Experiences of conflict, non-acceptance and discrimination are associated with poor mental well-being amongst LGBTQ-identified individuals in Singapore

Abstract

Purpose

Health disparities affecting lesbian, gay, bisexual, transgender and queer (LGBTQ) populations have been reported in many countries. For Singapore, no large quantitative studies on mental health and well-being in the local LGBTQ community have been published. The authors conducted a community-based survey (National LGBT Census Singapore, 2013; NLCS2013) that covered a comprehensive set of demographic, social and health indicators. Here, the authors investigated mental health status and its correlates in 2,350 LGBTQ individuals within the NLCS2013 sample.

Design/methodology/approach

The NLCS2013 was an anonymous online survey conducted amongst self-identified LGBTQ adults (aged ≥ 21 years) residing in Singapore. The survey included the World Health Organisation Well-being Index (WHO-5) as a measure of mental well-being, with low WHO 5 scores (<13/25) indicating poor mental well-being. The authors analysed relationships between low WHO-5 score and a range of respondent characteristics using multivariate logistic regression.

Findings

Strikingly, 40.9% of 2,350 respondents analysed had low WHO-5 scores, indicating poor mental well-being. Parental non-acceptance, experience of conflict at home and bullying/discrimination in the workplace or educational environments were all significantly associated with poor mental well-being. Conversely, community participation appeared protective for mental well-being, as respondents who participated in LGBTQ community organisations or events were less likely to have poor mental well-being than non-participants.

Originality/value

The NLCS2013 represents one of the first broad-based efforts to comprehensively and quantitatively capture the sociodemographic and health profile, including mental health status, within Singapore’s resident LGBTQ population. These findings affirm the need to address the mental health needs of LGBTQ individuals in Singapore and to foster safe spaces and allyship.

Citation

Toh, G.W., Koh, W.L., Ho, J., Chia, J., Maulod, A., Tirtajana, I., Yang, P. and Lee, M. (2023), "Experiences of conflict, non-acceptance and discrimination are associated with poor mental well-being amongst LGBTQ-identified individuals in Singapore", Equality, Diversity and Inclusion, Vol. 42 No. 5, pp. 625-655. https://doi.org/10.1108/EDI-10-2021-0270

Introduction

Mental health and mental well-being are widely recognised as integral to public health and sustainable development, being inextricably linked with physical and overall health outcomes (Prince et al., 2007). In recognition of this, the United Nations 2030 Agenda for Sustainable Development includes a commitment to promoting both “physical and mental health and well-being …” (United Nations Development Programme, 2015). Yet, poor mental health remains a leading cause of morbidity and disability worldwide and was one of the top 10 causes of disability-adjusted life-years for people aged 10–49 years in 2019 (GBD Diseases and Injuries Collaborators, 2020). Moreover, since the start of the coronavirus disease 2019 (COVID-19) pandemic in 2020, studies have documented worsened mental well-being in populations worldwide (Buspavanich et al., 2021; Covid-19 Mental Disorders Collaborators, 2021).

Mental and physical health disparities affecting lesbian, gay, bisexual, transgender and queer (LGBTQ) minority populations have been widely documented. Notably, LGBTQ individuals may be at higher risk of depression and poor mental well-being than the general population (Baptiste-Roberts et al., 2017; Moagi et al., 2021; Valentine and Shipherd, 2018). It is recognised that these mental health disparities are closely linked to experiences of stress across multiple life domains, including the family, educational and workplace environments, wherever an individual’s identity as a member of a sexual or gender minority comes into conflict with the dominant social environment. The majority of the extant research on minority stress and its consequences for impact on LGBTQ mental health has come from studies in North America and Europe. Little or no published data are available for many Asian countries, including Singapore. This lack of quantitative information makes it challenging for local LGBTQ communities to comprehensively articulate their unmet health needs and hinders health advocacy efforts.

The National LGBT Census Singapore [NLCS2013 (NLCS Research Network, 2016)] was an exploratory study conducted in 2013 to address the dearth of quantitative information about Singapore’s LGBT population. The NLCS2013 sought to survey the needs and status of self-identified LGBT Singapore citizens and residents in the domains of health, housing, education, employment and family. To the best of our knowledge, the NLCS2013 was then the first effort to investigate a comprehensive range of health indicators and sociodemographic factors in multiple subgroups within the LGBTQ community in a developed Asian country. The NLCS2013 data thus provided a quantitative and comprehensive description of health status, including mental health, amongst LGBT-identified individuals in Singapore. In the present work, we address one key sub-domain within the NLCS2013, that of mental health and well-being. We characterised the mental health status of individuals in the community using the WHO-5, a brief, well-validated questionnaire that measures subjective well-being and risk of depression (Topp et al., 2015; World Health Organization, 1998). We then analysed this status with respect to sociodemographic, psychosocial and relational characteristics to better understand determinants of mental health and well-being amongst LGBT-identified individuals in our local setting.

Literature review

Minority stress and mental health disparities

The minority stress model and related conceptual frameworks (Goldbach and Gibbs, 2017; Hendricks and Testa, 2012; Meyer, 2003) represent a prominent theoretical orientation that formulates how stigma, prejudice and discrimination related to an individual’s minority group status may become a stressor that contributes to poor mental health. These frameworks broadly characterise stressors as external and internal, or distal and proximal, with experiences of prejudice and discrimination being examples of the former and identity concealment/non-disclosure and internalised homophobia being examples of the latter. Numerous studies have linked experiences of discrimination, stigma, bullying, abuse and internalised homophobia with higher levels of psychological distress and morbidity in LGBTQ individuals (Alvarez-Galvez and Salvador-Carulla, 2013; Hatzenbuehler et al., 2009; Hatzenbuehler et al., 2010; Hatzenbuehler and Pachankis, 2016; Lea et al., 2014; Mays and Cochran, 2001; McConnell et al., 2018; Tan et al., 2021c). The minority stress frameworks also encompass stress-ameliorating factors such as social support, community engagement and sense of belonging, which are understood as critical mechanisms that help protect against poor mental health outcomes (Meyer, 2003). A range of studies point to the roles of social interaction, peer support and participation in community groups in promoting mental health and resilience, reducing depression risk and moderating the effects of stigma and other negative experiences (Bockting et al., 2013; Fish et al., 2019; Fredriksen-Goldsen et al., 2013; Garcia et al., 2020; McLaren et al., 2013; Roberts and Christens, 2021).

Being LGBTQ in Singapore: sociocultural context and lived experiences

Singapore is a densely-populated multi-ethnic city-state in Southeast Asia with a total population of 5.45 million and a land area of 733 km² (Singapore Department of Statistics, 2021). With respect to the racial/ethnic group categories customarily used in Singapore, around three-quarters of the resident population are Chinese, with Malay (14%) and Indian (9%) and other (3%) ethnic groups accounting for the remainder. In terms of attitudes towards LGBTQ issues, Singapore civil society is generally regarded as conservative. A 2018 survey on religion, morality and conservatism in Singapore reported that 64% of respondents viewed same-sex relationships as “always”/“almost always” wrong, down from 80% of respondents in an earlier 2013 wave of the survey (Mathews et al., 2019). These and earlier studies (Detenber et al., 2013; Lim, 2002) suggest gradual shifts in attitudes on LGBTQ issues over the last few decades, although the overall outlook remains conservative. Results from a 2022 survey indicate further shifts in public attitudes on LGBTQ issues but also reveal increasing polarisation of views, notably across age groups (Ipsos, 2022a, b).

In August 2022, amid these shifts, Singapore’s Prime Minister, Lee Hsien Loong, announced the government’s intention to repeal Section 377A of the Penal Code, a law criminalising homosexual relations between consenting adult males (Lee, 2022). At the same time, the government clarified its intention to preserve heteronormative policies relating to marriage, public housing, education, media and potentially other domains (Tham, 2022). LGBTQ groups expressed relief at the proposed move to eliminate the potential for criminal prosecution of LGBTQ persons who decide to come out in order to access LGBTQ-focussed services or to seek acceptance within their families, social circles, workplaces or schools (Iau, 2022a). Leaders from the major religious communities in Singapore, including the Christian, Muslim, Buddhist, Hindu and other communities, voiced their support for preserving heterosexual marriage laws and raised concerns about the repeal of Section 377A, but called for compassion and tolerance amongst their respective communities (Iau, 2022a, b). On the other hand, Singapore’s Association of Small and Medium Enterprises affirmed that the business community welcomed and supported the repeal of Section 377A (Iau, 2022a). This mix of responses to the proposed legislative change illustrates the wide range of views and the variability in acceptance that LGBTQ individuals in Singapore may experience, depending on the communities or organisations to which they belong.

LGBTQ individuals in Singapore commonly encounter negative reactions from sources ranging from family members or friends/acquaintances to members of the general public (Oogachaga Counselling and Support, 2012; Sayoni, 2011, 2018, 2019; TransgenderSG et al., 2020a). For example, in a 2012 survey on the impact of homophobia and transphobia on LGBTQ individuals in Singapore, 60% of respondents reported experiencing one or more forms of discrimination or abuse related to their sexual orientation or gender identity (SOGI) (Oogachaga Counselling and Support, 2012). Although this most often took the form of verbal abuse, some respondents (especially transgender individuals) also reported experiences of threats, physical aggression and sexual attacks or harassment. LGBTQ individuals in Singapore encounter SOGI-related discrimination in a range of social spaces, including workplaces, schools and other public or private institutions (Oogachaga Counselling and Support, 2012; Sayoni, 2011, 2018, 2019; Tan et al., 2021a; TransgenderSG et al., 2020a). These reports illustrate common sources of minority stress within the immediate environment, which in turn is shaped by the wider social environment. Research has shown that the effects of individual-level minority stressors may be compounded by those of structural or institutional stigma and discrimination (Hatzenbuehler et al., 2009, 2010). The lack of SOGI-specific anti-discrimination legislation, the legal status of same-sex partnerships, barriers to gender marker change for transgender persons and restrictive media content guidelines, are widely cited as examples of these structural or institutional issues in the Singapore context. The impact of these issues on LGBTQ individuals has been detailed in civil society stakeholder reports submitted for the most recent cycle of the United Nations Human Rights Council Universal Periodic Review on Singapore (IndigNation et al., 2020; Pink Dot SG and Oogachaga, 2020; TransgenderSG et al., 2020b).

Taken together, published and unpublished research shows that LGBTQ minority individuals in Singapore encounter stressors related to their SOGI that their non-LGBTQ peers do not (IndigNation et al., 2020; Oogachaga Counselling and Support, 2012; Pink Dot SG and Oogachaga, 2020; Sayoni, 2011, 2018; Tan, 2019; Tan et al., 2020; TransgenderSG et al., 2020a; TransgenderSG et al., 2020b). Although there is a growing body of published research addressing LGBTQ issues in Singapore from sociocultural, behavioural, legal, political, economic, human resource management, social work and media studies perspectives, for example (bin Ibrahim and Barlas, 2021; Chua, 2014; Detenber et al., 2014; Detenber et al., 2013; Goh, 2008; Lim et al., 2018; Lim and Ang, 2021; Maulod, 2021; Oswin, 2010; Radics, 2015; Ramdas, 2020; Tan, 2015; Tan, 2011; Tan and Lee, 2007; Teh et al., 2015; Yue, 2007), there is surprisingly little published research on issues related to LGBTQ health, particularly mental health. Consequently, the impact of SOGI-related stressors on mental, physical or other dimensions of health in the local LGBTQ population is not well characterised, at least within the published literature. Older research publications focussed on the developmental and psychological profiling of homosexual and transsexual individuals in Singapore (Kok et al., 1991; Tsoi, 1990, 1992). Recent published health research remains relatively scarce and much of it has focussed on sexual health or alcohol/substance use amongst gay, bisexual and other men who have sex with men (Choong et al., 2012; Chua et al., 2013; Ong et al., 2021; Tan et al., 2021a, b, c; Wong et al., 2011). Outside of academia, surveys and interview-based studies conducted by regional and local LGBTQ organisations have highlighted some of the health challenges faced by LGBTQ individuals in Singapore (Fridae, 2010; Oogachaga Counselling and Support, 2012; Sayoni, 2011, 2018; TransgenderSG et al., 2020a).

Mental health and well-being in Singapore

Published research on correlates and determinants of mental well-being as well as mental health and morbidity in Singapore has largely focussed on the general population (Chong et al., 2012b; Ho, 2015; Picco et al., 2017; Subramaniam et al., 2014, 2019; Vaingankar et al., 2013, 2018). A 2013 survey of Singapore youth (defined as those aged 16–35 years) reported that well-being in this population was significantly correlated with self-rated health, marital/relationship status, educational attainment and personal or combined parental income (Ho, 2015). Within adult community samples studied, ethnic group and age were reported to be correlated with positive mental health (Vaingankar et al., 2013, 2018). Periodic national-level health survey programmes include selected mental health measures and provide limited trend data by age, gender and ethnic group. However, these datasets do not include information on respondents’ SOGI, precluding use of these data to address similar research questions in LGBTQ minority populations. Some local LGBT-focussed mental health research has been published in recent years, analysing associations between mental health outcomes (depression severity, suicidal ideation) and factors such as experiences of stigma or homophobia in gay, bisexual and queer men (Ong et al., 2021; Tan et al., 2021a, b, c). However, published data on other groups remain scarce or absent. Our knowledge of the physical and mental health status and needs of the LGBTQ population in Singapore thus remains highly fragmented. This exacerbates the challenges of needs assessment and resource planning that local community organisations face.

Methods

Study design, survey population and data collection

The NLCS2013 was a cross-sectional, anonymous online survey conducted between May and August 2013. An informal community-based participatory approach was adopted through consultation with community groups and service organisations. The design and delivery of the survey questionnaire was informed by consultation with the three largest LGBT non-profit organisations in Singapore at the time (Pink Dot SG, Sayoni and Oogachaga). Prior to dissemination, the questionnaire was reviewed by these organisations.

The survey was publicised by various LGBT organisations through their websites, social media platforms and email newsletters, which provided the link to the survey website. Survey respondents were informed of the nature of the survey, its objectives, intended uses and the partner organisations on the landing page of the survey. No financial incentive was provided for either attempting or completing the online survey. Before beginning the survey, respondents provided informed consent for their responses to be analysed and published in an aggregate, non-identifiable manner. Survey responses were anonymised from the point of collection as respondents’ Internet Protocol address (IP address) were not collected and no tracking links were used.

Respondents answered up to 54 questions with conditional branching based on their responses. The questionnaire included questions used in recent national-level population health surveys, as well as custom questions developed in consultation with local community organisations and drawing on the published literature. As a broad-ranging survey covering diverse areas of life, it was not feasible to include detailed instruments for every possible research question. We sought to limit respondent burden by prioritising questions that would support comparisons with earlier national-level health surveys, including a clinically validated index of mental well-being. Based on a pre-survey pilot, the questionnaire was estimated to take 20–40 min to complete.

For the present analysis, we included respondents who met three criteria: Singapore Citizens/permanent residents or non-residents living in Singapore, self-identified as LGBT and aged 21 years or older. Non-residents of Singapore, individuals who identified as cisgender and heterosexual and individuals aged below 21 were excluded.

Measures

Demographic and socioeconomic characteristics

Respondents provided information on their age, ethnicity, gender identity, sexual orientation, relationship status and current employment. Three indicators of socioeconomic status (SES) were used: personal monthly income, highest education level attained and current housing type.

Gender identity categories included male, female, transgender (male-to-female), transgender (female-to-male), intersex and “Others”. A six-category measure of sexual orientation was used (homosexual, mostly homosexual, bisexual, mostly heterosexual, heterosexual, unsure/other).

Psychosocial and relational characteristics

Respondents were asked about the extent to which they revealed their gender and sexual identity to family and/or friends. We used Likert scales for ordinal responses and included “unsure” and “not applicable” response options.

Three indicators were used to characterise common stressors in the home environment: Parental non-acceptance of LGBT identity (5-point Likert scale); recent conflict at home, investigated using the question “In the past 6 months, have you experienced conflict, harassment, threats or felt unsafe at home?”; homelessness related to conflict at home was investigated using 2 questions, “Have you ever been homeless (e.g. stayed in a public space, temporarily sheltered by others, etc.)?” and “Have you ever left home due to conflict, harassment, threats or feeling unsafe? – Yes, by my own choice; Yes, I was made/asked to leave; No”.

Stressors in workplace and educational institutions were explored by asking respondents whether they had experienced or witnessed bullying or discrimination related to LGBT identity in the workplace (working respondents) or in educational institutions (full-time students) within the previous 12 months. Bullying/discrimination was defined as “verbal abuse, non-verbal bullying (e.g. being gossiped about or ostracised), physical or sexual assault, being asked to change appearance or behaviour, being excluded from job opportunities or dismissed from employment”.

Social participation was investigated by asking about the types of LGBT-oriented groups, events or businesses the respondent had ever participated in or patronised.

WHO-5 well-being index

The World Health Organisation Well-being Index (WHO-5), a widely used brief standard measure in public health, is a self-report global rating scale that measures subjective positive well-being related to quality of life (Bech et al., 2003; Topp et al., 2015; World Health Organization, 1998). Lack of positive well-being is an indicator of possible depression and the WHO-5 has been validated as a population screening tool for depression in a range of general adult, adolescent and paediatric populations, as well as an outcome measure for health interventions (Henkel et al., 2003; Krieger et al., 2014; Lowe et al., 2004; Sischka et al., 2020; Topp et al., 2015). The WHO-5 is considered to have high clinometric validity (Hall et al., 2011), as it can be used in many different settings, irrespective of the presence or absence of comorbid conditions.

The WHO-5 is a simple, non-invasive tool comprising five positively-worded statements: “I have felt cheerful and in good spirits”, “I have felt calm and relaxed”, “I have felt active and vigorous”, “I woke up feeling fresh and rested”and “My daily life has been filled with things that interest me”. Responses are scored on a 5-point Likert scale: “All of the time”; “Most of the time”; “More than half of the time”; “Less than half of the time”; “Some of the time”; “At no time”. The recall period is two weeks. The range of scores is 0–25 (worst to best possible well-being), with higher scores indicating better well-being. A score of <13 of 25 (<50%) indicates impairment of well-being severe enough to warrant diagnostic follow-up, including clinical screening for depression (Krieger et al., 2014; Topp et al., 2015). The WHO-5 has high sensitivity (>80%) and specificity (>80%) for population screening of depression using the cut-off score of 50% (Topp et al., 2015). The use of the WHO-5 in our survey also permitted some comparisons with data from the general population, since earlier national population health surveys included this tool and analysed the data with respect to the same cut-off. In this sample, the WHO-5 index showed excellent internal consistency (Cronbach’s alpha >0.9) and a one-factor structure was verified by confirmatory factor analysis.

Data handling and statistical analysis

Prior to analysis, the raw dataset was inspected and cleaned to resolve quality issues, such as duplicate responses. Sociodemographic and psychosocial characteristics within the sample were analysed descriptively. Categorical variables were summarised using counts and percentages. The percentage of respondents with low WHO-5 (<13 of 25) was determined for the overall sample and by subgroups defined by sociodemographic and psychosocial characteristics. Associations between categorical variables were identified using the chi-squared test or Fisher’s exact test. Multivariable logistic regression was used to evaluate relationships between low WHO-5 and sociodemographic or other factors. Odds ratios adjusted for age (AORs) were presented, along with the corresponding 95% confidence intervals (CIs). There was no imputation of missing data. All analyses were performed using IBM SPSS Statistics version 20 (IBM, Armonk, New York, USA) and R version 2.15.3 (R Core Team, 2013).

Results

Sample sociodemographic characteristics

A total of 3,119 respondents completed the online questionnaire and were screened for eligibility based on the following criteria: Singapore Citizens/Permanent Residents or non-residents living in Singapore; self-identified as LGBT; aged 21 years or older. The present analysis included 2,350 valid responses from individuals who met all these inclusion criteria.

Table 1 presents the sociodemographic characteristics of the 2,350 respondents included in this analysis. Mean age was 30.7 (SD = 8.1) years; 85.9% (n = 2017) of respondents were below 40 years of age. There was a higher proportion of male (61.1%, n = 1,435) than female (36.3%, n = 854) respondents. A small percentage identified as transgender, intersex or of other genders (2.6%, n = 61). Most respondents were Chinese (80.0%, n = 1881); the remainder were Malay (6.3%, n = 148), Indian (4.5%, n = 105) or of other ethnic groups (9.2%, n = 216).

Most respondents had received tertiary education (69.6%, n = 1,194); 26.4% (n = 453) had at least post-secondary education and 4.0% (n = 68) received secondary level education or below. Close to half of the respondents (45.6%, n = 1,072) were in a same-gender relationship or established partnership (e.g. civil union or marriage). The remainder were single (44.7%, n = 1,051), in a relationship or marriage with another gender (4.1%, n = 96) or in other types of relationships (5.6%, n = 131). When asked about social participation, 95.4% (n = 2,242) of respondents reported participating in at least 1 type of LGBT-oriented group, activity or business, with LGBT-oriented online networking platforms being most common (65.6%, n = 1,542).

With respect to the national population profile from the Singapore Census of Population 2010 (Department of Statistics, 2010), survey respondents on average were younger, of higher SES and higher education levels. The proportion of males was higher than the national average, whereas the proportions of Malay- and Indian-identified individuals were lower. It should be noted that because basic quantitative data for Singapore’s LGBT-identified resident population (e.g. population size and characteristics) are lacking, it is unknown whether these differences reflect true differences in population proportions or are related to other factors.

Association of WHO-5 scores with demographic and socioeconomic characteristics

Within this sample, the mean WHO-5 score was 13.5 (SD 5.4), very close to the threshold (<13 of 25) that indicates impaired mental well-being warranting clinical screening for depression. Strikingly, 40.9% (n = 961) of NLCS2013 respondents reported a WHO-5 score of <13 out of 25 (Table 1), indicating poor mental well-being. Table 2 shows the breakdown of percentages of respondents with low WHO-5 (<13) by sociodemographic subgroups. Age was strongly associated with WHO-5 score (p = 0.0016), with the percentage of those with low WHO-5 being highest amongst the youngest respondents. Low WHO-5 score was also more common amongst respondents who were transgender/intersex/of other genders (54.1%, n = 33), in relationships or established partnerships with another gender (49.0%, n = 47), single (46.3%, n = 487) and those who identified as “Bisexual” (48.3%, n = 113) or “Mostly heterosexual” (50.0%, n = 37). Consistent across all three socioeconomic indicators (housing type, education level and monthly income), low WHO-5 was more common amongst individuals with lower SES.

Association of WHO-5 scores with non-acceptance of LGBT identity and conflict in the home and workplace

Experience of non-acceptance and conflict in home/family environment

Less than a quarter (22.2%, n = 513) of respondents had revealed their LGBT identity to “most/all friends and family”. Around one-third (38.2%, n = 883) had revealed their LGBT identity to “some friends but no family”, 34.9% (n = 807) to “some friends and family” and 4.8% (n = 110) to neither friends nor family (Supplementary Table S1). The extent to which respondents revealed their LGBT identity showed an inverse relationship with low WHO-5.

Respondents who had not revealed their LGBT identity to any friends or family were more than twice as likely to have a low WHO-5 score than respondents who had revealed their LGBT identity to most or all friends and family (AOR 2.739 [95% CI: 1.798 – 4.172], p < 0.0001) (Figure 1). Relative to respondents who had revealed their identity to only some friends or family, or to some friends but not family, respondents who had not revealed their identity to friends or family had higher odds of low WHO-5 (AOR 1.597 [95% CI: 1.261 – 2.022] and AOR 1.771 [95% CI: 1.402 – 2.237], respectively, p < 0.0001 for both).

We asked respondents about the degree of acceptance their parents had towards their LGBT identity. A substantial proportion of respondents (40.4%, n = 949) reported that they were unsure about the degree of parental acceptance, and 3.9% (n = 92) indicated this was not applicable. Of the remaining 1,039 respondents, half (53.2%, n = 697) had parents who were “Accepting” or “Neutral”. Lack of parental acceptance was strongly associated with low WHO-5 scores (p < 0.001). Compared with respondents who had accepting parents, those with non-accepting parents had significantly greater odds of low WHO-5 (AOR 1.744 [95% CI: 1.355 – 2.244], p < 0.0001) (Figure 1).

Recent conflict at home (experiencing conflict, harassments, threats or feeling unsafe at home in the previous 6 months) was reported by 8.5% (n = 200) of respondents (Supplementary Table S1). Lifetime homelessness related to such conflict at home was reported by 17.8% (n = 418) of respondents. Specifically, 5.5% (n = 130) of respondents had “ever been homeless”, whilst 13.0% (n = 305) had ever chosen to leave home and 3.4% (n = 79) had “ever been asked to leave”. Both recent conflict (p < 0.0001) and lifetime homelessness related to conflict (p = 0.001) were highly significantly associated with the degree of parental acceptance of respondents’ LGBT identity (Supplementary Table S3).

Importantly, recent conflict and lifetime homelessness due to conflict were also each associated with significantly greater odds of low WHO-5 (Figure 1; Supplementary Table S1). Respondents with recent experience of conflict were more than twice as likely to have low WHO-5 than respondents who had not (AOR 2.53, 95% CI: 1.873 – 3.416, p < 0.0001). For those who experienced lifetime homelessness due to conflict, the AOR of low WHO-5 was 1.77 (95% CI: 1.429 – 2.192, p < 0.0001).

Experience of bullying and discrimination in the workplace and school environment

More than a quarter of working respondents (27.9%, n = 545) reported that they had not revealed their LGBT identity to anyone at the workplace, 46.3% (n = 904) had revealed their identity to some colleagues and 25.8% (n = 503) had revealed their identity to many or all colleagues. Mirroring the observations in the domain of home and family life, the extent to which respondents revealed their LGBT identity in the workplace was inversely associated with WHO-5 scores. Respondents who had not revealed their identity to any colleagues were twice as likely to have a low WHO-5 score (AOR 2.084 [95% CI: 1.614 – 2.692], p < 0.0001) than those who revealed their identity to many/all colleagues (Figure 2a).

One in eight (12.5%, n = 211) working respondents reported that they had experienced at least one form of LGBT-related workplace bullying/discrimination in the previous 12 months (Supplementary Table S1). Experiencing bullying/discrimination was strongly associated with low WHO-5 (p < 0.0001). Respondents who had experienced at least one form of workplace bullying/discrimination in the previous 12 months had significantly higher odds of low WHO-5 compared to respondents who had neither experienced nor witnessed such bullying/discrimination (AOR 1.846 [95% CI: 1.373 – 2.481], p < 0.0001) (Figure 2a).

Responses from full-time students showed that over half (57.7%, n = 211) had experienced bullying/discrimination within their educational institutions. We observed similar relationships between bullying/discrimination in educational institutions and low WHO-5 score as those for workplace bullying/discrimination. Specifically, those who had experienced bullying/discrimination were significantly more likely to have a low WHO-5 score (AOR 2.115 [95% CI: 1.174– 3.813], p = 0.013) as compared to respondents who had neither experienced nor witnessed bullying/discrimination (Figure 2b).

Association of WHO-5 scores with social support and community participation

Our analysis revealed potentially protective effects of community participation and social support. Low WHO-5 was significantly less common amongst those who participated in LGBT community organisations (p < 0.0001), community events (p < 0.0001), or patronised LGBT-oriented parties, clubs and bars (p = 0.004), compared with their respective reference groups (Supplementary Table S2). The largest reductions in odds of low WHO-5 were seen amongst those who participated in community organisations (AOR 0.635 [95% CI: 0.538 – 0.750], p < 0.0001), community events (AOR 0.635 [95% CI: 0.538–0.750], p < 0.0001), or clubs/bars/parties (AOR 0.792 [95% CI: 0.668 – 0.938], p = 0.007) (Figure 3). Interestingly, amongst the “online-only” group who participated in internet-based LGBT-oriented networking platforms but no other LGBT-oriented groups, events or businesses, almost half (49.0%) reported low WHO-5. This “online-only” group was also more likely to have a low WHO-5 score as compared with respondents who participated in both in-person and online communities (AOR 1.423 [95% CI: 1.092–1.855], p = 0.009).

Discussion

High prevalence of poor mental well-being amongst LGBT-identified individuals in Singapore

In this analysis of a subset of the NLCS Singapore 2013 data, we investigated positive mental well-being and quality of life (WHO-5 Well-being Index) and its relationships with sociodemographic and relational factors in a large community sample of over 2000 LGBT-identified individuals. Building upon earlier surveys that concentrated on defined communities/groups (Fridae, 2010; Sayoni, 2011) or on specific topics (Fridae, 2010; Oogachaga Counselling and Support, 2012), the NLCS was the first large local study to sample individuals across the entire spectrum of sexual orientation and gender identity in Singapore. This was achieved by publicising the survey through partnership with the largest non-profit organisations that served various segments within the community. In addition, the NLCS retrieved comprehensive data on self-reported physical and mental health, health-related behaviours and quality of life of LGBTQ individuals in Singapore, to help address important and long-standing data gaps in these areas. The present analysis of positive mental well-being used the WHO-5 Well-being Index, a validated instrument included in earlier national health survey programmes, thereby allowing some comparisons with available national-level data.

A WHO-5 score of <13 of 25 indicates poor mental well-being, warranting clinical follow-up for depression screening. In this sample, the mean score was 13.5, just above this clinically meaningful threshold. Out of 2,350 respondents analysed, 40.9% had a WHO-5 score of <13. Strikingly, this percentage of respondents with an at-risk WHO-5 score was nearly four times the general-population estimate (11.7%) in a nationally-representative periodic health survey, the 2007 National Health Surveillance Survey (NHSS2007) (Ministry of Health Singapore (Epidemiology and Disease Control Division), 2007). Within every age-group category, the proportion of individuals with poor mental well-being was much higher in the NLCS sample than that reported in the NHSS2007, particularly for those aged 18–29 (NLCS: 42.1%; NHSS2007: 9.2%). We noted similar trends across other sociodemographic variables where comparable categories were used. These included three major race/ethnicity categories customarily used in Singapore (Chinese, Malay and Indian) and Male/Female categories. Since the NHSS2007 did not survey respondents’ gender identity, direct comparisons are not possible for gender identity categories. Despite differences in sampling methodology, we believe that these strikingly large differences may reflect true disparities and are a cause for concern.

We observed associations between mental well-being and relationship/partnership status, educational attainment and income level that were consistent with correlates of well-being identified in a 2013 survey of a general-population sample of young adults (15–34 years) in Singapore in a similar time period (Ho, 2015). Apart from socioeconomic factors such as low income and unemployment (Chong et al., 2012a), that were associated with poorer mental health in adult community samples, a number of other proximal and distal risk factors related to the home and workplace environment emerged in our analyses. Specifically, poor mental well-being was strongly associated with respondents’ experience of limited or non-disclosure of their SOGI, parental non-acceptance of their SOGI, conflict at home and lifetime homelessness due to conflict. Poor mental well-being was also strongly associated with experience of bullying/discrimination in other major social spaces, namely workplaces and educational institutions. Our analyses also revealed potentially protective factors, notably social participation in LGBT-focussed community groups and activities. Taken together, these findings emphasise the relevance of minority stress as an additional social determinant of health for LGBT individuals in the local context.

Parental non-acceptance and family conflict are strongly associated with poor mental well-being

The strong relationships identified between home/family-related stressors and poor mental well-being in our sample are highly consistent with the processes described in minority stress frameworks (Goldbach and Gibbs, 2017; Hendricks and Testa, 2012; Meyer, 2003). The framework also posits interdependency amongst a number of stress processes and, indeed, our results suggest a similar inter-relatedness amongst the factors that we explored. For example, parental non-acceptance of respondents’ SOGI was strongly associated with limited or lack of identity disclosure and with experiencing conflict at home, and all three factors were strongly associated with poor mental well-being. Parental non-acceptance can manifest as psychological or physical conflict, ranging from verbalised disapproval of a child’s SOGI, to harassment or threats of physical violence, leading individuals to feel unsafe at home, as explored in our study. Our findings are corroborated by existing literature, which suggests that parents’ objection to the child’s identity and their attempts to influence or alter their children’s SOGI have a strong negative impact, such as increasing emotional stress, negativity towards their identity, suicidality and substance use amongst LGBTQ youths (D’Amico and Julien, 2012; D’Amico et al., 2015). Conversely, more accepting parental attitudes towards the child’s sexual orientation were associated with protective effects, specifically against proximal stressors (Feinstein et al., 2014).

In our analysis, less extensive disclosure of respondents’ SOGI to peers and family was associated with poorer mental well-being. Although this survey did not distinguish between non-disclosure and concealment of identity, the two concepts are related and share similar psychosocial components (Beals et al., 2009; Pachankis et al., 2020), with recent research specifically identifying concealment behaviour as a negative predictor of psychological well-being and self-identity (Jackson and Mohr, 2016). Consistent with the minority stress model, concealment or non-disclosure of SOGI status may be a significant source of proximal stress, whilst at the same time it can be a protective coping strategy, reducing victimisation and discrimination in some circumstances (Meyer, 2003).

Apart from psychological distress and poor mental health outcomes, our analysis suggests that SOGI-related conflict at home may have further potentially serious consequences. Close to one in five respondents (17.8%) reported lifetime homelessness related to conflict at home. Studies have consistently found higher rates of homelessness amongst LGBT youth than amongst their peers, which is cause for concern, since homelessness exposes LGBTQ individuals to complex health risks, as well as additional barriers to healthcare due to economic instability (Corliss et al., 2011; Rew et al., 2005; Whitbeck et al., 2004).

Experience of discrimination/bullying in workplaces and educational institutions is strongly associated with poor mental well-being

Experiences of discrimination, bullying, harassment and abuse, whether based on race, gender, sexuality, disability or other attributes, are associated with poorer health status and greater functional limitation (Burgess et al., 2007; Mays and Cochran, 2001; Okechukwu et al., 2014). In our sample, respondents who had recently experienced LGBT-related bullying/discrimination in the workplace were significantly more likely to have poor mental well-being than respondents without such experiences. In the workplace, organisation-wide adoption of inclusive, non-discriminatory policies, backed by strong support from management, would help promote psychological security and well-being amongst all employees, including LGBT individuals (Boekhorst, 2015; Webster et al., 2018). Additionally, numerous studies have reported the practical benefits of inclusive policies in creating economic value for both businesses and societies as a whole (Hossain et al., 2019; Lee Badgett et al., 2013, 2019), demonstrating that workplace inclusivity does indeed make good business sense.

In the educational setting, we found that full-time students who had experienced bullying/discrimination were more than twice as likely to have poorer mental well-being as compared to students who had not experienced bullying/discrimination. Bullying/discrimination, regardless of its basis, is well-known to be detrimental to mental health (Karanikola et al., 2018) and has also been linked to poorer academic performance (Samara et al., 2021).

Protective effects of community participation on mental health outcomes

Our study also adds to existing evidence on the protective effects of community participation and social support within the LGBTQ community (Frost and Meyer, 2012; McConnell et al., 2015, 2018), corresponding to the ameliorating factors and coping mechanisms described in the minority stress framework. Respondents who participated in community organisations or events had significantly lower odds of poor mental well-being compared with non-participants. This reinforces the potential importance of community connectedness and relationships with other LGBTQ individuals for social support and empowerment (Garcia et al., 2020), especially for those without supportive families. While we did not measure these specific attributes in this survey, we note with interest that various self-protective mechanisms, including in-group comparisons and selective devaluing, may provide a buffering or self-protective effect against the negative impacts of social stigma (Crocker and Major, 1989). Overcoming isolation, connecting with a community and sharing similar lived experiences can help individuals normalise their identities in a healthy and empowering way.

Interestingly, amongst respondents who participated only in online LGBT-oriented networking platforms, a significantly higher proportion (49.0%) had WHO-5 scores indicating poor mental well-being than those who participated in in-person community organisations/groups (33.6%) or community activities/events (36.0%). Although existing research points to the benefits of virtual communities in providing social support to LGBTQ individuals (Garcia et al., 2020; Wilson and Cariola, 2019), our results add an extra dimension, suggesting that in-person interactions still matter. We acknowledge that the differences between those reporting exclusively-online participation versus a mixture of in-person and online participation could be attributed to factors such as fear of negative consequences and/or internalised homophobia. This could result in bias favouring community participation amongst individuals with greater resilience to these factors.

Limitations and future research

A number of limitations should be noted. Given the one-time cross-sectional nature of the survey, no formal causal inferences can be made based solely on these data. In view of the conservative social climate and practical resource constraints, the NLCS2013 was conducted as a single online anonymous survey, using snowball sampling and leveraging on the social reach of the largest LGBTQ community organisations in Singapore. If resources permit, future surveys should be designed to allow ongoing data collection and analysis, which could help to mitigate the limitations of a single-administration design. Although we readily acknowledge the inherent limitations of the non-probability sampling method employed, under the circumstances we consider it a reasonable and practical alternative to traditional population-based methods, especially for reaching sufficient numbers of individuals within the target population (Hidaka and Operario, 2006). In the presence of social stigma and the retention of anti-homosexuality laws, attempting to use traditional methods such as interview-based surveys and random household sampling would create practical challenges and introduce other sources of selection and non-response bias that are equally challenging to account for.

As with any self-report survey involving voluntary participation, the potential influence of selection and non-response bias must be considered. However, we note that the lack of accurate knowledge of the sociodemographic profile and other characteristics of the underlying target population presents practical difficulties for estimating the impact of such bias. The majority of respondents were Chinese, relatively young (<40 years), well-educated and identified as homosexual/mostly homosexual. The numbers of transgender or other-gender identity respondents and non-Chinese respondents were relatively small, limiting interpretation of the data for these groups. Given the online-only nature of the survey and the survey language (English only), there could also have been a bias towards individuals who were English-literate and comfortable with the use of the internet and online media. This could have contributed to an underrepresentation of older individuals, those with lower income or education levels, as well as other groups with limited access to online media. Considering the negative influence of factors such as low SES or education levels on mental health and well-being (Alegria et al., 2018; Ng et al., 2014; Wee et al., 2014, 2017), the proportion of individuals with poor mental well-being within the LGBT-identified population as a whole could have been even higher than we have reported here.

The main limitations are shared with many studies on LGBT health conducted in earlier decades in North America and Europe (Coyle, 1993; King et al., 2003; Lock and Steiner, 1999; Meyer, 1995), as well as more contemporary studies in other Asian countries where anti-LGBT stigma remains prevalent (Feng et al., 2012; Hidaka and Operario, 2006; Patel et al., 2013; Sivasubramanian et al., 2011). Such research has historically relied on internet-based or other non-random convenience samples, such as attendees at medical clinics, LGBT events or patrons of LGBT-oriented businesses. Despite such constraints, it is noteworthy that early studies on community samples were often able to identify trends consistent with those from population samples in later and more broadly representative surveys (Gonzales et al., 2016; Sandfort et al., 2006). This was possible because these later surveys included questions pertaining to respondents’ sexual orientation and/or gender identity, which are demographic dimensions that may be as keenly relevant to health as race/ethnicity (Cahill and Makadon, 2014; Streed et al., 2020). We recommend that all health-related surveys, particularly those that seek to achieve population-level coverage, offer respondents the option to provide SOGI information alongside other demographic data. Collection of accurate and meaningful SOGI data, with appropriate safeguards in place, is a crucial first step towards understanding and addressing health disparities in sexual and gender minority populations.

In Asia, published research on the health needs of LGBT individuals, especially at a nationally representative level, remains relatively scarce. Our results point to a disproportionately high prevalence of poor mental well-being and possible risk of depression amongst LGBT-identified people in Singapore, similar to what has been described for sexual and gender minorities in a number of other developed Asian societies (Chan et al., 2020; Hidaka and Operario, 2006; Yi et al., 2017). The cross-sectional study by Hidaka et al. in 2006 illustrates the challenges faced in early-phase research on mental health in sexual minority populations. Although focussing on correlates of attempted suicide in gay/bisexual/queer Japanese men, this study identified common and persistent themes in LGBTQ mental health research, including high levels of depression, verbal harassment and bullying related to sexual orientation (Hidaka and Operario, 2006). A decade later, in South Korea, Yi et al. estimated a 5–7-fold higher prevalence of depressive symptoms amongst LGB individuals relative to the general population (Yi et al., 2017). Of note, their paper discusses the challenges involved in making comparisons between the LGB and general population, related to the lack of measures to capture SOGI in nationally-representative surveys on adults. A 2019 survey of LGBT adults in Hong Kong reported that 29.8% of respondents met criteria for probable clinical depression, which was double the proportion estimated for the general population in a separate population-level survey that covered a similar time-frame (Chan et al., 2020).

To move beyond the limitations of earlier studies, a number of challenges need to be overcome, most notably that of obtaining suitable population-based samples. As discussed above, inclusion of appropriate SOGI variables in population-level surveys would facilitate the identification of individuals from the group(s) of interest and would also support analyses that use appropriate non-LGBTQ comparison groups to contextualise findings. To date, the number of countries where this has been implemented is relatively small. In the meantime, one priority for local research is to improve basic estimates of the size, sociodemographic profile and other characteristics of the LGBTQ-identified resident population in Singapore. This would help researchers to understand and estimate the impact of various types of bias in local samples. The findings could also inform aspects of research design for future studies, such as meaningful category definitions for certain variables or identifying certain population subgroups that may need to be over-sampled, depending on the specific study objectives.

Within this context, the present study of a large community sample is best understood as early-phase work in this population, primarily descriptive and hypothesis-generating. This comprehensive dataset documented multiple aspects of health (self-reported physical and mental health, health-related behaviours and quality of life) amongst LGBTQ individuals, providing a valuable reference point for this time period in the community’s history. Despite the limitations, the implications of the findings still appear relevant, as a number of the observed trends emerged in later research studies as well. Our analysis identified both risk and protective factors, showing parallels with research on LGBTQ populations elsewhere and indicating important areas of unmet need in our local setting. In particular, steps to improve the home and workplace climate are still urgently needed.

Implications for stakeholders and society

Despite indications of gradual shifts in societal attitudes on LGBT issues, recent studies show that LGBTQ individuals in Singapore remain vulnerable to violence and discrimination in the home/family environment (Sayoni, 2018; TransgenderSG et al., 2020a). In summing up observations from interviews of 40 LBTQ individuals, researchers have reported that the threat of or actual violence and abuse very often comes from immediate family members or relatives who “believe that homosexuality is wrong and that LBTQ individuals must be punished or ‘cured’ of their homosexual tendencies” (Sayoni, 2018). Arguably, the greatest impact is on younger individuals who are not financially independent and lack the means to move out of the family home. Of more than 200 trans/non-binary individuals surveyed in 2020, nearly a quarter reported that they had experienced violence from a family member or intimate partner (TransgenderSG et al., 2020a). In view of the above, our findings on mental well-being and conflict at home and “lifetime” homelessness due to such conflict serve to emphasise the importance of the family environment as a safe space, physically and psychologically. Access to safe and adequate housing remains a major concern for LGBTQ individuals, due to a combination of factors such as the high cost of housing and public housing subsidy policies that favour heteronormative family units, which have been examined in detail elsewhere (Oswin, 2010, 2019).

Similarly, studies show that continued efforts are needed to address discrimination against LGBTQ individuals in the workplace. A 2017 audit in four South-East Asian countries reported high levels of pre-employment discrimination against transgender people in Singapore, finding that cisgender job applicants were 80% more likely to get a positive response to their applications and 100% more likely to be invited for job interviews than transgender applicants with equivalent qualifications and experience (Winter et al., 2018). Apart from barriers in seeking employment, about a quarter of transgender individuals surveyed reported negative workplace experiences such as repeated or intentional mis-gendering, being gossiped about or asked to change their behaviour/appearance or “lifestyle” (TransgenderSG et al., 2020a). A Singapore study found evidence of discrimination by hiring personnel against gay and lesbian job applicants for task-interdependent occupations, which require greater interaction with co-workers. Countering assumptions that workers would be uncomfortable working closely with gay or lesbian colleagues, the study found that gay men and lesbians in high-task-interdependent jobs were actually more—not less—likely to be invited by co-workers to socialise outside of work (Lim et al., 2018). Such research provides encouraging evidence that normalising interactions and interpersonal contact with LGBTQ individuals in the workplace, such as working closely with a gay or lesbian co-worker, can help reduce prejudice and enhance social interactions within and outside of the workplace.

As for educational settings, no official data on SOGI-related discrimination or bullying in Singapore schools or higher education institutions are available. However, our findings and those of others indicate that SOGI-related bullying/discrimination in educational institutions has been and remains common (Oogachaga Counselling and Support, 2012; Sayoni, 2018, 2019; TransgenderSG et al., 2020a), leading many students to feel that they lack safety and support in these spaces. In response to the identified needs of LGBTQ students in Singapore’s higher education institutions, student-run volunteer organisations have developed sexual health and mental health resources, signalling that these are still significant areas of concern (Inter University LGBT Network, n.d.).

Both empirical and theoretical research underscore the importance of support and acceptance from parents, families and society at large, as these factors are strongly protective against poor mental health and well-being in LGBTQ individuals (Garcia et al., 2020; Hatzenbuehler et al., 2011; Simons et al., 2013). We would expect these factors to be just as relevant in the Singapore context, where it has been shown that well-being in young adults is strongly correlated with family support (Ho, 2015). Given that many LGBTQ individuals remain vulnerable to factors within the family/home environment that adversely affect their mental health, our findings remain relevant and highlight the importance of mitigating the negative impact of family non-acceptance and managing conflict situations, particularly for younger individuals. Enhanced access to LGBTQ-friendly and culturally-sensitive mental health and social services could benefit LGBTQ individuals of all ages as well as their families. The findings also emphasise the need to improve the handling of incidents of SOGI-related bullying and harassment in educational institutions, which would ultimately help improve the well-being of LGBTQ students. Implementation of anti-bullying guidelines and policies with SOGI-specific clauses, could empower educators and other staff to support students who report SOGI-related bullying or discrimination and promote a greater sense of safety within educational institutions.

The recent announcement of the intention to repeal Section 377A in Singapore (Lee, 2022) raises interesting questions about what this change might mean for LGBTQ individuals in everyday life. In other Asian jurisdictions, such as Hong Kong and India, there was increased demand for mental health support and social services in LGBTQ communities following the repeal of anti-homosexuality laws. A similar trend is likely to be observed following Singapore’s repeal of Section 377A, once LGBTQ individuals can be assured of not facing prosecution if they seek access to LGBTQ-focussed services and support (Oogachaga Counselling and Support, personal communication to the authors). This trend is likely to be compounded by the effects of the Covid-19 pandemic, which has greatly increased awareness of mental well-being and its importance in Singapore. In August 2021, the Singapore government announced the formation of the Inter-Agency Taskforce on Mental Health and Well-being. In July 2022, following a public consultation exercise, the Taskforce announced three focus areas, namely “(a) improve accessibility, coordination and quality of mental health services; (b) strengthening of services and support for youth mental well-being; (c) improve workplace well-being measures and employment support” (Ministry of Health Singapore, 2022).

These shifts towards promoting individual and community mental health and towards greater acceptance of LGBTQ individuals, coupled with the high prevalence of poor mental well-being in the LGBTQ community, signal the urgency of capacity- and capability-building for mental health and social work professionals and their accompanying support systems. Studies of family physicians and social workers in Singapore have revealed considerable needs in terms of training and preparedness to serve LGBTQ clients (Lim and Ang, 2021; Teh et al., 2015) and such needs are likely to exist within other medical and social service disciplines as well. Our research provides empirical knowledge of factors that contribute to poor mental well-being amongst LGBTQ individuals in Singapore, which will be useful for planning of mental well-being community interventions and individual treatment plans and for the training of service providers. Attitudes and beliefs associated with parental non-acceptance, or bullying and discrimination in schools or the workplace, may be slow to shift, even with the repeal of Section 377A, so are likely to remain relevant. In the long term, broad-based efforts to reduce stigmatisation at all levels may be the most effective and sustainable way of mitigating health disparities amongst stigmatised groups (Hatzenbuehler et al., 2013; Lim et al., 2014; Mule et al., 2009).

Conclusions

Our study of mental well-being in LGBTQ individuals in Singapore adds to the body of evidence on the health status of this population. Together with prior and subsequent research, our findings highlight a number of unmet health needs in this population, including but not limited to mental well-being. Our findings illustrate different ways in which SOGI minority status may contribute to disparities in mental well-being and thus point to the relevance of this social determinant of health and well-being in the local context. The overall well-being of our society as a whole depends on the well-being of all its communities and individuals. The success of broad-based efforts to improve health in the population as a whole thus depends upon understanding and addressing the full range of applicable social determinants of health.

Further targeted research to capture information on the LGBTQ population’s overall health status and unmet physical and mental health needs is essential to guide interventions that are effective and sustainable in the long term. Over the period since this research was conducted, dedicated community groups have continued their efforts to document and address the immediate mental health and related needs of the communities they serve, whilst also engaging with a range of stakeholders within the wider community in Singapore. In the long term, what will ultimately be needed to address such health disparities is concerted and continued commitment to education, destigmatisation and promoting acceptance across the multiple social spaces that LGBTQ individuals must navigate, including families, educational institutions and workplaces.

Figures

Figure 1

Home/family setting: factors associated with poor mental well-being (WHO-5 <13 of 25)

Figure 2

Workplace and educational setting: factors associated with poor mental well-being (WHO-5 <13 of 25)

Figure 3

Protective effect of community and social participation on mental well-being

Table 1

Characteristics of the survey respondents (N = 2,350)

Variable	Total analysed (N = 2,350)
Age (years)
Mean (SD)	30.7 (8.1)
Median (Q1, Q3)	29 (24, 35)
21–29	1,268 (54.0)
30–39	749 (31.9)
40–49	254 (10.8)
>50	79 (3.4)
Gender identity
Male	1,435 (61.1)
Female	854 (36.3)
Transgender, intersex, other gender	61 (2.6)
Sexual orientation
Homosexual	1,555 (66.2)
Mostly homosexual	398 (16.9)
Bisexual	234 (10.0)
Mostly heterosexual	74 (3.1)
Other or unsure	89 (3.8)
Ethnic identity
Chinese	1881 (80.0)
Malay	148 (6.3)
Indian	105 (4.5)
Other	216 (9.2)
Relationship status
Single	1,051 (44.7)
Same-gender relationship or civil union/marriage	1,072 (45.6)
Opposite-gender relationship or civil union/marriage	96 (4.1)
Other2	131 (5.6)
Housing type
Public housing 1–3 room)	366 (15.7)
Public housing (4–5 room)	1,179 (50.7)
Private property	706 (30.4)
Other	75 (3.2)
Education level
Tertiary (university and post-graduate)	1,194 (69.6)
Post-secondary	453 (26.4)
Secondary or below	68 (4.0)
Monthly income (Singapore dollars)
<2000	290 (17.0)
2001–4,000	640 (37.5)
4,001–6,000	367 (21.5)
>6,000	410 (24.0)
Current employment status
Local organisation employee	846 (48.2)
Multinational organisation employee	484 (27.5)
Self-employed or business owner	238 (13.5)
Seeking employment	147 (8.4)
Retired, homemaker, or intentionally not working2	42 (2.4)
Social participation
Community groups	556 (23.7)
Community events	1,293 (55.0)
Parties, clubs, bars (LGBT-oriented)	1,475 (62.8)
Saunas or spas (LGBT-oriented)	605 (25.7)
Online networking platforms (LGBT-oriented)	1,542 (65.6)
Only online networking platforms (LGBT-oriented)	247 (10.5)
At least 1 type of LGBT-oriented group, activity, or business	2,242 (95.4)
WHO-5 score (range: 0 to 25)
Mean (SD)	13.5 (5.4)
Median (Q1, Q3)	14 (10, 18)
WHO-5 score <13 of 25	961 (40.9)
WHO 5 score ≥13 of 25	1,389 (59.1)

Note(s): ¹ Includes individuals in open or multiple relationships, separated/divorced/widowed from a marriage/civil union, other

² Includes retirees, homemakers and those intentionally not working

Table 2

Demographic and socioeconomic factors associated with low WHO-5 (<13 of 25)

Variable	WHO-5 ≤13 N (%)	p-value1	Age-adjusted OR (95% CI)2	p-value3
Age (years)		0.0016
21–29 years	534 (42.1)
30–39 years	312 (41.7)
40–49 years	99 (39.0)
>50 years	16 (20.3)
Gender identity		0.0399
Male	566 (39.4)		1
Female	362 (42.4)		1.077 (0.904, 1.283)	0.408
Transgender, intersex, other gender	33 (54.1)		1.730 (1.032, 2.898)	0.038
Sexual orientation		0.0079
Homosexual	598 (38.5)		1
Mostly homosexual	171 (43.0)		1.146 (0.914, 1.438)	0.238
Bisexual	113 (48.3)		1.419 (1.073, 1.875)	0.014
Mostly heterosexual	37 (50.0)		1.519 (0.950, 2.429)	0.081
Other/unsure	42 (47.2)		1.339 (0.870, 2.062)	0.185
Ethnic identity		0.0018
Chinese	786 (41.8)		1
Malay	65 (43.9)		1.058 (0.704, 1.590)	0.787
Indian	47 (44.8)		0.825 (0.496, 1.372)	0.459
Other	63 (29.2)		0.578 (0.402, 0.833)	0.003
Relationship status		<0.0001
Single	487 (46.3)		1
Same gender relationship or established partnership	373 (34.8)		0.639 (0.535, 0.762)	<0.0001
Opposite gender relationship or established partnership	47 (49.0)		1.135 (0.746, 1.726)	0.554
Other4	54 (41.2)		0.883 (0.607, 1.284)	0.514
Housing type		0.0066
Public housing (1–3 room)	160 (43.7)		1
Public housing (4–5 room)	505 (42.8)		0.921 (0.725, 1.169)	0.498
Private property	249 (35.3)		0.706 (0.545, 0.915)	0.008
Other	31 (41.3)		0.878 (0.530, 1.456)	0.615
Education level		0.0001
Tertiary	457 (38.3)		1
Post-secondary	220 (48.6)		1.425 (1.141, 1.779)	0.002
Secondary or below	36 (52.9)		1.929 (1.177, 3.162)	0.009
Monthly income (Singapore dollars)		<0.0001
<2000	157 (54.1)		1
2001–4,000	286 (44.7)		0.694 (0.525, 0.918)	0.010
4,001–6,000	132 (36.0)		0.496 (0.360, 0.684)	<0.0001
>6,000	135 (32.9)		0.459 (0.325, 0.648)	<0.0001
Current employment status		0.0003
Local organisation employee	365 (43.1)		1
Multinational organisation employee	189 (39.0)		0.872 (0.694, 1.097)	0.243
Self-employed or business owner	75 (31.5)		0.649 (0.476, 0.885)	0.006
Seeking employment	78 (53.1)		1.438 (1.010, 2.047)	0.044
Not working5	14 (33.3)		0.814 (0.414, 1.601)	0.550

Note(s): ¹ p-value from chi-squared (χ²) test

²Multivariable logistic regression was used to estimate the odds ratio of low WHO-5 (<13 of 25), adjusted for age

³ p-value from Wald test

⁴Other includes individuals in open or multiple relationships, separated/divorced/widowed from a marriage/civil union

⁵Includes retirees, homemakers and those intentionally not working

Table S1

Psychosocial and relational factors associated with poor mental well-being (WHO-5 <13 of 25)

Variable	Total N (%)	WHO-5 < 13 N (%)	p-value1	Adjusted OR (95% CI)2	p-value3
Home and family setting
Revealed LGBT identity to			<0.0001
Most or all friends and family	513 (22.2)	155 (30.2)		1
At least some friends or some family	807 (34.9)	337 (41.8)		1.597 (1.261, 2.022)	<0.0001
At least some friends but not any family	883 (38.2)	395 (44.7)		1.771 (1.402, 2.237)	<0.0001
Neither friends nor family	110 (4.8)	60 (54.5)		2.739 (1.798, 4.172)	<0.0001
Parental acceptance of LGBT identity4			<0.0001
Accepting	524 (40.0)	162 (30.9)		1
Neutral	173 (13.2)	60 (34.7)		1.153 (0.800, 1.662)	0.446
Non-accepting	612 (46.8)	276 (45.1)		1.744 (1.355, 2.244)	<0.0001
Recent conflict at home (past 6 months)5			<0.0001
No	2,150 (91.5)	836 (38.9)		1
Yes	200 (8.5)	125 (62.5)		2.530 (1.873, 3.416)	<0.0001
Lifetime homelessness related to conflict at home6			<0.0001
No	1932 (82.2)	744 (38.5)
Yes	418 (17.8)	217 (51.9)		1.770 (1.429, 2.192)	<0.0001
Workplace setting
Revealed LGBT identity to peers			<0.0001
Many or all	503 (25.8)	151 (30.0)		1
Some or a few	904 (46.3)	398 (44.0)		1.812 (1.436, 2.286)	<0.0001
None	545 (27.9)	254 (46.6)		2.084 (1.614, 2.692)	<0.0001
Bullying/discrimination related to LGBT identity (past 12 months)7
Neither experienced nor witnessed bullying/discrimination	1,251 (74.3)	486 (38.8)	<0.0001
Witnessed bullying/discrimination only	222 (13.2)	94 (42.3)		1.150 (0.860, 1.538)	0.345
Experienced bullying/discrimination	211 (12.5)	116 (55.0)		1.846 (1.373, 2.481)	<0.0001
Educational setting
Revealed LGBT identity to peers			0.1562
Many or all	74 (20.4)	27 (36.5)		1
Some or a few	168 (46.3)	76 (45.2)		1.429 (0.813, 2.513)	0.215
None	121 (33.3)	42 (34.7)		0.921 (0.503, 1.686)	0.790
Bullying/discrimination related to LGBT identity (past 12 months)7
Neither experienced nor witnessed bullying/discrimination	67 (18.3)	20 (29.9)	0.0029
Witnessed bullying/discrimination only	88 (24.0)	26 (29.5)		0.986 (0.492, 1.976)	0.968
Experienced bullying/discrimination	211 (57.7)	100 (47.4)		2.115 (1.174, 3.813)	0.013

Note(s):¹ p-value from chi-squared (χ²) test

²Multivariable logistic regression was used to estimate the odds ratio of low WHO-5 (<13 of 25), adjusted for age

³p-value from Wald test

⁴Excludes “Not applicable”, “unsure” responses (n = 1,401)

⁵Experienced conflict, harassment, threats, or felt unsafe at home in the past 6 months

⁶Ever made homeless or ever left home due to conflict, harassment, threats or feeling unsafe at home

⁷Experienced or witnessed at least one form of bullying/discrimination related to LGBT identity (including verbal abuse, bullying, being asked to change appearance or behaviour, physically or sexually assaulted, discriminated against or excluded from job assignments/promotions or dismissed from employment)

Table S2

Protective effects of community and social participation on mental well-being

Social participation1	Participated: Yes	Low WHO-5 (<13 of 25) by participation status (yes/No)			Odds of low WHO-5 (<13 of 25)
		Yes (%)	No (%)	p-value2	Adjusted OR3,4	p-value5
At least 1 type of LGBT-oriented organisation, activity, or business	2,242 (95.4)	40.5	48.1	0.116	0.746 (0.506, 1.100)	0.139
LGBT community organizations or groups	556 (23.7)	33.6	43.1	<0.0001	0.684 (0.560, 0.835)	<0.0001
LGBT community activities or events	1,293 (55.0)	36.0	46.9	<0.0001	0.635 (0.538, 0.750)	<0.0001
Parties, clubs, bars (LGBT-oriented)	1,475 (62.8)	38.6	44.7	0.004	0.792 (0.668, 0.938)	0.007
Saunas or spas (LGBT-oriented)	605 (25.7)	40.5	41.0	0.817	1.062 (0.874, 1.291)	0.545
Online networking platforms (LGBT-oriented)	1,542 (65.6)	41.8	39.1	0.203	1.151 (0.966, 1.371)	0.116
Only online platforms (LGBT-oriented)	247 (10.5)	49.0	39.9	0.006	1.423 (1.092, 1.855)	0.009

Note(s): ¹ Respondents selected one or more options as applicable

² p-value from chi-squared (χ²) test

³ AOR, odds ratio of low WHO-5 (<13 of 25), adjusted for age

⁴ Reference group: Did not participate

⁵ p-value from Wald test

Table S3

Relationship between parental acceptance of LGBT identity and experience of conflict at home

Variable	Recent conflict (past 6 months)1			Lifetime homelessness related to conflict at home2
	Total N (%)	Yes N (%)	p-value3	Total N (%)	Yes N (%)	p-value3
Parental acceptance of LGBT identity4			<0.0001			0.001
Accepting	524 (40.0)	24 (4.6)		524 (40.0)	88 (16.8)
Neutral	173 (13.2)	13 (7.5)		173 (13.2)	39 (22.5)
Non-accepting	612 (46.8)	90 (14.7)		612 (46.8)	157 (25.7)

Note(s): ¹ Experienced conflict, harassment, threats, or felt unsafe at home in the past 6 months

² Ever made homeless or ever left home due to conflict, harassment, threats, or feeling unsafe at home

³ p-value from chi-squared (χ²) test

⁴ Analysis excludes “Not applicable”, “unsure” responses for the question about parental acceptance of LGBT identity