Association between Community Collectivization and Mental Depression among Men Who Have Sex with Men in Andhra Pradesh, India

Mental health problems are among the most significant contributors to the global burden of disease and disability with a global prevalence of 13% [1]. According to the WHO report, individual-level attributes and behaviors (low self-esteem, emotional immaturity, medical illness and substance use), environmental factors (poor access to basic services, discrimination, social and gender inequalities) and social and economic circumstances (loneliness, family conflict/neglect, exposure to violence/abuse, low income and poverty) are contributing factors to mental health and wellbeing [2]. In addition, the mental health issues, especially among the marginalized population of female sex workers (FSWs), Men Who have Sex with Men (MSM) and injecting drug users (IDUs) and people living with HIV (PLHIV) is a serious concern for the effective HIV prevention efforts. The studies among FSWs revealed that high prevalence of violence, drug use and discrimination and HIV risk affect the mental health of this population [3-5]. However, the research among people living with HIV (PLHIV) in low and middle income countries signify the prevalence of mental health problems like depression ranging from 11% to 63% and depression is linked to poor adherence to antiretroviral therapy [6-8]. Globally, as compared to the general population, a community based sample of MSM is at increased risk of experiencing mental health problems. The factors contributing to poor mental health of MSM are depressive symptoms, substance use disorders, mental disorders-anxiety, panic, mood, suicidal ideation, discrimination, suicidal symptoms among individuals reporting same sex behavior or identifying as a homosexual, partner violence and childhood sexual abuse. These hazardous psychosocial health problems had often been considered as a source of explaining the distribution and consequences of HIV infection among MSM [9-18]. Abstract

In India, a range of 10-12% of lifetime prevalence of mental health disorders is observed and of which MSM faces many psychosocial challenges due to their stigmatized sexual orientation and are highly vulnerable to a variety of societal stressors resulting in poor mental health [19]. Despite the extensive HIV prevention efforts, the MSM continues to constitute a high-risk group (HRG) with national HIV prevalence of 4.4% and 10.1% in Andhra Pradesh [20]. Factors such as stigma, discrimination, and violence lead to several co-occurring mental health problems among MSM, which in turn restrict their ability to benefit from risk reduction counselling and moderate the beneficial impact of HIV prevention approaches/efforts [21]. Recent studies have shown the interrelationship between psychosocial factors and elevated HIV risk among MSM [21,22]. In order to address these elevated levels of HIV risk among MSM, a community-led structural intervention (CLSI) plays an architect role to address stigma, violence and harassments of MSM in public spaces and helps to create an enabling environment for them. The CLSI stands on three major pillars: creating an enabling environment, control over access and utilization of services and community mobilization. Of which, community mobilization plays a prime role in the success of any development initiative [23]. In India, as a part of CLSI, community collectivization processes through peerled approach was intervened to position the community as a leading actor in order to tackle the structural issues to reduce individual and communities' vulnerability to HIV among them through the Avahan program, its India AIDS initiative [24,25]. Avahan worked with highrisk communities to strengthen their individual and collective agency so that they can adopt and sustain the practice of safer behaviors [26]. As a part of the Avahan intervention among MSM a 'crisis response strategy' was devised to deal with violence, discrimination, abuse and harassment (which leads to co-occurring mental health problems among MSM) from the family members, clients, police and members of the community and to assist the victim and enhance the confidence and sense of safety among the MSM community [25].
Globally, studies among MSM and HIV positive gay couples had demonstrated the impact of informal social support from the sex partner or community member to enhance the mental health status to build the emotional well-being of the individual and support from family and friends to reduce the risky HIV behaviors [27][28][29]. This signifies the role of social support and involvement of community members in the improvization of individuals' health. In addition, studies on community led structural intervention among marginalized groups in India had demonstrated the positive outcome on reducing not only the HIV/STI risk behaviors [30][31][32], but also the structural factors [33][34][35][36][37][38][39]. Earlier studies had focused mainly on the correlates of depression among key populations [40][41][42]; however the role of community collectivization on the mental health of key populations is still understudied in India. This creates an opportunity to study the association between community led structural interventions and mental depression among key populations. With this background, the objective of this paper is to study the prevalence of mental depression among MSM. Secondly, to examine the association between community collectivization indicators and mental depression among MSM. Thirdly, to examine the combine association of community collectivization indicator (e.g. collective agency) and violence with mental depression among MSM.

Study settings, study design and sampling
This study used the data from the Behavioral Tracking Survey (BTS), a cross-sectional, survey conducted between January and February, 2012 among MSM in Andhra Pradesh (the state was divided into two separate states as Andhra Pradesh and Telangana from June 2, 2014). The objective of the survey was to track the key components of Avahan program, its India AIDS initiative, such as community mobilization, safer sexual practices and STI treatment seeking behaviors. The survey tool covered the information on condom use, STI treatment, community mobilization and also mental health using the Patient Health Questionnaire -2 (PHQ-2) depression scales.
Three districts namely Khammam (n=388), Kurnool (n=396) and Ananthapur (n=392) were covered in this BTS survey from 23 districts where Avahan program was implemented. A sample size of 400 was calculated for each district based on prevalence of consistent condom use (CCU) and expected level of change. A rapid mapping exercise was done at each hot spot (the place where the MSM congregate to solicit clients) through key informants such as local community members, police and social workers to validate a list of hot spots. The hot spots were then divided into two categories -non-public (hotels, lodges, roadside cafes) and public (streets, market areas, cinemas). The sampling frame was developed and probability-based sampling was used to select the respondents. A conventional cluster sampling was used for non-public hot spots and time-location cluster sampling was used for public hot spots. In time-location cluster sampling hot spots were divided into clusters according to time slots at different locations. As a first stage of time-location cluster sampling, the required number of clusters (time-location cluster of hot spots) were randomly selected and at second stage, the respondents within each selected cluster were randomly selected to reach the sample size of 400 per each district. A total sample of 1176 MSM was collected in this survey. As a selection criteria "males aged 18 or older identified to have sex with another male in exchange of cash/kind and cruising from one place to another, soliciting clients, or hanging out at any suitable place, including street corners, highways and pick up points within operational area" were eligible to participate in the survey. All interviews were conducted by the trained field staff and were held in private locations convenient to study participants.

Avahan program
In 2003, Bill and Melinda Gates Foundation began its India AIDS Initiative, known as Avahan, to curb the spread of HIV in India. The first phase of the Avahan program (2003-2007) was perceived as a focused prevention program, offering a standardized package of proven prevention interventions for high-risk groups (including MSM) and bridge populations in the geographic areas most affected by the epidemic. Avahan program was intervening in six high-prevalence states in southern (Andhra Pradesh, Karnataka and Tamil Nadu), Western (Maharashtra) and North-east (Manipur and Nagaland) regions of India. The Avahan package of prevention intervention was based on five program components -peer-led outreach -to provide support and information to improve their ability to negotiate condom use and to encourage for attendance at the STI clinic; program-supported clinical services -to treat STIs other than HIV; commodity distribution; community mobilization and ownership of the program and advocacy for an enabling environment. The second phase of the Avahan program (2007-2012) was designed with a special focus on community mobilization, safe sex behavior, STI treatment-seeking behavior, financial security, and social entitlements among key populations in India.

Ethics statement
The study design and questionnaires were approved by the institutional review boards of Family Health International and the Karnataka Health Promotion Trust. Verbal consent was obtained from all respondents prior to participation in the interview. No names and addresses of participants were recorded on the questionnaires. This survey has tried to take care of all the ethics measures of the study and study participants. This study protects the rights of privacy, confidentiality, and well-being of FSWs. Participants were not provided any compensation for their time in the study, but were referred to local project services run by implementing agency in the survey districts.

Measures Background characteristics
The survey included the socio-demographic measures such as age, educational attainment (illiterate, primary, secondary and secondary+); marital status (never married, currently married and formerly married); living status (living alone, living with spouse/female partner, living with male partner/friend/colleague and living with family/others); current place of residence (urban, semi-urban and rural); duration of practicing anal sex work with other men (<=5 years, 6-10 years and 10+ years); and mobile for sex work in past 2 years (non-mobile, less mobile and more mobile).

Dependent variable
The key outcome variable was mental health status and measured based on Patient Health Questionnaire (PHQ-2) depression scale.
Respondents were asked two questions-"Over the last 2 weeks, how often have you been bothered by any of the problems-1) Little or no interest or pleasure in doing things 2) feeling down, depressed or hopeless" with responses as not at all, several days, more than half days and nearly every day. A composite index was constructed with scale reliability (Cronbach's alpha) of 0.84. The index score was ranging from 0-6 and divided into two categories as 0 'No depression' for score 0 and for score 1 to 6 as 1 ' Any depression' .

Independent variables
The survey also collected the information on enabling environment indicators such as experienced physical or sexual violence in the past 6 months (no, yes). The other independent variables related to individuallevel and collective-level community led structural intervention (CLSI) indicators included in the analysis are defined and described as below.

Individual-level CLSI variables
Individual agency: This individual level indicator was measured using the item: Did you ever refuse a client in the last one month, if he insisted on sex without a condom? The responses were yes and no. The response 'yes' indicates the individual control, whereas response 'no' indicates no individual control.

Collective-level CLSI variables
Collective identity: Collective identity was measured using the statements: a) Individuals need not be ashamed to be MSM? b) Other MSMs have thought the same things as me? The responses were strongly agreed, agree, disagree, and strongly disagree. A composite index was constructed with scale reliability (Cronbach's alpha) of 0.60. The index score was further divided into two equal categories (using median) of 'low' with coding value '0'for the score between 0-1 and 'high' with coding value '1' for the score 2 or more.
Collective efficacy: Collective efficacy was measured using the responses to the item: How confident are you that FSWs in your community can work together to achieve the following goals: keep each other safe from harm; increase condom use with clients; speak up for your rights; and improve your lives? The responses were not at all, somewhat, very, and completely confident. A composite index was constructed with scale reliability (Cronbach's alpha) of 0.95. The index score was further divided into two equal categories (using median) as 0 'low' with the score 0-5 and 1 as 'high' with score 6 or more.
Collective action: Collective action was measured using the items: a) The last time you were arrested, did the other MSM help you? b) The last time a client or partner was violent; did the other MSM help you? The responses were dichotomous as no and yes. A composite index was constructed with scale reliability (Cronbach's alpha) of 0.65. The index score was further divided into two equal categories (using median) as 0 'low' with score 0 and 1 as 'high' with score 1 or more.
Collective agency: Collective agency was measured using the item: Are you a member of a community group? (Self-help group or community based organization) with the response as (0=No and 1=Yes).

Statistical analysis
Descriptive statistics were carried out to calculate proportions, means and standard deviations to describe the profile of MSM. Bivariate analysis  was performed to assess the association between key outcome variable, i.e. mental health status of the MSM and independent variables using the chi-square test statistics. Multivariate logistic regression models were used to estimate the adjusted odds ratios (AORs) with 95% confidence interval to examine the factors associated with mental health (any depression) among MSM in Andhra Pradesh. Multiple logistic regression models were used to assess the relationship between community collectivization indicators with mental health by controlling the socio-demographic characteristics such as age, current marital status, living status, current place of residence, duration of practicing anal sex work with other men and mobility for sex work. The p-value was set at 5% level of significance and all the analyses were conducted using STATA 11.2.

Profile and CLSI characteristics of sampled population
The average age of the MSM was 28.2 years (± 6.2 years) with the highest proportion of attaining the secondary school (38%). More than three-fifths (61%) of MSM were never married. Sixty four percent of MSM were living with family or others, and more than two-fifths (41%) were currently residing in urban places. More than forty-five percent of MSM reported to be engaged in anal sex for more than 10 years of duration and almost half of the MSM (48%) were non-mobile (Table 1).
As a part of CLSI strategy, enabling environment indicators described that more than half (57%) of MSM reported to be under debt at the time of the survey and almost one-tenth of MSM (9%) had experienced either physical or sexual violence in the past 6 months. Among the MSM, about 72% have individual agency (e.g. 'high' level of community involvement). At collective-level, more than half of the MSM were at a high level of community involvement for CLSI indicators such as collective identity (77%), collective efficacy (54%), collective action (48%) and collective agency (92%). More than onethird of MSM reported of having any depression (35%) in Andhra Pradesh (Tables 2 and 3).  at collective-level, collective identity (33% vs. 41%, AOR: 0.54, p=0.01) and collective agency (34% vs. 38%, AOR: 0.46, p=0.01) were inversely associated with mental depression. In addition, the composite variable of collective agency and experience of violence (31% vs. 37%, AOR: 0.44, p=0.012) was also significantly associated with mental health status, i.e., those were members of the community group and had not experienced any violence, were less likely to have any depression.

Discussion
This study among MSM in Andhra Pradesh demonstrates the evidence of association between individual-level and collective-level collectivization of community on mental health in the context of community led structural approaches. The findings from the study show that the community collectivization factors show a negative relationship with the mental depression of MSM. In addition, the study shows an inverse relationship between individual agency and mental depression, however it is not statistically significant. The experience of physical or sexual violence among MSM were positively associated with poor mental health which plays a 'dual' role at an individual level as well as the dimension of enabling environment among the community.
Previous studies show the experience of physical or sexual violence among same-sex relationship is higher than opposite-sex relationship [43,44]. A systematic review and meta-analysis among MSM, by Buller et al. illustrate a strong association between partner violence and depressive symptoms which in turn leads to poor mental health. The study further concludes that men who are depressed and are part of violent relationship finds it difficult to seek help or to adhere the treatment [15]. The previous studies in different types of populations suggest that physical or sexual violence are associated with poor mental health at an individual level [45][46][47].
This study contends the poor enabling environment among MSM. In addition, not having an individual agency (refused the client when insisted on sex without a condom) also shows the poor mental health among MSM. Though, the statistical significance is not established in this relationship, it throws the light on the poor condom negotiation power among the MSM. The reason for such low condom negotiation power is the intimate partner violence which affects the condom negotiating skills among them [48] and leads to poor mental health due to violence. This implies the importance of individual level structural indicators in enhancement of mental health among the most-at-risk populations. This is supported by the Coates et al. in their review of individual behavior change strategies and demonstrated that when HIV prevention strategies address the broader structural factors that constrain individual behavior (such as power, policy, and poverty) the propensity of success in reducing risk behavior substantially improved [49]. Thus, a positive enabling environment among the MSM community is one of the most important structural factors to keep themselves mentally healthy to practice safer sexual behaviors. Since the last few decades, both individual-level structural interventions and community-led structural interventions are effective in improving the outcomes of HIV interventions by altering the social, economic, political or environmental factors. These altering factors further accentuate the HIV risk and vulnerability. In this context, community collectivization is central and key to success of HIV prevention program worldwide. It is also evident from the previous studies that structural approaches and community mobilization are an integral part of successful interventions in reducing the HIV risk vulnerability among key populations (e.g. female sex workers, men having sex with men and injecting drug users) [30,33,36,37,[50][51][52][53]. When mental health in the context of community collectivization approach among MSM is discussed, there is negligible evidence in this sphere of mental health. This study demonstrated a positive association of community collectivization with enhanced mental health status (i.e., less mental depression) among MSM. In other words, being part of a community group (i.e., collective agency) has a positive association with good mental health among MSM. Similarly, the findings of the study also show that high degree of collective identity is negatively associated with mental depression among MSM. These crucial findings of this study provide an evidence of the importance of community ownership and community collectivization. Since community collectivization is an integral component of any public health intervention and has shown a positive effect on a variety of sexual, reproductive and child health problems. Moreover, this collectivization is characterized by a strong sense of collective identity and collective agency; which would further help to MSM community to work together to achieve their goals and stand up for each other. This study is unique in establishing the relationship between community collectivization and mental depression among MSM in India and recommends for further research on this. Further, the composite measure of collective agency and violence is also significantly associated with mental depression among MSM. It  concludes that MSM, those were members of community group and not experienced any physical or sexual violence were less likely to have mental depression; which provides an evidence of collective agency as a mediator for good mental health among MSM.
Although the findings of this study have important programmatic implications, they must be interpreted with some certain limitations. First, the design of the study is cross-sectional and hence causality cannot be established in this study. Secondly, the key independent variables were based on the self-reported answers which are exposed to social desirability and recall bias. Thirdly, the creation of the mental health status variable is based on 2-items from PHQ scale; hence need careful consideration while interpreting the results. Lastly, the study settings for this analysis are Andhra Pradesh and not the entire community of MSM in India and there is a state to state variation due to socio-demographics, regional, and cultural difference among MSM. Hence, results must be interpreted carefully and cannot be generalized to the entire MSM population in India. Despite these limitations, the present study is the first to examine the role of community-led structural intervention or participation in mental health. The present study shows that individual-level variables related to enabling environment and community collectivization indicators significantly associated with mental depression among MSM. It highlights that community-led structural interventions are more successful and effective in changing the attitudes, behavior and knowledge of the key population in terms of HIV-risk reduction along with creating a positive environment for mental health among MSM. This study demands for more community engagement activities and community participation in order to deal with mental health problems among the key population. This study recommends for further research and to explore the new community led structural approaches with innovative ideas for integrated mental health counselling services among MSM.