1. Community Based Research Centre
For Gay Men’s Health
Pride, Prejudice &
Determinants of Health
Gay Health @ Work
Terry Trussler EdD
2. Why Gay Health @ Work?
• Social location of “gender order”
• Workplace prejudice lore
• Population Health Policy
• Whitehall Study-social determinants
gradient
social status health status
3. Supporting Theories
• Population Health Marmot et al.
• Minority Stress Meyer et al.
• Intersectionality Hankivsky et al.
• Syndemic Production Stall et al.
4. Sex Now 2011
• September 2011- February 2012
• Men seeking men online
• n=8,607 participants
• All provinces/territories FSA
• Average age: 43
• Range: 13-84
9. Minority Stress
Meyer et al. (2010)
Social Disparity Hypothesis
“disadvantaged fare worse”
Analytical Strategy
1. Within group analysis
2. Between group analysis
3. Outcome analysis
10. Measurement Challenges
• Event recall
• Nonevents potent but difficult
• Covert discrimination
• Systemic prejudice
• Accepting professions
• Living “the life”
32. Axes of Social (dis)Advantage
• Gender (in-gender status)
• Age
• Ethnicity
• Education
• Income
• Sexual Orientation
33. Axes of social (dis)advantage
employment discrimination vs no discrimination
significance odds ratio confidence
Age < 30 <.001 1.783 1.488-2.136
Age 30-44 <.001 1.786 1.494-2.092
Higher Ed .002 1.224 1.079-1.390
Income <$50K <.001 1.311 1.154-1.491
Gay Orientation <.001 4.408 3.738-5.198
34. Stress Factors
• Orientation Disclosure (Out/Not)
• Body Type (BMI, self image)
• Prior prejudice (bashed, bullied)
• Gender presentation (piercings, tattoos)
• Work environment (supportive, cover)
35. Mediating (stress) Factors
employment discrimination vs no discrimination
significance odds ratio confidence
Out @ work <.001 2.845 2.384-3.395
Cover <.001 1.881 1.650-2.144
Unsupportive <.001 1.515 1.299-1.767
Body Type BMI .995 1.002 .588-1.707
36. Health Outcomes
• Suicidality (thoughts & attempts)
• Anxiety + Rx (care seeking)
• Depression + Rx (care seeking)
• Sad, lonely (design questions)
• Sexual risk (UAI)
• STI (last 12 months)
37. Health Outcomes
employment discrimination vs no discrimination
significance odds ratio confidence
Hurt Career <.001 7.310 6.098-8.763
Suicidal 23% <.001 1.872 1.576-2.223
Depression .123 .904 .796-1.028
Sexual Risk .002 1.252 1.095-1.431
Gonorrhea <.001 1.635 1.210-2.210
38. Intersectional Difference
employment discrimination vs no discrimination
Gay Bisexual
BMI II .810 .008
Suicidal <.001 .095
Hurt <.001 <.001
UAI <.001 .755
Gonorrhea .010 .466
41. Intersections
• Age
• Income, education/class
• Ethnicities/race
• HIV status
42. Limitations
• Cross sectional survey
• No population sample (structural)
• Internet sampling
• Self report
• Unmeasurables
• Good Questions
43. Acknowledgements
Vancouver Foundation
Public Health Agency of Canada
Province of British Columbia
44. Acknowledgements
CBRC
Olivier Ferlatte, Travis Salway Hottes, Rick Marchand,
Craig Phillips, David Ham & the CBRC Board
Investigaytors
Alex Chen, Joshun Dulai, Darren Ho, Trevor Hodges, David Le,
Daniel McGraw, Keith Reynolds, Jordan Sang, Jaedyn Starr
Dialogue
Moffat Clarke, Jeff Dodds, Mark Gilbert, Olena Hankivsky,
Ilan Meyer, aaron poirier… and many more
Notas del editor
As a student I never once thought I would one day become a gay men’s health researcher. But today we are seeing a posse of young men at this conference who clearly have that possibility ahead of them. That’s just one small indicator of social movement around the acceptance of gay men in the last few years. But this warming trend exists in an apparent paradox. We have pockets of acceptance and a backdrop of hostility at the same time. Many of the old values are still present practically everywhere but expressed in sometimes deliberately covert ways…
We investigated the relationship between employment experience and health outcomes in pursuit of our overall investigation of social determinants affecting gay men’s health. The workplace is a key social setting shaping most people’s lives. For this very reason, the workplace was the setting of the original research from which Population Health theory emerged: the organizing framework of Canada’s health policies. Michael Marmot’s “Whitehall Study” exposed the social determinedness of health as a gradient defined by social status at work. Higher status, better health.
The results of the research to be presented here are grounded in supporting theories woven together for their capacities to account for, interpret and describe the population health effects of gay men’s workplace experience.
Sex Now 2011 was Canada’s first survey on social determinants of health in gay men. Survey recruitment and participation was unprecedented, the largest survey of Canadian gay and bisexual men to date (an increase of nearly 10% over our pilot study in 2010).
The working corps of the survey was our team of young “Investigaytors” – CBRC’s latest community program development project. Without any previous experience, these young men 19-22, learned research basics first-hand as we navigated the stages of a survey project together. In the beginning all these young would-be investigaytors had was curiosity, but 18 months later they have most of the skills needed to run a survey. And analyze it. And interpret it.
Recruitment on a survey of this size doesn’t just happen. It requires a plan, a promotional campaign and recruitment drive. Our young Investigaytors played a central role in connecting with gay men’s groups and networks across Canada to stimulate participation. We exploited Canada’s lack of census data on gay men as promotional gimmick.
The survey’s recruitment achieved widespread coverage of Canada, including all provinces and territories and a diverse mix of urban, suburban and rural men.
To make sense of social determinants in a survey of this scope requires an organizing model. We’ve been working with Ilan Meyer’s Minority Stress Model shown here –a dynamic model that illustrates how minority groups can be caught in conditions that sustain their social disadvantage.
The practical advantage of the Minority Stress model is that it helps to systematicallysort through data to expose the origins of health problems in social conditions. The basic hypothesis is that the socially disadvantaged ultimately fare worse: sharing an unequal burden of harmful exposures and stress produces an unequal burden of mental disorders. The model also maps out an analytical plan to account for this hypothesis in a given set of data.
Ilan Meyer outlined some of the complications with measuring the health effects of stigma and prejudice at Summit 2010. People remember events that didn’t happen and forget those that did. Oppressive atmospheres can have potent effects on people but are difficult to name, let alone calibrate. As gay discrimination becomes less politically correct it goes behind the scenes. Some gay prejudice is structural like absence of accommodation and in some settings enforced (Boy Scouts of America). Some men may plan their careers (i.e., interior design) specifically to avoid gender confrontation. And some gay men may have chosen to abandon conventional careers altogether. All of this affects the kind of questions we can ask and so the answers we can expect to get.
Within group analysis looks at the whole population and divisions in experience…
For example, a majority of men in the survey said that they think of their workplaces as supportive of their sexuality, but 2 in 5 disagreed or weren’t sure.
Yet, almost as many felt that privacy about their sexuality was important or very important in the workplace.
The population was a bit divided on the question of how much sexuality played a role in defining one’s career.
As many felt that sexuality had hurt their career as helped. Another group reported having had both experiences. But a large majority felt sexuality had little affect on their careers at all.
The largest group were keeping their sexuality a secret but another large group seemed to be out to everyone in their environment.
We used a high standard of job discrimination: not just workplace sexism but real consequences to one’s job or advancement. The overall rate of 17% seems at first to be a small minority – that is until we see differences among groups in the sample…
The idea of between group analysis is to uncover inequities that remain concealed by the within group analysis.
Michael Marmot led the Whitehall study of employees in the British bureaucracy of the 1960’s (replicated in the 90’s and ongoing). The results were similar for chronic heart disease and mental distress. A gradient was apparent according to job status. Executives having high control over their own work had the lowest rates of heart disease and mental distress, line staff had the highest rates and middle management somewhere between. The question for us was, would we see such a pattern in what we learned about gay men in the workplace from Sex Now 2011?
We need comparison groups to conduct between group analysis. Intersectionality theorizes that people have multiple identities tuned to different social locations in their everyday life (work, family, church, club etc.), so analytical categories are always somewhat suspect as they tend to conceal as much as they reveal. Non gay men who have sex with men while married or partnered with a woman, the MSM category, exemplifies Intersectionality’s critique of ”fixed” identities. Using a set of analytical questions that we designed into the survey we found that “men who have with sex with men while married to woman (MSM)” as a group is constructed from at least 3 orientation identities. Each of “gay” and “bisexual” have their own variations and anomalies (gay man married to a lesbian partner having sex with other men, bisexual man partnered with a man having sex with women). So we’ve approached this analysis with three consciously constructed categories to explore between group disparities that might exist in workplace experience.
We noticed this basic inequity (among many others) between gay men and married MSM in the pilot study SN 2010. But we didn’t have our in-gender analysis worked out in the pilot so we lacked some critical data.
Here we apply what we learned in SN 2011 with a battery of questions to assess sexual orientation, attraction and gender identity. We see a gradient and it looks a lot like the Whitehall gradient. We saw the same disparity between gay men and married MSM in 2010 but here we have a middle ground occupied by bisexual men without consideration of who they may be partnered with.
We don’t see the same gradient pattern with education, but as shown in SN 2010, on average, gay men were more highly educated than bisexual men or MSM.And, on average, they also earned substantially less annual income with their education than bisexual or married MSM.
Interestingly, on the issue of privacy at work we see a slightly skewed Whitehall gradient show up again -- defined by the need to protect sexual privacy. Foucault notes in the History of Sexuality that as far back as Greek classical times, higher social rankings had increasingly greater need for sexual discretion to maintain their power image. Plato references homoeroticism specifically in Symposium. The point is made statistically in this chart – exposing social rank by desire for discretion.
A recent Angus Reid poll of 983 LGBT Canadians found that 93% considered their workplace “tolerant”. We asked whether participants felt their workplace to be “supportive” of their sexuality – a different kind of question. Here we see that gay men were far more likely than either bisexual men or married MSM to perceive that their work environment is supportive of their sexuality. We noticed elsewhere in the survey that bisexual men and married MSM were also more likely to perceive their neighborhood as dangerous and more worried about violence than gay men, yet it was gay men who were the main targets of violence. Might misapprehension also be true of the workplace?
Here we see the gradient defining the effect that sexuality has had in channeling career toward accepting environments or in avoidance of hostile ones. You begin here to see that there may be a social disadvantage for gay men in the workplace (Minority Stress Model) when seen relative to other men who have sex with men but who conceal their sexuality in the work environment.
We used this question analytically to assess how much “gender cover” was required in workplaces and because it seemed such a common experience among gay men who work in office environments. Again we see a social gradient based on “need to conceal” in reverse of “desire for privacy”.
This takes us back to being “out” at work – other people in the workplace knowing your sexuality. This chart exposes the very large cultural distance among men who have sex with men based on identity disclosure. The majority of gay men are out at work because it is a cultural imperative among gay men to be out everywhere, a social attitude that barely exists among bisexual men.
But this is where reality bites back. This chart is showing us something even more dramatic than the Whitehall gradient. We see here that at least 1 in 5 gay men (nearly 1 in 4) have had a serious problem with discrimination over their sexuality in the workplace compared to less than 1 in 20 bisexual men or 1 in 45 married MSM.
Here again we see the gradient show up as a disparity among men who feel their career was damaged by discrimination in the workplace.
So… what happens to men who suffer discrimination at work due to their sexuality? Is there a health impact that can be measured? That’s what we want to know in Outcome Analysis. We will use Logistic Regression to model the multiple factors that may be involved in answering this question.
First we will look at socio-demographic features often called “controls” in social analysis. Intersectionality theorizes that these are much more than just controls, they are intersecting axes defining social advantage and disadvantage. Intersectionality’s critique of Logistic Regression as it is used in epidemiology and the social sciences includes which items should be on this list. Sexual orientation for example is often left out of social epidemiology – a major fault of the Population Health model. But gender has to be considered as well, even in a one gender study, because, as Intersectionality theorizes, significant in-gender status differences would otherwise remain concealed, masking relevant social inequities that may be contributing to health outcomes.
Here we are looking at the results of the socio-demographic block of the regression model. Considering all social factors together this table is showing who has more than average disadvantage with employment discrimination and to what degree. The odds are greater for younger men who have higher education and incomes less than $50K and elevated for gay men (as we might have expected from the between group analysis). In other words gay men have more than 4 times greater odds than the other sexual identities of job discrimination.
This phase of the regression model explores the contribution of various socio environmental factors to minority stress (Meyer’s model).
In this block, we examine socio environmental factors associated with employment discrimination (with controls). The results showed that being “out” to anyone at work was the strongest contributing factor. Out men had nearly 3 times the odds of employment discrimination as those undisclosed. Unsupportive environments such as those where gay men need the cover of woman to pass also contributed significantly. Body Type contributed no more than average to employment discrimination similar to many other variables we tested in the model. We had expected to see gender performance measures stand out: tattoos and piercings or prior victimization, bullying or harassment. But it was simply being “out” and therefore “targetable” in the workplace that made the strongest contribution.
The third block of the regression model considers health outcomes (mental disorders in Meyer’s model). We had already explored the relationship between mental and sexual health in the pilot study with variables for suicidallity, depression care seeking, and measures of sadness and loneliness. To these we added care seeking for anxiety and whether medications were prescribed for either anxiety or depression. We also investigated reported sexual risk and diagnosed STI.
It should not have been too surprising to see that those who had faced employment discrimination ended up with such high odds of perceived career damage over those who never had to endure such an experience. There were strong associations with both suicidal thinking and attempts. In fact, 23% of the men who faced employment discrimination reported attempting suicide. An important observation by our epidemiologist here, we miss counting successful suicides in this model: these are survivors. Depression and anxiety care seeking and prescribed medications appeared to be no more significant than average. The low significance of depression was somewhat unexpected as we had found a strong association between bullying and depression in the pilot study. We were also not expecting to find increased odds of sexual risk and STI in the results. People who counsel gay men have observed that increased sex seeking is a more common coping strategy for emotional upheaval than psychological care seeking. We may be seeing this exemplified in this regression table.
Most studies in most of the relevant journals would report the findings we have discussed thus far and therefore base their interpretation and discussion on the preceding results.: “young gay men, who are ‘out’ in the workplace have high odds of career damage and suicidaltiy”. Which leads to another point of contention for Intersectionality. Since so much depends on the axes of social (dis)advantage, Intersectionality questions what may be concealed within a “main effects” model. One way to overcome this is to return to the socio-demographic controls and re-examine the main effects model, comparing between groups. In this case we have returned to sexual orientation and divided the model to look at gay and bisexual men comparatively. Interestingly, as predicted, significant detail for both gay and bisexual men was concealed by the “main effects” model. Here we see that both sexual risk and suicidality were more significant for gay men than bisexual. But having an athletic body type appears to have been a potential contributor to bisexual men being targeted, though no more than average among gay men. We could take this further and look at other socio-demographic factors but the main point to see here is that Intersectionality is quite clearly giving us critical information that would otherwise be missed. *Note data truncated in this slide.
Returning to “gender” in the main effects model can’t be accomplished in the same straight forward way since we have an all male sample. Instead it is more important to assess status differentials within gender. As shown in the between group analysis there is a strong pattern of inequity throughout most measures that consistently affords gay men the least social status and highest vulnerability or risk. The pattern is so consistent that it exposes a hidden structural reality that is difficult to dismiss…
Which is a real concern when it comes to health care policy and funding. Here we see another gradient, resembling Whitehall, that accounts for the distribution of funding within the Global Program on HIV/ AIDS. Virtually everywhere in the world gay men are the most affected yet least resourced in what appears more and more to be a socially structured and produced epidemic….
We have not fully completed our study until we look into further axes of disadvantage. Some we might explore within our own data concern age, income groups, ethnicities and HIV status. Future studies might also look into the effects of types of employment and the added complications of disability.
Its important to note that we are working with self report survey data not clinical or population data. The strength of our approach is acquiring large national samples with relatively convenient survey questions. Thus the information we get from ordinary gay men across the country helps to theorize relationships that would not be otherwise known – data to compare and corroborate with other sources and surveys.