Health Aging and Sexuality in Marginalized Communities LGBT

Health, Aging and Sexuality in Marginalized Communities: LGBT Older Adults Emerging from the Margins Karen I. Fredriksen-Goldsen, Ph. D University of Washington (Funded in part by NIH/NIA, 1 R 01 AG 026526 -01 A 2; 2 R 01 AG 026526 -03 A 1) CCBAR Annual Meeting ▪ University of Chicago ▪ October 17, 2013

Marginalization, Health & Aging Health and Aging in Marginalized Communities HIV and Medication Adherence in China Healthy Hearts in Tulalip Native Community National Health, Aging and Sexuality: Caring and Aging with Pride Over Time

Diversity in Aging Global worldwide aging By 2050, 130 million 50 and older (U. S. Census, 2012) 42% people of color (Vincent & Velkoff, 2010) Two million LGBT adults, age 50 and older Increasing number of diverse LGBT older adults

Research Gaps Sexual orientation key gap in health research (NIH, 2012; CDC, 2011) LGBT people at-risk and underserved (Institute of Medicine, 2011) First time in Healthy People 2020 (DHHS, 2011) Rapidly changing social context, policies and laws

Closing the Gap Behavioral Risk Factor Surveillance System (BRFSSWA) Caring and Aging with Pride: Community-based 2, 560 LGBT older adults, age 50 to 95 Testing of Complex Social Network Driven Sampling Continuation: Longitudinal Study

Figure 1. Non-response: Sexual Identity Question Over Time Trends in Rates of “Refuse to Answer” on Sexual Orientation by Age: Washington State Behavioral Risk Factor Surveillance System (BRFSS-WA), 2003 -2010

Table 1. Non-response to Sexual Orientation Questions by Race and Ethnicity Responses to sexual orientation question Other Not sure/Don’t know Refused to answer AOR(95% CI) Non-Hispanic White 1. 00 (ref) African American 1. 28 (0. 41, 3. 99) 2. 63** (1. 36, 5. 11) 1. 67 (0. 99, 2. 78) Asian American 1. 67 (0. 53, 5. 33) 12. 50*** (8. 68, 18. 02) 4. 42*** (3. 15, 6. 20) American Indian or Alaskan Native 0. 92 (0. 32, 2. 65) 1. 02 (0. 48, 2. 16) 1. 34 (0. 79, 2. 25) Hispanic 0. 63 (0. 30, 1. 35) 6. 43*** (4. 93, 8. 39) 2. 02*** (1. 47, 2. 78) Note. ref=the reference group; AOR=adjusted odds ratio; CI=confidence interval; those who selfidentified as “heterosexual or LGB” were treated as the baseline group; the analysis controlled for age, income, education, and year of interview. *P<. 05; **P<. 01; ***P<. 001 Kim. H. -J. & Fredriksen-Goldsen, K. I. (2013). Nonresponse to a Question on Self-Identified Sexual Orientation in a Public Health Survey and its Relationship to Race and Ethnicity. American Journal of Public Health, 103(1), 67 -69. doi: 10. 2105/AJPH. 2012. 300835. NIHMSID: NIHMS 503953

Table 2. BRFSS-WA Health Disparities: Disability and Mental Distress by Sexual Orientation and Gender Women Men Heterosexual Lesbian Bisexual % % AOR Disability 36. 9 44. 3 1. 5*** Frequent Mental Distress 9. 4 15. 9 1. 4* Heterosexual Gay Bisexual % AOR 34. 0 38. 3 1. 26* 6. 9 13. 1 1. 8** AOR = Adjusted Odds Ratio. Reference Group = Heterosexuals *p <. 05. **p <. 01. ***p <. 001 Data source: Washington State Behavioral Risk Factor Surveillance System, 2003 -2010. Fredriksen-Goldsen K. I. , Kim, H. -J. , Barkan, S. E. , Muraco, A. , & Hoy-Ellis, C. P. (2013). Health Disparities among Lesbian, Gay, and Bisexual Older Adults: Results from a Population-Based Study, American Journal of Public Health, 103(10), 1802 -1809. doi: 10. 2105/AJPH. 2012. 301110

Health Disparities: Distinct Risks ► Lesbians and bisexual women: CVD risk and obesity Gay and bisexual men: Poor physical health and living alone “LGBT” is often used in research and services yet they are distinct groups with specific needs

Figure 2. Health Equity Model Social Positions (intersectionality) Multi-level context Structural levels (social exclusion, institutional heterosexism) Individual levels (microaggressions, discrimination, victimization) Health Adverse and Health. Promoting Pathways Psychosocial (distinct social relations, social support, social network, LGBT community integration) Behavioral (exercise, diet, preventative care, sexual behavior, smoking) Biological (higher cortisol levels, allostatic load) Life course (risks and opportunities) Physical (physical healthrelated quality of life, HIV, obesity, cancer, CVD, disability) Mental (mental healthrelated quality of life, anxiety, depression, suicidal ideation)

Life Course Perspective Social context Cultural meaning Structural location Pre-and Post-Stonewall Silent Generation Baby-Boomer

Research Hypotheses Based on the Health Equity Development model: ► We hypothesize that when controlling for social positions, elevated degrees of discrimination, stigma and non-disclosure of sexual orientation will be significantly associated with lower levels of physical and mental health related quality of life. ► We hypothesize that the configuration of structural, psychological, and behavioral explanatory factors predicting physical and mental health related quality of life, including degrees of discrimination, stigma and non-disclosure of sexual orientation, will be dissimilar for three age groups (age 50 -64, age 65 -79, and age 80 and older) of LGBT midlife and older adults.

CAP Survey: Risk & Protective Factors Community Participatory Integrated Research (CPIR) Eleven community partners West, central and east regions Reach across 48 states 2, 560 LGBT older adults, age 50 to 95 Response rate: 63% Service users: 28%

Table 3. Measures PHQOL & MHQOL Lifetime Discrimination Measured using the SF-8 Health Survey (Ware et al. , 2001), which consists of two components: Physical and mental quality of life. The summary scores were standardized. Cronbach’s α for PHQOL=. 888; Cronbach’s α for MHQOL=. 860 A 16 -item measure based on the Lifetime Victimization Scale (D'Augelli & Grossman, 2001) and Discrimination Scale (Consortium for Political/Social Research, 2010). The summed score range 0 to 48. Cronbach’s α =. 865 Measured using modified items from the Outness Inventory Scale (Mohr & Disclosure Fassinger, 2000). The summary scores ranges 1 to 4 with higher scores indicating higher degree of disclosure. Cronbach’s α =. 917 Length of being The difference between age of first sexual orientation disclosure – age of first sexual orientation awareness was divided by age (D'Augelli & Grossman, 2001) closeted Assessed by a five-item scale based on the Homosexual Stigma Scale (Liu, Feng, Internalized & Rhodes, 2009). Summary scores range from 1 to 4, with higher scores stigma indicating higher degrees of internalized stigma. Cronbach’s α =. 784 Social support The 4 -item social support scale (Sherbourne & Stewart, 1991). Summary scores range from 1 to 4, with higher scores indicating higher degrees of social support. Cronbach’s α =. 851 Social network How many friends, family members, colleagues, and neighbors they interact with in a typical month? The total size of social network was calculated and size summarized by quartiles, with 1 indicating small social network (bottom 25%) and 4 indicating very large social network (top 25%).

Table 3. Measures (cont) Collective Integration Lack of physical activities Substance use Regular checkup Chronic conditions Background characteristics Measured to what extent participants have positive feeling of belonging to LGBT communities. The range is 1 to 4 with higher values indicating better feeling. Assessed by whether or not participants engaged in moderate activities for at least 10 minutes at a time, such as brisk walking, bicycling, vacuuming, gardening, or anything else that causes some increase in breathing or heart rate in a usual week (CDC, 2012). Current smoking (having ever smoked 100 or more cigarettes and currently smoke every day or some days), excessive drinking (having five or more drinks on one occasion during the past 30 days), or drug use (CDC, 2011; NIAAA, 2004) Assessed by asking participants whether or not they had a routine checkup within the past year (CDC) Participants were asked whether they had ever been told by a doctor that they had the following conditions: high blood pressure, high cholesterol, heart attack, angina, stroke, cancer, arthritis, diabetes, asthma, or HIV/AIDS. The number of chronic health conditions was summed, with a range of 0 to 10. Standardized measures were used to assess background characteristics, including sexual orientation (0 = bisexual, 1 = gay or lesbian), gender (0 = men, 1 = women), gender identity (0=non-transgender, 1=transgender), race/ethnicity (0 = non-Hispanic White, 1 = African Americans, 2= Hispanics, 3 = others), income (0 = above 200% of the federal poverty level [FPL], 1 = at or below 200% FPL), unemployment, education (0 = some college or more, 1 = high school or less), relationships status (0 = married or partnered, 1 = other), and residency (0=rural, 1=urban).

Table 4. Background Characteristics All Sample Age 50 -64 Age 65 -79 Age 80 + p Lesbian/Gay (vs. Bisexual) 92. 94% 91. 37% 94. 64% 91. 90% 0. 008 Female 37. 15% 44. 78% 32. 99% 22. 35% 0. 000 Transgender 6. 83% 10. 51% 4. 23% 2. 35% 0. 000 White 86. 50% 84. 22% 87. 14% 93. 70% 0. 006 Black 3. 50% 4. 34% 3. 11% 1. 57% Hispanic 5. 59% 6. 12% 5. 78% 2. 36% Others 4. 41% 5. 32% 3. 97% 2. 36% Poverty 30. 72% 26. 76% 32. 97% 38. 77% 0. 000 Unemployment 56. 11% 34. 77% 70. 34% 86. 22% 0. 000 High school or less 7. 93% 6. 98% 8. 60% 9. 06% 0. 268 Partnered/married 44. 32% 48. 19% 43. 45% 30. 95% 0. 000 Urban 96. 63% 96. 44% 96. 60% 97. 61% 0. 712 1. 955 (. 029) 1. 666 (. 042) 2. 165 (. 042) 2. 286 (. 090) . 000 Race # of Chronic Conditions

Table 5. Explanatory and Outcome Variables Explanatory Variables Lifetime Discrimination Disclosure Closeted Stigma Social Support Collective Integration Network Size, Small Medium Large Very large Lack of Physical Activity Substance Use Regular Health Check # of Chronic Conditions Outcome Variables Physical Quality of Life Mental Quality of Life All Sample M(SE) or % PHRQOL β (SE) MRQOL β (SE) 6. 518 (. 150) 3. 481 (. 012). 103 (. 004) 1. 470 (. 011) 3. 092 (. 017) 3. 416 (. 017) 24. 44% 25. 25% 25. 29% 25. 02% 15. 06% 25. 05% 82. 32% 1. 955 (. 029) -. 028 (. 003)**. 052 (. 030)† -. 030 (. 110) -. 169 (. 037)**. 222 (. 026)**. 094 (. 028)** Ref. 134 (. 056)*. 306 (. 060)**. 276 (. 059)** -. 659 (. 057)** -. 084 (. 047)†. 011 (. 053) -. 280 (. 012)** -. 034 (. 003)**. 082 (. 031)** -. 282 (. 121)* -. 341 (. 037)**. 449 (. 027)**. 180 (. 029)** Ref. 293 (. 059)**. 461 (. 062)**. 580 (. 059)** -. 602 (. 061)** -. 249 (. 049)**. 163 (. 051)** -. 180 (. 013)* . 001. 266. 015. 000 (. 020) --- --- p . 002. 000. 290. 000. 00

Table 6. Explanatory and Outcome Variables Explanatory Variables Lifetime Discrimination Disclosure Closeted Stigma Social Support Collective Integration Network Size, Small Medium Large Very large Lack of Physical Activity Substance Use Regular Health Check # of Chronic Conditions Outcome Variables Physical Quality of Life Mental Quality of Life All Sample M(SE) or % Age 50 -64 M(SE) or % Age 65 -79 M(SE) or % Age 80 + M(SE) or % 6. 518 (. 150) 3. 481 (. 012). 103 (. 004) 1. 470 (. 011) 3. 092 (. 017) 3. 416 (. 017) 24. 44% 25. 25% 25. 29% 25. 02% 15. 06% 25. 05% 82. 32% 1. 955 (. 029) 7. 637 (. 225) 3. 611 (. 016). 097 (. 005) 1. 463 (. 018) 3. 113 (. 025) 3. 451 (. 024) 22. 95% 24. 32% 25. 29% 27. 44% 13. 82% 31. 64% 76. 18% 1. 666 (. 042) 6. 051 (. 216) 3. 430 (. 019). 105 (. 006) 1. 456 (. 017) 3. 089 (. 023) 3. 418 (. 024) 24. 28% 25. 87% 26. 46% 23. 39% 14. 06% 21. 05% 87. 09% 2. 165 (. 042) 3. 592 (. 293) 3. 123 (. 047). 122 (. 014) 1. 569 (. 039) 3. 007 (. 047) 3. 241 (. 050) 32. 99% 26. 90% 19. 29% 20. 81% 25. 30% 12. 61% 87. 90% 2. 286 (. 090) . 000. 178. 023. 118. 001. 015 . 000 (. 020) . 059 (. 029) -. 043 (. 030) -. 004 (. 030). 054 (. 030) -. 250 (. 060) -. 057 (. 051) . 000. 029 p . 000

Table 7. Model Fitting for Physical HRQOL Age 50 -64 β (SE) Demographic Factors Female Poverty Unemployment Partnered LGB-Related Experiences Lifetime discrimination Network size Medium vs. Small Large vs. Small Very large vs. Small Health Related Conditions Lack of Physical Activity # of Chronic Conditions Physical Quality of Life Age 65 -79 β (SE) Age 80+ β (SE) -0. 319** -0. 244** -0. 378** 0. 060 (0. 057) (0. 073) (0. 062) (0. 064) -0. 362** -0. 164* -0. 121* 0. 033 (0. 063) (0. 066) (0. 062) (0. 066) -0. 135 -0. 438** -0. 266 -0. 310** (0. 151) (0. 157) (0. 204) (0. 148) -0. 021** (0. 004) -0. 016** (0. 004) -0. 045** (0. 015) -0. 013 0. 139 0. 059 (0. 081) (0. 083) (0. 082) 0. 118 0. 203* 0. 194* (0. 084) (0. 085) (0. 091) 0. 006 0. 094 0. 236 (0. 202) (0. 233) (0. 215) -0. 431** -0. 207** (0. 081) (0. 021) -0. 357** -0. 238** (0. 083) (0. 020) -0. 546** -0. 192** (0. 178) (0. 049)

Figure 3. Lifetime Discrimination and PHQOL by Age Groups Slope Comparisons Wald Test p value Between Group 1 & 2 χ2(1) = 0. 71 0. 3985 Between Group 1 & 3 χ2(1) = 3. 01 0. 0828 Between Group 2 & 3 χ2(1) = 4. 41 0. 0358

Table 8. Model Fitting for Mental HRQOL Mental Quality of Life Age 65 -79 Age 50 -64 Demographic Factors Female Race Black vs. White Hispanic vs. White Others vs. White Poverty Unemployment Partnered LGB-Related Experiences Lifetime discrimination Stigma Social Support Network size Medium vs. Small Large vs. Small Very large vs. Small Collective Integration Health Related Conditions Lack of Physical Activity Substance Use Regular Health Check # of Chronic Conditions Age 80+ β (SE) -0. 206** (0. 057) -0. 258** (0. 061) -0. 006 (0. 154) -0. 006 -0. 064 -0. 154 -0. 366** -0. 341** -0. 004 (0. 134) (0. 119) (0. 121) (0. 073) (0. 062) (0. 064) 0. 224 -0. 232* -0. 139 -0. 188** -0. 057 0. 047 (0. 168) (0. 115) (0. 156) (0. 063) (0. 060) (0. 064) 0. 661 -0. 101 0. 397 -0. 476** -0. 164 -0. 254* (0. 502) (0. 663) (0. 577) (0. 154) (0. 185) (0. 131) -0. 024** -0. 109* 0. 317** (0. 004) (0. 056) (0. 044) -0. 012** -0. 111 0. 249** (0. 004) (0. 059) (0. 045) -0. 034* -0. 120 0. 244 (0. 014) (0. 128) 0. 163* 0. 105 0. 307** 0. 029 (0. 081) (0. 083) (0. 082) (0. 041) 0. 143 0. 268** 0. 358** 0. 082* (0. 082) (0. 088) (0. 040) 0. 159 0. 176 0. 456* 0. 070 (0. 195) (0. 224) (0. 227) (0. 139) -0. 469** -0. 256** -0. 036 -0. 100** (0. 081) (0. 059) (0. 064) (0. 021) -0. 295** -0. 142* 0. 172* -0. 150** (0. 080) (0. 068) (0. 084) (0. 019) -0. 474** 0. 087 0. 300 -0. 066 (0. 181) (0. 256) (0. 283) (0. 049)

Figure 4. Lifetime Discrimination and MHQOL by Age Groups Slope Comparisons Wald Test p value Between Group 1 & 2 χ2(1) = 4. 12 0. 0423 Between Group 1 & 3 χ2(1) = 0. 77 0. 3814 Between Group 2 & 3 χ2(1) = 3. 531 0. 0602

Summary of Key Findings ► ► ► Identifies potentially modifiable factors associated with physical and mental health related quality of life of LGBT mid-life and older adults. Lifetime discrimination, lack of physical activities and poverty are common correlates of physical and mental health quality of life of LGBT adults, midlife and older. Social positions, structural and psychosocial processes and health behaviors differ by age groups. Even among the common correlates, the degree of influence may differ by age groups. The degree of lifetime discrimination is lower for older age groups, while the degree of internalized stigma is higher and the degree of sexual orientation disclosure is lower for older age groups. Non-disclosure for LGBT older adults may reduce risk of discrimination. Younger age group has higher degrees of discrimination, lower stigma and more disclosure. Limitations of study: cross-sectional, not generalizable, participants connect to community agencies, self-report measures.

Moving Forward National Health, Aging and Sexuality Study: Caring and Aging with Pride Over Time Next phase: LGBT mid-life and older adults, over time in order to test theoretically specified model to understand the temporal relationships that may be amenable to change through targeted interventions. Participants, 50 and older, 3 points in time

Cohorts Baby Boom Generation (born between 1947 -1963) Silent Generation (born before 1947) Cohort differences and changing social context Multiple birth cohort design Analysis of cohort effects from age effects

Sample Hard to reach communities Some subgroups hidden within hidden populations Goal: Obtain a demographically diverse sample of LGBT older adults Ensure coverage of the heterogeneous nature of the populations Address noncoverage, overrepresentation, and other selection biases

Longitudinal Model Age effect, Cohort effect, Social positions (Gender and Race/Ethnicity)

Biological Measures ► Investigate link of poor physical health via allostatic load (AL), a physiological stress-related mechanism linking the psychosocial environment to physiological dysregulations ► AL measures: waist-to-hip ratio, blood pressure, cortisol, DHEA-S, total cholesterol, HDL cholesterol, hemoglobin A 1 c (blood sugar), and Creactive protein. ► Non-invasive dried blood spots (DBS) ► Hypothesis: Controlling for lifetime victimization and other confounding variables, changes in physiological response to stress and health behaviors will partially mediate the effect of change in discrimination on subsequent health and QOL

Next Steps Discussion – How can we maximize the use of biomeasures in this study and obtain quality information given limited resources?