equally

    Corresponding Author:

    Yan Guo, PhD

    Department of Epidemiology and Biostatistics

    School of Public Health

    Sun Yat-sen University

    #74 Zhongshan 2nd Road

    Guangzhou,

    China

    Phone: 86 020 87334202

    Email: guoy8@mail.sysu.edu.cn


    Background: People living with HIV and depression have high rates of suicide. Studies of mobile health (mHealth) interventions have shown feasibility, acceptability, and efficacy in improving mental health in people living with HIV and depression. However, few studies have examined the mechanisms and effects of mHealth interventions on suicide.

    Objective: This study was designed to examine the mechanisms and effects of a WeChat-based intervention, Run4Love, on suicide among people living with HIV and depression in China, while considering perceived stress and depressive symptoms as mediators.

    Methods: A sample of 300 People living with HIV and depression was recruited from the outpatient clinic of a large HIV or AIDS treatment hospital and was randomized to the Run4Love group or a control group. Data were collected at baseline, 3-, 6-, and 9-month follow-ups. Path analysis modeling, with longitudinal data, was used in data analyses.

    Results: The Run4Love mHealth intervention had a direct effect on reducing suicide rate at the 6-month follow-up (beta=−.18, P=.02) and indirect effect through reducing perceived stress and depressive symptoms at the 3-month follow-up (beta=−.09, P=.001). A partial mediating effect between perceived stress and depressive symptoms accounted for 33% (–0.09/–0.27) of the total effect.

    Conclusions: Through path analyses, we understood the mechanisms and effects of an mHealth intervention on suicide prevention. The findings underscored the importance of stress reduction and depression treatment in such a program. We call for more effective suicide prevention, especially mHealth interventions targeting the vulnerable population of people living with HIV and depression.

    Trial Registration: Chinese Clinical Trial Registry ChiCTR-IPR-17012606; http://www.chictr.org.cn/showprojen.aspx?proj=21019

    J Med Internet Res 2019;21(11):e14729

    doi:10.2196/14729

    Keywords

    HIV; mHealth; depression; suicide 


    Background

    People living with HIV have high rates of depressive symptoms. For example, a nationally representative survey in the United States reported that 36% of people living with HIV had depression [1]. In China, the rate of depression among people living with HIV was 50.8%, according to a recent review study [2]. Out of 37.9 million people living with HIV in the world [3], more than 12 million are living with depression [4]. The literature has shown a causal relationship between depression and suicide [5,6]. Previous studies have also reported high suicide rates in people living with HIV. A survey of 1560 people living with HIV in the United States revealed that 26% of the participants had a suicidal ideation and 13% had a suicidal attempt [7]. According to a recent survey in China, 32.4% of people living with HIV had had suicidal ideation or behavior since HIV diagnosis [8]. Such high rates of suicide and depression warrant effective interventions targeting the vulnerable population of people living with HIV and depression, especially the interventions that can reach a large number of people living with HIV and depression efficiently, such as mobile health (mHealth) interventions.

    Reviews

    According to a recent systematic review, the majority of the existing mHealth interventions targeting people living with HIV were focused on medication adherence or retention in care [9]. Only a small number of mHealth interventions were implemented to improve mental health outcomes or decrease suicidal risks among people living with HIV and depression [9-11]. In a pilot mHealth intervention, Swendeman and colleagues developed a self-monitoring program, which showed initial efficacy in reducing stress in people living with HIV [12]. Van Luenen and colleagues tested a Web-based cognitive behavioral therapy program via a randomized controlled trial (RCT) with 188 people living with HIV and depression. The intervention successfully reduced depressive symptoms and anxiety of the intervention group at 8-week and 12-week follow-ups compared with the control group [13]. Despite the growing interest and initial efficacy of mHealth interventions to deliver mental health services and prevent suicide among people living with HIV and depression, no study existed on the potential mechanism of the intervention in mental health outcomes, especially on the basis of a longitudinal design.

    Aims

    Accordingly, in this study, we aimed to examine the mechanisms of the mHealth intervention Run4Love in a suicide outcome in a sample of people living with HIV and depression by conducting path model analyses, using longitudinal data from an RCT [14]. The Run4Love was a WeChat-based intervention, comprising 2 major components: the adapted cognitive-behavioral stress management (CBSM) course and physical activity promotion. It was designed to reduce stress through multiple coping strategies.

    The literature has documented the importance of stress reduction in depression prevention [15-18]. Previous studies also showed that depressive symptoms mediated perceived stress and suicidal behaviors [19-21]. However, no study has investigated the mechanisms of stress and depressive symptoms in suicide in the context of mHealth intervention. On the basis of the associations among perceived stress, depressive symptoms, and suicide illustrated in the previous studies [19-21], we hypothesized that perceived stress and depressive symptoms at a 3-month follow-up served as mediators on the effects of the mHealth intervention on suicide at a 6-month follow-up. The hypothesized model is depicted in Figure 1.

    ‎
    Figure 1. Hypothesized path model of intervention, perceived stress, depressive symptoms, and suicide in people living with HIV and depression.
    View this figure

    Research Setting

    A total of 300 participants were recruited from an outpatient clinic of the only designated hospital for HIV or AIDS treatment in Guangzhou, China. As the capital city of Guangdong Province, Guangzhou is the third largest city in China. According to the AIDS Prevention and Treatment Report in 2017, more than 53,600 people living with HIV live in Guangdong province [22].

    Participant Eligibility and Recruitment

    We recruited participants in the waiting room of an outpatient clinic. A member of the research team invited patients to participate in a research study. Those who showed initial interest were invited to a private space for further explanation and screening. The participation criteria included the following: (1) at least 18 years old, (2) HIV-seropositive status (registered in the hospital system or with an official document), (3) having clinically significant depressive symptoms (Center for Epidemiologic Studies-Depression; CES-D score≥16), and (4) having a mobile phone and WeChat account. Patients were excluded if they were (1) currently under psychiatric or psychological treatment, (2) unable to finish the questionnaire, (3) unable to read or listen to the materials on WeChat, and (4) unable to engage in physical activities because of medical reasons.

    Participants who met the eligibility criteria and provided written informed consent were invited to participate. They completed a baseline survey before they were randomized to the intervention group or control group of the Run4Love trial. All participants received breakfast (milk and bread) as an incentive. The study protocol was approved by the Institutional Review Board of the Sun Yat-sen University.

    The Run4Love Intervention

    The Run4Love intervention was adapted from the CBSM program [23], which was designed as a face-to-face intervention program for people living with HIV, and it has demonstrated good efficacy in various populations and settings [24-26]. We adapted the CBSM into a multimedia format and delivered it via WeChat, the most popular social media platform in China, with over 1 billion active users [27]. The intervention protocol is detailed elsewhere [14]. Briefly, participants assigned to the intervention group received a series of adapted CBSM courses, and they were encouraged to do regular physical activities for 3 months. Multimedia information, with automatic progress tracking and personalized feedback, was delivered via the Run4Love WeChat account (Figure 2). The main purpose of the program was to alleviate people living with HIV’s depressive symptoms and improve their quality of life by reducing perceived stress. Participants in the wait-list control group received a brochure on HIV-related nutrition. Outcomes were assessed at baseline, 3, 6, and 9 months. All participants in the control group would receive the intervention after the completion of the trial.

    ‎
    Figure 2. WeChat users’ interfaces in the Run4Love Program. CBSM: cognitive-behavioral stress management.
    View this figure

    Measurements

    Demographic Characteristics

    Participants’ demographic characteristics assessed in the study included age, gender, sexual orientation, marital status, education, and duration since HIV diagnosis.

    Depressive Symptoms

    Depressive symptoms were assessed using the Chinese version of the CES-D scale at baseline and each follow-up [28]. The scale has been validated in various Chinese chronic patients, and it demonstrated high levels of validity and reliability [29,30]. The CES-D comprises 20 items, such as My appetite was poor and I could not shake off the blues. Participants were asked to select the frequencies of physiological symptoms during the past week. The total scores ranged from 0 to 60, with a higher score indicating a higher level of depressive symptoms. Participants with scores of 16 or higher were considered as having clinically significant depressive symptoms [31]. The Cronbach alpha of the scale was .77 in this study.

    Perceived Stress

    Perceived stress was assessed with the 10-item Chinese version of Perceived Stress Scale (PSS-10) at baseline and each follow-up [32]. The scale was the most widely used instrument to measure perception of stress; the validity and reliability of the scale had been validated and established in the Chinese population [33-35]. Participants were asked about their feelings and thoughts in the past month (eg, How often have you felt nervous and stressed?). The total scores of the scale ranged from 0 to 40, with higher scores indicating higher levels of perceived stress. Scores ranging from 0 to 13, 14 to 26, and 27 to 40 were considered as low, moderate, and high levels of perceived stress, respectively [36]. In this study, the Cronbach alpha of the PSS-10 was .67.

    Suicide

    Suicide was defined as having had suicidal ideation or attempt in this study. Suicidal ideation referred to the planning or thinking of committing suicide, and suicidal attempt was the actual behavior of committing suicide. Participants were asked how many times they had had thought about committing suicide and how many times they had actually attempted suicide in the past 3 months at baseline and each follow-up. Participants who answered yes to either of the 2 questions were considered as having had suicide. Those who answered no to both of the questions were considered as not having had suicide (dummy coded as 0=not having had suicide, 1=having had suicide).

    Data Analysis

    First, descriptive analyses were used on demographic characteristics, perceived stress, depressive symptoms, and suicide. The continuous variables with skewed distribution (eg, age, duration since HIV infection, perceived stress, and depressive symptoms) were described using median (interquartile range; IQR), and categorical variables were described using frequencies and percentages.

    Second, treatment group comparisons for the outcome and mediating variables at 3 assessment points were conducted using Wilcoxon rank-sum test for continuous variables with skewed distribution (eg, perceived stress and depressive symptoms) and Chi-square test for categorical variables (eg, suicide). Third, bivariate analyses of suicide were performed using Wilcoxon rank-sum test (for continuous variables with skewed distribution) and Chi-square test (for categorical variables).

    To further examine the mechanisms of how participants’ suicide changed as a result of the intervention, we conducted a path analysis after controlling for the baseline suicide and potential confounders, such as demographics. A mediation model of the path analysis was used to test the hypothesis whether the intervention effect on suicide could be explained by the mediating factors of perceived stress and depressive symptoms. Such modeling was designed to explore the mechanisms and direct and indirect effects of the intervention on suicide. The pathways of intervention→suicide, intervention→depressive symptoms→suicide, intervention→perceived stress→suicide, and intervention→perceived stress→depressive symptoms→suicide were examined, respectively (dummy coded as 0=control group, 1=intervention group). The statistical significance was defined as P<.05.

    The model was estimated by weighted least squares means and variance–adjusted estimation. Multiple indicators were used to evaluate the goodness of fit of the model, including Chi-square statistic, Comparative Fit Index (CFI), Tucker-Lewis Index (TLI), root mean square error of approximation (RMSEA), and weighted root mean square residual (WRMR). Smaller Chi-square value indicates better model fit. CFI≥0.95, TLI>0.90, RMSEA≤0.06, and WRMR≤1.00 indicate good model fit [37,38]. Descriptive statistics, bivariate statistics, and correlation analyses were performed using SAS version 9.4 (SAS Institute, Inc). Path analysis was performed using Mplus version 7.0 (Muthen and Muthen) [39].


    Baseline Characteristics of the Participants

    As shown in Table 1, the median (IQR) age of the 300 participants was 27.5 (24.5, 31.3) years, with a range from 18 to 51. Most (277/300, 92.3%) of the participants were male. More than four-fifths (245/300, 81.7%) of the participants were homosexual, bisexual, or uncertain of their sexual orientations, and 87.3% (262/300) were unmarried. About 60.7% (182/300) of the participants had completed at least some college education. The median duration of HIV infection was 1.7 years (IQR 0.6, 3.8).

    At baseline, a majority (256/300, 85.4%) of the participants had a moderate level of perceived stress, followed by 8.3% (25/300) having a high level and 6.3% (19/300) a low level of stress. The median (IQR) of the CES-D score was 23.0 (19.0, 28.0). Approximately 44.0% (132/300) of the participants reported having considered committing suicide and 9.7% (29/300) having tried to attempt suicide in the past 3 months. In total, about 45.0% (135/300) of the participants had suicidal ideation or attempts in the past 3 months.

    The proportion of homosexual, bisexual, or sexual orientation–uncertain participants in the intervention group was slightly higher than the control group (130/150, 86.7% vs 115/150, 76.7%; P=.03). Other demographic characteristics, mental health outcomes (depressive symptoms and perceived stress), and suicide were balanced in the 2 groups at baseline.

    Table 1. Sample characteristics of people living with HIV and depression (N=300).
    CharacteristicsValue
    Age (years), median (IQRa)27.5 (24.5, 31.3)
    Gender, n (%)

    		Male277 (92.3)

    		Female23 (7.7)
    Sexual orientation, n (%)

    		Heterosexual55 (18.3)

    		Homosexual/bisexual/uncertain245 (81.7)
    Education, n (%)

    		≤High school118 (39.3)

    		>High school182 (60.7)
    Marital status, n (%)

    		Single262 (87.3)

    		Married38 (12.7)
    Employment status, n (%)

    		Unemployed49 (16.3)

    		Employed251 (83.7)
    Duration since HIV diagnosis (years), median (IQR)1.7 (0.6, 3.8)
    Perceived stress, median (IQR)20.0 (18.0, 23.0)
    Levels of perceived stress, n (%)

    		Low perceived stress19 (6.3)

    		Moderate perceived stress256 (85.3)

    		High perceived stress25 (8.3)
    Depressive symptoms, median (IQR)23.0 (19.0, 28.0)
    Suicidal ideation (times in the last 3 months), n (%)

    		0168 (56.0)

    		1-284 (28.0)

    		≥348 (16.0)
    Suicidal attempt (times in the last 3 months), n (%)

    		0271 (90.3)

    		1-219 (6.3)

    		≥310 (3.3)
    Suicide (in the last 3 months), n (%)

    		Yes135 (45.0)

    		No165 (55.0)

    aIQR: interquartile range.

    Changes in Outcomes and Mediating Variables Over Time

    The intervention had significant effects on the mental health outcomes over time. A total of 274 participants (91.3%, 274/300) remained at the 3-month follow-up (139 in the intervention group; 135 in the control group), and 265 (88.3%, 265/300) participants remained at the 6-month follow-up (132 in the intervention group; 133 in the control group). As reported in Table 2, the median score of the perceived stress was 20 at baseline in the intervention and control groups; it changed to 17 in the intervention group compared with 19 in the control group at the 3-month follow-up. Similarly, the depressive symptoms in the intervention group reduced from 23 at baseline to 17 at the 3-month follow-up, whereas in the control group, it remained unchanged at 23. The rate of clinically significant depressive symptoms reduced from 100% to 56.8% (79/139) in the intervention group compared with 79.3% (107/135) in the control group at the 3-month follow-up. The rate of suicide reduced from 44.0% (66/150) at baseline to 31.7% (44/139) at the 3-month follow-up, and it further reduced to 21.2% (28/132) at the 6-month follow-up in the intervention group. The difference in the rate of suicide between the 2 groups was close to 20% (21.2% in intervention group vs 40.6% in control group) at 6 months. The intervention group had improved significantly in 3 variables—perceived stress at the 3- and 6-month follow-ups (both P<.001), depressive symptoms at the 3- and 6-month follow-ups (both P<.001), and suicide at the 6-month follow-up (P<.001).

    Table 2. Mental health and behavior outcomes at baseline and follow-ups.
    VariablesBaseline (T0)3 months (T1)6 months (T2)
    Ia (n=150)Cb (n=150)P valueI (n=139)C (n=135)P valueI (n=132)C (n=133)P value
    Perceived stress, median (IQRc)20 (17, 22)20 (18, 23).14d17 (12, 20)19 (16, 22)<.001d17 (13, 20)19 (16, 23)<.001d
    Depressive symptoms, median (IQR)23 (19, 28)23 (19, 27).81d17 (11, 24)23 (18, 31)<.001d17 (10, 23)24 (16, 32)<.001d
    Clinically significant depressive symptoms, n (%)150 (100.0)150 (100.0)—e79 (56.8)107 (79.3)<.001f68 (51.5)102 (76.7)<.001f
    Suicide, n (%)66 (44.0)69 (46.0).82f44 (31.7)49 (36.3).42f28 (21.2)54 (40.6)<.001f

    aIntervention group.

    bControl group.

    cIQR: interquartile range.

    dWilcoxon rank-sum test.

    eChi-square test is not applicable in this cell. All participants had depressive symptoms in both the intervention and control groups.

    fChi-square test.

    Bivariate Analysis of Suicide

    Results of bivariate analyses between demographic characteristics and suicide indicated that only employment was significantly associated with suicide. Specifically, employed participants reported lower rates of suicide than the unemployed (104/251, 41.4% vs 31/49, 63.3%, P=.005). Therefore, employment should be controlled as a covariate in the hypothesized path model.

    Path Model

    After 3 months of the Run4Love intervention, 16 (5.3%) participants were lost to follow-ups, without their data of perceived stress and depressive symptoms at the 3- and 6-month assessments. The characteristics of the 16 participants were not significantly different from the remaining participants. The path model was estimated using data from the remaining 284 participants. The hypothesized model was examined, and there were 2 significant pathways: intervention→suicide and intervention→perceived stress→depressive symptoms→suicide. In this model, perceived stress did not appear to be significantly related to suicide (beta=−.05, P=.71), and the intervention did not affect depressive symptoms directly either (beta=−.06, P=.08). In addition, both employment and suicide at baseline were controlled as potential confounders in the hypothesized model, whereas only suicide showed statistical significance (beta=.44, P<.001). Consequently, employment as a covariate and the 2 nonsignificant pathways were removed in the final model.

    The final path model showed good model fit (χ24= 6.1, P=.19, CFI=0.99, TLI=0.97, RMSEA=0.04, and WRMR=0.55). Standardized regression coefficients for the final model are reported in Table 3 and Figure 3. Results indicated that the intervention reduced participants’ perceived stress at the 3-month follow-up (beta=−.32, P<.001), which was positively associated with depressive symptoms at the 3-month follow-up (beta=.83, P<.001); reduced depressive symptoms consequently resulted in reduced suicide at 6 months (beta=.34, P<.001). The pathway from perceived stress to depressive symptoms had the strongest effect size in the mediation model (beta=.83, P<.001).

    The direct, indirect, and total effects of the path model are summarized in Table 3. The direct effect of the intervention on suicide at the 6-month follow-up was significant (beta=−.18, P=.02). The indirect effect of the intervention on suicide via perceived stress and depressive symptoms was also significant (beta=−.32*.83*.34=−.09; P<.001). In summary, there was a partial mediating effect of perceived stress and depressive symptoms, accounting for 33% (−0.09/−0.27) of the total effect of the intervention on suicide.

    Table 3. Coefficients of the pathways in the final model (n=284).
    PathwaysCoefficient (beta)Standardized coefficient (beta)95% CISEP value
    Intervention→ perceived stressa−3.67−.32−5.03 to −2.310.69<.001
    Intervention→ suicideb−.41−.18−0.76 to −0.070.18.02
    Perceived stressa → depressive symptomsa1.44.831.31 to 1.570.07<.001
    Depressive symptomsa → suicideb.04.340.02 to 0.060.01<.001
    Total effect−.62−.27−0.96 to −0.280.17<.001
    Direct effect

    		Intervention→ suicideb−.41−.18−0.76 to −0.070.18.02
    Indirect effect

    		Intervention→ perceived stressa→ depressive symptomsa → suicideb−.21−.09−0.33 to −0.090.06.001

    a3-month follow-up.

    b6-month follow-up.

    ‎
    Figure 3. Estimation of the final path model of intervention, perceived stress, depressive symptoms, and suicide in people living with HIV and depression.
    View this figure

    Principal Findings

    This study was among the first efforts to explore the mechanisms of how an mHealth intervention reduced suicide in people living with HIV and depression, with a longitudinal design. Understanding the mechanisms is important to discern the processes between an intervention and mental health outcomes [40,41]. Better understanding of such mechanisms can facilitate targeted mHealth intervention to prevent suicide among people living with HIV and depression. Our data analyses revealed that the mHealth intervention had both direct and indirect effects on suicide, suggesting that the mHealth intervention had effectively reduced the participants’ suicide. These findings demonstrated that suicide in people living with HIV and depression could be reduced through a WeChat-based CBSM intervention, such as Run4Love. The mechanisms of the indirect effects were mediated through reduced perceived stress and depressive symptoms, which consequently resulted in reduced suicide.

    Implications of the Path Model

    Consistent with the existing literature [15-18], perceived stress had a strong positive association with depressive symptoms, as evidenced by the large effect size shown in this study. This finding underscored the importance of reducing perceived stress to alleviate depressive symptoms in people living with HIV and depression. Many effective mHealth interventions that targeted depressive symptoms in people living with HIV have included stress management as a main component of the intervention, such as Web-based cognitive behavioral therapy [42].

    The results also indicated the causal relationship between depressive symptoms at the 3-month follow-up and suicide at the 6-month follow-up. Consistent with previous studies conducted in the general population or psychiatric outpatients [5,6], the findings demonstrated that suicide might be effectively reduced through mitigating depressive symptoms. Furthermore, this study confirmed this relationship in the vulnerable population of people living with HIV and depression, who needed targeted intervention because of high rates of suicide. Some studies screened people living with HIV for depression before targeted intervention, and these demonstrated good efficacy [43-45]. Our data also suggested the importance of targeted suicide prevention intervention in people living with HIV and depression.

    The direct effect of the mHealth intervention on suicide might also be explained by factors other than perceived stress and depressive symptoms, such as reduction of stigma and improvement of self-efficacy, positive coping, and social support. Further research is needed to better understand the roles and mechanisms of these potential factors on suicide.

    Another finding of this study that needs to be noted is that people living with HIV and depression who committed suicide reported lower level of employment. Previous studies found that unemployment was a risk factor for suicide [46]. As a stressful event, unemployment could lead to financial distress and social isolation, which was associated with elevated stress and depression and, eventually, suicide [47,48]. Although the relationship was significant in bivariate analysis, it was no longer significant when other factors (perceived stress, depressive symptoms, and suicide at baseline) were controlled.

    The data showed that the Run4Love mHealth intervention alone was effective, compared with the blank control group. Currently, the professional psychologists cannot meet the treatment demands of large number of people living with HIV and depression, especially in resources-limited settings, such as China. Therefore, an mHealth intervention can be an alternative resolution for people living with HIV and depression who cannot access psychotherapy. Alternatively, such an mHealth intervention can also be integrated with other traditional psychotherapies because of its easy access and adaptability. Future studies are needed to examine the effect of a stand-alone mHealth intervention compared with the one adjunct to psychotherapy.

    This study suggested that public policies and coordinated efforts should be made to improve mental health outcomes, especially to reduce suicide among people living with HIV and depression. At the health care level, depression treatment and suicide prevention for people living with HIV are urgently needed. Although the guideline of HIV treatment has recommended integrating mental health screening into regular HIV care [49], such a practice was rarely implemented because of limited psychological recourses, perceived stigma of people living with HIV, and extra commute burden of face-to-face psychotherapy [42]. In middle- and low-income countries, such as China, mHealth interventions such as the WeChat-based Run4Love can reach more people in a cost-effective manner [13]. Run4Love platform also provides useful functions, such as automatic progress tracking and personalized feedback, which are necessary for an effective intervention. Our data underscore the importance of integrating mental health services into routine HIV health care by mobile apps, such as WeChat. At the community level, we call for more mHealth interventions, such as Run4Love, to reduce stress and improve mental health outcomes of people living with HIV and depression.

    Limitations

    Several limitations of this study should be noted. First, the participants were recruited from 1 hospital in a metropolitan area, and the majority of them were men. Therefore, the results might not be generalizable to all people living with HIV and depression in China, especially women or those in rural areas. Second, the measures in this study were self-reported, with potential recall or self-report biases. We did not include a standardized measurement of suicide, and we only had 2 items about suicidal ideation and suicidal attempt. Nevertheless, self-reported suicide assessment has shown a high level of agreement with a clinician-delivered face-to-face assessment, and this could serve as an efficient and reliable method to assess suicide outcome [50]. Third, like all modeling studies, assumptions were embedded in the current path model, for example, linear relationship among variables. Finally, the pathway model included only a limited number of covariates, and other potential factors that might affect mental health outcomes were not included.

    Conclusions

    In conclusion, this was the first study to examine the mechanisms of an mHealth intervention, Run4Love, in reducing suicide through both direct and indirect pathways, using longitudinal data from an RCT. Besides direct effect, the indirect effect was mediated by reduced perceived stress and depressive symptoms, which consequently resulted in reduced suicide. To reduce suicide in people living with HIV and depression, perceived stress and depressive symptoms are both key intervention targets. We call for targeted intervention to prevent suicide in people living with HIV and depression, especially mHealth programs that can reduce perceived stress.

    Acknowledgments

    This study was supported by a grant from the National Natural Science Foundation of China (Grant Number: 71573290) and China Medical Board (Grant Number: 17-271). The funder had no role in study design, data collection and analysis, or writing of the manuscript.

    Authors' Contributions

    YL analyzed the data and drafted the paper. YG and YAH contributed to funding obtaining, study design, and manuscript revision. MZ and CZ helped in study concept. JQ, ZX, HZ, and YZ contributed to clinical trial and data acquisition. WC, LL, and CL provided administrative, technical, and material support of the clinical trial.

    Conflicts of Interest

    None declared.

    ‎
    Multimedia Appendix 1

    CONSORT-EHEALTH checklist (V 1.6.1).

    PDF File (Adobe PDF File), 2277 KB
    1. Bing EG, Burnam MA, Longshore D, Fleishman JA, Sherbourne CD, London AS, et al. Psychiatric disorders and drug use among human immunodeficiency virus-infected adults in the United States. Arch Gen Psychiatry 2001 Aug;58(8):721-728. [CrossRef] [Medline]
    2. Wang T, Fu H, Kaminga AC, Li Z, Guo G, Chen L, et al. Prevalence of depression or depressive symptoms among people living with HIV/AIDS in China: a systematic review and meta-analysis. BMC Psychiatry 2018 May 31;18(1):160 [FREE Full text] [CrossRef] [Medline]
    3. UNAIDS. 2018. Global HIV & AIDS Statistics—2019 Fact Sheet.   URL: http://www.unaids.org/en/resources/fact-sheet [accessed 2019-09-13]
    4. Bengtson AM, Pence BW, Mimiaga MJ, Gaynes BN, Moore R, Christopoulos K, et al. Depressive symptoms and engagement in human immunodeficiency virus care following antiretroviral therapy initiation. Clin Infect Dis 2019 Jan 18;68(3):475-481 [FREE Full text] [CrossRef] [Medline]
    5. Brown GK, Beck AT, Steer RA, Grisham JR. Risk factors for suicide in psychiatric outpatients: a 20-year prospective study. J Consult Clin Psychol 2000 Jun;68(3):371-377. [CrossRef] [Medline]
    6. Zhang Y, Law CK, Yip PS. Psychological factors associated with the incidence and persistence of suicidal ideation. J Affect Disord 2011 Oct;133(3):584-590. [CrossRef] [Medline]
    7. Badiee J, Moore DJ, Atkinson JH, Vaida F, Gerard M, Duarte NA, et al. Lifetime suicidal ideation and attempt are common among HIV+ individuals. J Affect Disord 2012 Feb;136(3):993-999 [FREE Full text] [CrossRef] [Medline]
    8. Zeng C, Li L, Hong YA, Zhang H, Babbitt AW, Liu C, et al. A structural equation model of perceived and internalized stigma, depression, and suicidal status among people living with HIV/AIDS. BMC Public Health 2018 Jan 15;18(1):138 [FREE Full text] [CrossRef] [Medline]
    9. Cooper V, Clatworthy J, Whetham J, Consortium E. mHealth interventions to support self-management in HIV: a systematic review. Open AIDS J 2017;11:119-132 [FREE Full text] [CrossRef] [Medline]
    10. Cao B, Gupta S, Wang J, Hightow-Weidman LB, Muessig KE, Tang W, et al. Social media interventions to promote HIV testing, linkage, adherence, and retention: systematic review and meta-analysis. J Med Internet Res 2017 Nov 24;19(11):e394 [FREE Full text] [CrossRef] [Medline]
    11. Kempf M, Huang C, Savage R, Safren SA. Technology-delivered mental health interventions for people living with HIV/AIDS(PLWHA): a review of recent advances. Curr HIV/AIDS Rep 2015 Dec;12(4):472-480 [FREE Full text] [CrossRef] [Medline]
    12. Swendeman D, Ramanathan N, Baetscher L, Medich M, Scheffler A, Comulada WS, et al. Smartphone self-monitoring to support self-management among people living with HIV: perceived benefits and theory of change from a mixed-methods randomized pilot study. J Acquir Immune Defic Syndr 2015 May 1;69 Suppl 1:S80-S91 [FREE Full text] [CrossRef] [Medline]
    13. van Luenen S, Garnefski N, Spinhoven P, Kraaij V. Guided internet-based intervention for people with HIV and depressive symptoms: a randomised controlled trial in the Netherlands. Lancet HIV 2018 Sep;5(9):e488-e497. [CrossRef] [Medline]
    14. Guo Y, Hong YA, Qiao J, Xu Z, Zhang H, Zeng C, et al. Run4Love, a mHealth (WeChat-based) intervention to improve mental health of people living with HIV: a randomized controlled trial protocol. BMC Public Health 2018 Jun 26;18(1):793 [FREE Full text] [CrossRef] [Medline]
    15. Metalsky GI, Joiner TE. Vulnerability to depressive symptomatology: a prospective test of the diathesis-stress and causal mediation components of the hopelessness theory of depression. J Pers Soc Psychol 1992 Oct;63(4):667-675. [CrossRef] [Medline]
    16. Brown TA, Rosellini AJ. The direct and interactive effects of neuroticism and life stress on the severity and longitudinal course of depressive symptoms. J Abnorm Psychol 2011 Nov;120(4):844-856 [FREE Full text] [CrossRef] [Medline]
    17. Felsten G. Stress reactivity and vulnerability to depressed mood in college students. Pers Indiv Differ 2004;36(4):789-800. [CrossRef]
    18. Caspi A, Sugden K, Moffitt TE, Taylor A, Craig IW, Harrington H, et al. Influence of life stress on depression: moderation by a polymorphism in the 5-HTT gene. Science 2003 Jul 18;301(5631):386-389 [FREE Full text] [CrossRef] [Medline]
    19. Smith SS, Smith Carter J, Karczewski S, Pivarunas B, Suffoletto S, Munin A. Mediating effects of stress, weight-related issues, and depression on suicidality in college students. J Am Coll Health 2015;63(1):1-12. [CrossRef] [Medline]
    20. Anastasiades MH, Kapoor S, Wootten J, Lamis DA. Perceived stress, depressive symptoms, and suicidal ideation in undergraduate women with varying levels of mindfulness. Arch Womens Ment Health 2017 Feb;20(1):129-138. [CrossRef] [Medline]
    21. Gill SK, Muñoz RF, Leykin Y. The influence of perceived stress and depression on suicide-related beliefs in caucasian and Indian adults. Crisis 2018 Mar;39(2):127-136. [CrossRef] [Medline]
    22. Guangdong Provincial Center for Disease Control and Prevention. Guangdong Provincial Center for Disease Control and Prevention. 2017. AIDS Prevention and Treatment in 2017 of Guangdong Province (in China).   URL: http://cdcp.gd.gov.cn/hhff/inbfo/content/post_1107865.html [accessed 2019-09-13] [WebCite Cache]
    23. Antoni M, Ironson G, Schneiderman N. Cognitive-behavioral Stress Management For Individuals Living With HIV (Treatments That Work). Oxford: Oxford University Press; 2007.
    24. Cruess DG, Antoni MH, Schneiderman N, Ironson G, McCabe P, Fernandez JB, et al. Cognitive-behavioral stress management increases free testosterone and decreases psychological distress in HIV-seropositive men. Health Psychol 2000 Jan;19(1):12-20. [CrossRef] [Medline]
    25. McGregor BA, Dolan ED, Murphy KM, Sannes TS, Highland KB, Albano DL, et al. Cognitive behavioral stress management for healthy women at risk for breast cancer: a novel application of a proven intervention. Ann Behav Med 2015 Dec;49(6):873-884 [FREE Full text] [CrossRef] [Medline]
    26. Gaab J, Blättler N, Menzi T, Pabst B, Stoyer S, Ehlert U. Randomized controlled evaluation of the effects of cognitive-behavioral stress management on cortisol responses to acute stress in healthy subjects. Psychoneuroendocrinology 2003 Aug;28(6):767-779. [CrossRef] [Medline]
    27. Tencent. 2018. Tencent Announces 2018 Second Quarter and Interim Results.   URL: https://www.tencent.com/en-us/articles/8003521534381984.pdf [accessed 2019-05-07] [WebCite Cache]
    28. Radloff LS. The CES-D Scale:A self-report depression scale for research in the general population. ‎Appl Psychol Meas 2016 Jul 26;1(3):385-401. [CrossRef]
    29. Wang M, Armour C, Wu Y, Ren F, Zhu X, Yao S. Factor structure of the CES-D and measurement invariance across gender in mainland Chinese adolescents. J Clin Psychol 2013 Sep;69(9):966-979. [CrossRef] [Medline]
    30. Zhang Y, Ting RZ, Lam MH, Lam S, Yeung RO, Nan H, et al. Measuring depression with CES-D in Chinese patients with type 2 diabetes: the validity and its comparison to PHQ-9. BMC Psychiatry 2015 Aug 18;15:198 [FREE Full text] [CrossRef] [Medline]
    31. Zich JM, Attkisson CC, Greenfield TK. Screening for depression in primary care clinics: the CES-D and the BDI. Int J Psychiatry Med 1990;20(3):259-277. [CrossRef] [Medline]
    32. Lu W, Bian Q, Wang W, Wu X, Wang Z, Zhao M. Chinese version of the Perceived Stress Scale-10: A psychometric study in Chinese university students. PLoS One 2017;12(12):e0189543 [FREE Full text] [CrossRef] [Medline]
    33. Leung DY, Lam T, Chan SS. Three versions of Perceived Stress Scale: validation in a sample of Chinese cardiac patients who smoke. BMC Public Health 2010 Aug 25;10:513 [FREE Full text] [CrossRef] [Medline]
    34. Wang Z, Chen J, Boyd JE, Zhang H, Jia X, Qiu J, et al. Psychometric properties of the Chinese version of the Perceived Stress Scale in policewomen. PLoS One 2011;6(12):e28610 [FREE Full text] [CrossRef] [Medline]
    35. Ng SM. Validation of the 10-item Chinese perceived stress scale in elderly service workers: one-factor versus two-factor structure. BMC Psychol 2013;1(1):9 [FREE Full text] [CrossRef] [Medline]
    36. Olpin M, Hesson M. Stress Management For Life: A Research-based Experiential Approach. Boston: Cengage Learning; 2015.
    37. Hu L, Bentler PM. Fit indices in covariance structure modeling: sensitivity to underparameterized model misspecification. ‎Psychol Methods 1998;3(4):424-453. [CrossRef]
    38. Hu L, Bentler PM. Cutoff criteria for fit indexes in covariance structure analysis: conventional criteria versus new alternatives. Struct Equ Modeling 1999 Jan;6(1):1-55. [CrossRef]
    39. Kline RB. Latent variable path analysis in clinical research: a beginner's tour guide. J Clin Psychol 1991 Jul;47(4):471-484. [CrossRef] [Medline]
    40. Silverstein M, Cabral H, Hegel M, Diaz-Linhart Y, Beardslee W, Kistin CJ, et al. Problem-solving education to prevent depression among low-income mothers: a path mediation analysis in a randomized clinical trial. JAMA Netw Open 2018 Jun 1;1(2):e180334 [FREE Full text] [CrossRef] [Medline]
    41. Carrico AW, Ironson G, Antoni MH, Lechner SC, Durán RE, Kumar M, et al. A path model of the effects of spirituality on depressive symptoms and 24-h urinary-free cortisol in HIV-positive persons. J Psychosom Res 2006 Jul;61(1):51-58. [CrossRef] [Medline]
    42. Karyotaki E. Internet-based interventions for people with HIV and depression. Lancet HIV 2018 Sep;5(9):e474-e475. [CrossRef] [Medline]
    43. Kelly JA, Murphy DA, Bahr GR, Kalichman SC, Morgan MG, Stevenson LY, et al. Outcome of cognitive-behavioral and support group brief therapies for depressed, HIV-infected persons. Am J Psychiatry 1993 Nov;150(11):1679-1686. [CrossRef] [Medline]
    44. Lee MR, Cohen L, Hadley SW, Goodwin FK. Cognitive-behavioral group therapy with medication for depressed gay men with AIDS or symptomatic HIV infection. Psychiatr Serv 1999 Jul;50(7):948-952. [CrossRef] [Medline]
    45. Kraaij V, van Emmerik A, Garnefski N, Schroevers MJ, Lo-Fo-Wong D, van Empelen P, et al. Effects of a cognitive behavioral self-help program and a computerized structured writing intervention on depressed mood for HIV-infected people: a pilot randomized controlled trial. Patient Educ Couns 2010 Aug;80(2):200-204. [CrossRef] [Medline]
    46. Milner A, Page A, LaMontagne AD. Cause and effect in studies on unemployment, mental health and suicide: a meta-analytic and conceptual review. Psychol Med 2014 Apr;44(5):909-917. [CrossRef] [Medline]
    47. Yur'yev A, Värnik A, Värnik P, Sisask M, Leppik L. Employment status influences suicide mortality in Europe. Int J Soc Psychiatry 2012 Jan;58(1):62-68. [CrossRef] [Medline]
    48. Schneider B, Grebner K, Schnabel A, Hampel H, Georgi K, Seidler A. Impact of employment status and work-related factors on risk of completed suicide. A case-control psychological autopsy study. Psychiatry Res 2011 Dec 30;190(2-3):265-270. [CrossRef] [Medline]
    49. World Health Organization. Consolidated Guidelines On The Use Of Antiretroviral Drugs For Treating And Preventing HIV Infection: Recommendations For A Public Health Approach. Geneva: WHO press; 2016.
    50. Safer DJ. Self-reported suicide attempts by adolescents. Ann Clin Psychiatry 1997 Dec;9(4):263-269. [Medline]

    ‎
    CBSM: cognitive-behavioral stress management
    CES-D: Center for Epidemiologic Studies-Depression
    CFI: Comparative Fit Index
    IQR: interquartile range
    mHealth: mobile health
    PSS: Perceived Stress Scale
    RCT: randomized controlled trial
    RMSEA: root mean square error of approximation
    TLI: Tucker-Lewis Index
    WRMR: weighted root mean square residual

    Edited by G Eysenbach; submitted 18.05.19; peer-reviewed by R Melia, L Elopre; comments to author 03.08.19; revised version received 25.08.19; accepted 31.08.19; published 27.11.19

    Copyright

    ©Yiran Li, Yan Guo, Y Alicia Hong, Mengting Zhu, Chengbo Zeng, Jiaying Qiao, Zhimeng Xu, Hanxi Zhang, Yu Zeng, Weiping Cai, Linghua Li, Cong Liu. Originally published in the Journal of Medical Internet Research (http://www.jmir.org), 27.11.2019.

    This is an open-access article distributed under the terms of the Creative Commons Attribution License (https://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work, first published in the Journal of Medical Internet Research, is properly cited. The complete bibliographic information, a link to the original publication on http://www.jmir.org/, as well as this copyright and license information must be included.