. 2019 Sep 5;6(6):271–288. doi: 10.1089/lgbt.2018.0204

Table 2.

Summary of Studies Comparing Mental Health Outcomes Among Sexual Minority and Heterosexual Cancer Survivors

Study	Study description	Sample recruitment and characteristics	Study methods and outcomes	Results	Quality assessment^a
Studies conducted with sexual minority women
Arena et al., 2006⁴⁶	Cross-sectional study comparing psychosocial responses to breast cancer treatment between lesbian and heterosexual women	Lesbian women (N = 39) recruited nationwide using flyers distributed via physicians, women's networks, and lesbian community resource centers Mean age: 48.3 (SD = 8.3) Race/ethnicity: Non-Hispanic White: 92.3% Hispanic: 5.1% Black: 2.6% Heterosexual women (N = 39) participating in one of two existing studies recruited from Miami-area hospitals and medical practices Mean age: 49.4 (SD = 8.9) Race/ethnicity: Non-Hispanic White: 74.4% Hispanic: 10.2% Black: 15.4%	Matched lesbian sample to heterosexual sample based on age, ethnicity, stage of disease, and time since diagnosis Compared samples on depression using CES-D scale Conducted analysis of covariance controlling for medical and demographic variables that significantly differed between groups (education and receipt of radiation therapy)	No statistically significant difference in depression symptoms was found between lesbian women (M = 11.9, SD = 2.1) and heterosexual women (M = 15.8, SD = 2.1) when controlling for education and radiation status	Six of six criteria met
Boehmer et al., 2012⁴⁷	Cross-sectional study comparing anxiety and depression between SMW and heterosexual women breast cancer survivors	SMW (N = 69) recruited from Massachusetts Cancer Registry (RSMW) Mean age: 55.9 (SD = 8.3) Race/ethnicity: White: 94.2% SMW (N = 112) convenience sample recruited nationwide via community outreach, internet advertising, and print publications Mean age: 55.1 (SD = 8.7) Race/ethnicity: White: 86.6% Heterosexual women (N = 257) recruited from Massachusetts Cancer Registry (RH) Mean age: 62.7 (SD = 11.0) Race/ethnicity: White: 85.2%	Compared RSMW with RH on depression and anxiety using the HADS; conducted t-tests for each subscale Compared self-reported use of antidepressants and anxiety medications among RSMW and RH using χ² tests Also used stepwise regression to examine the relationship between sexual orientation and HADS subscales for both registry (RSMW vs. RH) and combined samples (RSMW + SMW convenience sample vs. RH), including numerous demographic and clinical variables in the model	For HADS depression subscale, no statistically significant difference was found between RSMW (M = 3.1, SD = 3.2) and RH (M = 3.2, SD = 3.0) For HADS anxiety subscale, no statistically significantly difference was found between RSMW (M = 5.3, SD = 3.7) and RH (M = 5.2, SD = 3.9) Significantly more RSMW reported current use of antidepressant medication compared to RH (40.6% vs. 21.0%) No statistically significant difference was found between RSMW and RH women on current use of anti-anxiety medication (8.7% vs. 7.4%) For stepwise regression analyses, no statistically significant relationship was found between sexual orientation and HADS depression for the registry or combined sample when controlling for age, marital status, education, income, insurance, co-morbidities, stage, receipt of radiation, use of mood stabilizer, discrimination, and four interactions between sexual orientation and stage, radiation, mood stabilizer, and discrimination There was a significant relationship between sexual orientation and HADS anxiety for the combined sample, but not for the registry sample, when controlling for age, marital status, income, employment, neighborhood household income, discrimination, use of mood stabilizer, and two interactions between sexual orientation and marital status and sexual orientation and employment, with the combined SMW sample reporting significantly less anxiety than RH women	Six of six criteria met
Boehmer et al., 2012⁴⁸	Cross-sectional study examining the association between sexual orientation and QOL among SMW and heterosexual women breast cancer survivors	SMW (N = 69) recruited from Massachusetts Cancer Registry (RSMW) Mean age: 55.9 (SD = 8.3) Race/ethnicity: White: 94.2% SMW (N = 112) convenience sample recruited nationwide via community outreach, internet advertising, and print publications Mean age: 55.1 (SD = 8.7) Race/ethnicity: White: 86.6% Heterosexual women (N = 257) recruited from Massachusetts Cancer Registry (RH) Mean age: 62.7 (SD = 11.0) Race/ethnicity: White: 85.2%	Compared RSMW with RH on mental QOL using the Medical Outcomes Study Short Form Health Survey (SF-12) MCS; conducted t-test Used stepwise regression to examine the relationship between sexual orientation and MCS for both registry (RSMW vs. RH) and combined samples (RSMW + convenience sample vs. RH), including age, employment, and income in the model	On the MCS, no statistically significant difference was found between RSMW (M = 48.8 SD = 10.1) and RH (M = 50.3, SD = 10.4) In stepwise regression analyses, sexual orientation was not significantly associated with MCS for either the registry or the combined sample when controlling for age, employment, and income	Six of six criteria met
Jabson and Bowen, 2014⁴⁹	Cross-sectional study comparing perceived stress between SMW and heterosexual breast cancer survivors	SMW (defined as lesbian, bisexual, queer, and women-loving) breast cancer survivors (N = 68) recruited online through breast cancer survivor groups, websites, discussion boards, and organizations; through lesbian-specific health organizations, discussion boards, and newsletters; as well as using hard-copy materials distributed at women's health centers, book stores, cafes, Pride festivals, lesbian health centers and clinics, and community events Mean age: 56; range 38–74 Race/ethnicity^b: White: 60.3% Latina: 2.9% Black/African American: 1.5% Heterosexual female breast cancer survivors (N = 143) recruited online through breast cancer survivor groups, websites, discussion boards, and organizations; as well as using hard-copy materials distributed via women's health centers, book stores, and cafes Mean age: 54; range 30–79 Race/ethnicity^b: White: 60.1% Latina: 2.8% American Indian/Alaska Native: 2.8% Asian American/Pacific Islander: 1.4%	Compared samples on perceived stress using the PSS-4 Conducted unadjusted linear regression between sexual orientation and perceived stress Conducted linear regression between sexual orientation and perceived stress controlling for age (dichotomized into <55 and ≥55 years of age)	There was a statistically significant association between sexual orientation and perceived stress for both the unadjusted model and the model adjusted for age. SMW breast cancer survivors reported more perceived stress (M = 8.5, SD = 1.3) than did heterosexual breast cancer survivors (M = 8.1, SD = 1.5)	Six of six criteria met
Jabson et al., 2011⁵⁰	Cross-sectional study examining the association between QOL and sexual orientation among lesbian and heterosexual breast cancer survivors	Lesbian breast cancer survivors (N = 61) recruited online through breast cancer survivor groups, websites, and discussion boards, social networking sites, regional news outlets, lesbian-health focused newsletters, organizations, and discussion boards, as well as through hard-copy materials distributed at women's health centers and hospitals, book stores, cafes, community message boards, Pride festivals, and community events Mean age: 55.7; range 38–74 Race/ethnicity^b: White: 55.7% Latina: 1.6% Heterosexual female breast cancer survivors (N = 143) recruited online through breast cancer survivor groups, websites, and discussion boards, social networking sites and news outlets, as well as through hard-copy materials distributed at women's health centers and hospitals, book stores, cafes, and community message boards Mean age: 54.1; range 30–79 Race/ethnicity^b: White: 60.1% Latina: 2.8% American Indian/Alaska Native: 2.8% Asian American/Pacific Islander: 1.4%	Compared samples on psychological well-being using Ferrell and Hassey Dow's QOL Cancer Survivors psychological well-being domain subscale Domain scores were compared across groups using t-test	No statistically significant difference in psychological well-being was found between lesbian (M = 6.3, SD = 0.9) and heterosexual (M = 6.4, SD = 0.1) breast cancer survivors	Five of six criteria met (“No” on “Strategies used to address confounders”)
Studies conducted with sexual minority men
Allensworth-Davies, 2012⁵¹	Cross-sectional dissertation study examining QOL among gay men after treatment for prostate cancer, and comparing QOL between gay prostate cancer survivors and a comparison sample	Gay men older than the age of 50 (N = 111) recruited nationwide via LGBT print publications, community newspapers, and word of mouth Age range: 50–64: 41.4% 65–74: 41.4% ≥75: 17.2% Race/ethnicity White: 89.2% Non-White: 7.2% “Talcott sample” (N = 235) used as comparison group; Boston clinic-based sample of prostate cancer survivors assumed to be predominantly heterosexual Age range: 50–64: 27.0% 65–74: 40.8% ≥75: 32.2% Race/ethnicity: White: 99.1%	Compared mental QOL of gay sample to comparison sample using the SF-12 MCS Conducted Kruskal-Wallis one-way analysis of variance	No statistically significant difference in MCS scores was found between gay (M = 48.6, SD = 10.1) and comparison (M = 47.3, SD = 5.9) samples	Four of six criteria met (“No” on “Inclusion criteria clearly defined” and “Strategies used to address confounders”)
Hart et al., 2014⁵²	Cross-sectional study comparing health-related QOL and sexual behavior between GB prostate cancer survivors and a normative group of prostate cancer survivors	GB men (N = 92) from the United States and Canada recruited using electronic listservs for prostate cancer survivors, flyers in community centers and support groups, and local media advertising Mean age: 57.8 (SE = 1.0) Race/ethnicity: Caucasian: 91.3% African American: 5.4% Asian: 1.1% Other: 2.2% CaPSURE study cohort (N = 730) used as comparison group; national registry of prostate cancer survivors assumed to be predominantly heterosexual Demographic information not available	Compared GB survivors with published CaPSURE norms on mental QOL using SF-36 MCS Conducted independent samples t-test	GB prostate cancer survivors had significantly lower MCS scores (M = 43.9, SE = 1.4) compared with the CaPSURE study cohort (M = 51.9, SE = 1.4)	Three of six criteria met (“No” on “Subjects and setting described in detail,” “Confounding factors identified,” and “Strategies used to address confounders”)
Kamen et al., 2014⁵³	Cross-sectional study examining prevalence of cancer diagnosis and health risk behavior in a national sample of gay and heterosexual men, including non-survivors, as well as disparities in psychological distress between subsamples of gay and heterosexual cancer survivors	Gay male cancer survivors (N = 53) surveyed in the 2009 BRFSS conducted by the CDC in Arizona, California, Massachusetts, Ohio, and Wisconsin Heterosexual male cancer survivors (N = 1545) surveyed in the 2009 BRFSS conducted in Arizona, California, Massachusetts, Ohio, and Wisconsin Participants were part of a broader study sample, including men who had never been diagnosed with cancer. Demographic information was not provided for the subsample of cancer survivors Cancer types included prostate cancer, colon cancer, bladder cancer, melanoma, and “other cancers”	Psychological distress was compared between the gay and heterosexual subsamples of cancer survivors using a single-item BRFSS measure asking participants to report the number of days over the past 30 days in which they experienced poor mental health Conducted independent samples t-test for this comparison, accounting for sample weights and yielding weighted mean estimates and standard errors	Gay male cancer survivors endorsed significantly more days of poor mental health in the past 30 days (weighted M = 5.3, weighted SE = 9.5) than did heterosexual male cancer survivors (weighted M = 2.4, weighted SE = 6.8)	Two of six criteria met (“No” on “Subjects and setting described in detail,” “Confounding factors identified,” and “Strategies used to address confounders”; “Unclear” on “Outcome measures valid and reliable”)
Ussher et al., 2016⁵⁴	Cross-sectional study comparing GB prostate cancer survivors with heterosexual prostate cancer survivors on health-related QOL, psychological distress, and sexual health	GB men (N = 124) recruited using an information sheet distributed via cancer research volunteer databases, physicians' offices, prostate cancer support groups, GB community organizations, GB social media, and electronic listservs serving prostate cancer survivors Mean age: 64.3 (SD = 8.2) Race/ethnicity: Anglo-Celtic 67.7% Other: 32.3% Heterosexual men (N = 225) recruited using an information sheet distributed via cancer research volunteer databases Mean age: 71.5 (SD = 9.0) Race/ethnicity: Anglo-Celtic: 88.0% Other: 11.1% Sample predominantly recruited from Australia, with a minority recruited from the United States and the United Kingdom	Compared samples using independent-sample t-tests for the following outcomes: Emotional QOL was assessed using the FACT-P Psychological distress was measured by the BSI-18 Cancer-related distress was measured using the MAX-PC	GB prostate cancer survivors endorsed significantly worse emotional well-being on the FACT-P (M = 17.1, SD = 4.5) compared with heterosexual prostate cancer survivors (M = 18.3, SD = 4.0) GB prostate cancer survivors reported significantly more psychological distress on the BSI-18 (M = 10.7, SD = 12.4) than heterosexual prostate cancer survivors (M = 7.0, SD = 10.4) GB men were also significantly more likely to meet BSI-18 criteria for distress (13.7%) than heterosexual men (7.1%) GB men scored significantly higher on the BSI-18 depression subscale (M = 4.7, SD = 5.4) than heterosexual men (M = 2.7, SD = 4.4) GB men scored significantly higher on the BSI-18 anxiety subscale (M = 2.0, SD = 2.6) compared with heterosexual men (M = 1.25, SD = 2.2) GB men endorsed significantly greater cancer-related distress (M = 13.6, SD = 10.6) compared with the heterosexual sample (M = 8.9, SD = 8.9)	Five of six criteria met (“No” on “Strategies used to address confounders”)
Wassersug et al., 2013⁵⁵	Cross-sectional study comparing diagnostic and outcome differences between nonheterosexual prostate cancer survivors and heterosexual prostate cancer survivors after cancer treatment	Nonheterosexual men (N = 96) recruited by distributing study information via websites, e-mail lists, and newsletters of prostate cancer information and support organizations in the United States, Australia, Canada, the United Kingdom, and New Zealand, as well as targeted Facebook advertising Age range: <46 years: 3.0% 46–55 years: 29.0% 55–65 years: 42.0% 66–75 years: 26.0% Country of residence: United States: 83.0% Canada: 9.0% Other: 7.0% Heterosexual men (N = 460) recruited by distributing study information via websites, e-mail lists, and newsletters of prostate cancer information and support organizations in the United States, Australia, Canada, the United Kingdom, and New Zealand, as well as targeted Facebook advertising Age range: <46 years: 2.0% 46–55 years: 23.0% 55–65 years: 43.0% 66–75 years: 26.0% >76 years: 4.0% Country of residence: United States: 59.0% Australia: 21.0% Other: 20.0%	Measured self-reported current antidepressant use (yes/no) as a proxy for mental health/depression Conducted a logistic regression analysis comparing use of antidepressants between nonheterosexual and heterosexual men controlling for demographic variables that differed across groups (country of residence and relationship status)	There was no statistically significant difference found in current antidepressant use between nonheterosexual (15%) and heterosexual (15%) samples when controlling for country of residence and relationship status	Four of six criteria met (“No” on “Inclusion criteria clearly defined” and “Outcome measures valid and reliable”)
Studies conducted with mixed-gender samples
Kamen et al., 2015⁵⁶	Cross-sectional study comparing psychological distress between LGBT and heterosexual cancer survivors	LGBT cancer survivors (N = 207) were recruited online and through LIVESTRONG partner organizations for a national dataset collected online by the LIVESTRONG Foundation Mean age: 47.6 (SE = 11.8) Self-reported sex: Male: 41.5% Female: 58.5% Race/ethnicity: Non-Hispanic White: 87.0% Hispanic/Latino: 5.3% Non-Hispanic Black: 2.9% Asian/Pacific Islander: 0.5% Other: 4.3% Sexual orientation/gender identity: Lesbian/gay: 74.4% Bisexual: 24.6% Transgender: 2.9% Heterosexual cancer survivors (N = 621) were recruited online and through LIVESTRONG partner organizations for a national dataset collected online by the LIVESTRONG Foundation. Survivors were culled from a larger dataset (N = 4889) and selected using propensity matching Mean age: 48.5 (SE = 12.5) Self-reported sex: Male: 40.6% Female: 59.4% Race/ethnicity: Non-Hispanic White: 87.4% Hispanic/Latino: 5.8% Non-Hispanic Black: 1.1% Asian/Pacific Islander: 0.5% Other: 4.5% Most frequently reported cancer types included breast cancer, prostate cancer, colorectal cancer, and testicular cancer	LGBT cancer survivors were matched to heterosexual cancer survivors using a propensity-matching approach. Survivors were matched on age, self-reported sex, race, income, education, relationship status, insurance status, state, type of cancer, age at diagnosis, and time since treatment Using 35 LIVESTRONG dichotomous (yes/no) items about “emotional concerns,” conducted principal component analysis to extract the empirically strongest symptom clusters (eigenvalues ≥2), including a “depression related to cancer” cluster “Depression related to cancer” symptom cluster was compared across samples using Poisson regression analysis to compare count of depression symptoms (ranging from 0 to 6) between LGBT and heterosexual groups Poisson regression analyses comparing depression symptoms between LGBT and heterosexual groups were then stratified by self-reported sex	LGBT identity was significantly associated with greater symptoms of depression, with LGBT cancer survivors reporting more symptoms (M = 2.4, SD = 1.8) than heterosexual survivors (M = 2.0, SD = 1.7) Stratified analyses revealed an interaction between depression symptoms and self-reported sex Among those identifying as male, GBT survivors reported significantly more depressive symptoms (M = 2.2) than heterosexual survivors (M = 1.7) Among those identifying as female, there was no statistically significant difference found between LBT (M = 2.5) and heterosexual (M = 2.4) survivors on count of depressive symptoms	Four of six criteria met (“No” on “Inclusion criteria clearly defined,”; “Unclear” on “Outcome measures valid and reliable”)
Kamen et al., 2016⁵⁷	Pilot randomized controlled trial of a dyadic exercise intervention for psychological distress, partner support, and exercise adherence among lesbian, gay, and heterosexual cancer survivors Analyses were conducted to examine differences in psychological distress between lesbian/gay survivors and heterosexual survivors at baseline	Gay and lesbian cancer survivors (N = 10) referred by nurses and oncologists at Wilmot Cancer Center at the University of Rochester, through word of mouth, and by outreach to local groups serving the lesbian and gay communities Mean age: 54.0 (SE = 4.8) Gender: Male: 50.0% Female: 50.0% Race/ethnicity: Non-Hispanic White: 90.0% Hispanic/Latino: 10.0% Heterosexual cancer survivors (N = 12) referred by nurses and oncologists at Wilmot Cancer Center at the University of Rochester Mean age: 58.0 (SE = 2.3) Gender: Male: 25.0% Female: 75.0% Race/ethnicity: Non-Hispanic White: 100.0% Cancer types included breast, prostate, rectal, testicular, esophageal, ovarian, pancreatic, sinus, stomach, thyroid, and tongue	Compared samples on depression symptoms at baseline using the CES-D Compared samples on anxiety symptoms at baseline using the Spielberger STAI Form Y-1, State Form For each comparison, conducted independent samples t-test	Gay and lesbian cancer survivors reported significantly more depression symptoms (M = 12.4, SE = 3.3) than heterosexual cancer survivors (M = 6.3, SE = 0.8) There was no statistically significant difference in anxiety symptoms between gay and lesbian cancer survivors (M = 30.5, SE = 2.7) and heterosexual cancer survivors (M = 25.8, SE = 1.9)	Four of six criteria met (“No” on “Confounding factors identified” and “Strategies used to address confounders”)

^{^a}

Conducted using the Joanna Briggs Institute Analytical Cross Sectional Studies Critical Appraisal Tool. Ratings for identification and management of confounders, outcome measures, and analytic techniques were assigned at the outcome level rather than the study level. See Table 3 for further details.

^{^b}

These studies had a large proportion of missing data on demographic variables.

BRFSS, Behavioral Risk Factor Surveillance System; BSI-18, Brief Symptom Inventory-18; CaPSURE, Cancer of the Prostate Strategic Urologic Research Endeavor; CDC, Centers for Disease Control and Prevention; CES-D, Center for Epidemiological Studies-Depression; FACT-P, Functional Assessment of Cancer Therapy-Prostate; GB, gay and bisexual; GBT, gay, bisexual, trans; HADS, Hospital Anxiety and Depression Scale; LBT, lesbian, bisexual, trans; LGBT, lesbian, gay, bisexual, or transgender; M, mean; MAX-PC, Memorial Anxiety Scale for Prostate Cancer; MCS, Mental Component Summary; N, number of participants; PSS, Perceived Stress Scale; QOL, quality of life; RH, registry heterosexual; RSMW, registry SMW; SD, standard deviation; SE, standard error; SMW, sexual minority women; STAI, State-Trait Anxiety Inventory.