This project demonstrates applied regression modeling, effect modification assessment, and transparent reporting using large-scale public health survey data.
- Overview: The analysis investigates relationships between key health outcomes and demographic factors, utilizing a subset of BRFSS data with 106,285 observations across 12 variables, including sexual orientation, mental health, and health-related covariates. https://www.cdc.gov/brfss/
- Dataset: A subset of the BRFSS 2018 dataset with 106,285 observations and 12 variables, collected from noninstitutionalized U.S. adults aged 18+ via telephone surveys.
- Variables:
- Exposure: Sexual Orientation (SO) as the primary exposure variable.
- Covariates: Sex, Race, Age Category, Income Category, BMI Category, Presence of Health Plan, and Smoking Status.
- Outcomes: Number of Poor Mental Health Days (MENTHLTH) and Depressive Disorder (ADDEPEV2).
- Data Handling: Missing values were excluded, and categorical variables were formatted for analysis.
- Sexual Orientation: Lesbian/Gay, Straight/not gay, Bisexual, Something else.
-
Poor Mental Health Days: Continuous (0-30 days).
-
Depressive Disorder: Binary (Yes/No).
The following table outlines the key variables used in this analysis, derived from the BRFSS 2018 dataset:
| Variable Name | Description |
|---|---|
| SO | Sexual Orientation (e.g., Straight, Lesbian/Gay, Bisexual, Something Else) |
| SEX1 | Sex (Male, Female) |
| RACE | Race/Ethnicity (e.g., White Non-Hispanic, Black Non-Hispanic, etc.) |
| AGECAT | Age Category (e.g., 18-29, 30-39, 40-59, 60+) |
| INCOMECAT | Income Category (e.g., < $50,000, ≥ $50,000) |
| BMICAT | BMI Category (e.g., Underweight, Normal, Overweight, Obese) |
| MENTHLTH | Number of Poor Mental Health Days |
| ADDEPEV2 | Depressive Disorder (Yes, No) |
| POORHLTH | Number of Poor Physical Health Days |
| GENHLTH | General Health Status (e.g., Excellent, Poor) |
| HLTHPLN1 | Presence of Health Plan (Yes, No) |
| SMOKER | Smoking Status (e.g., Current, Former, Never) |
The primary goals of this analysis include:
- Investigate the correlation between sexual orientation and mental health metrics.
- Measure disparities in mental health across varying sexual orientation categories.
- Explore age as a potential modifier in the sexual orientation-mental health relationship.
- Descriptive Statistics: Means, SDs, frequencies, percentages by sexual orientation (ANOVA, chi-square, α = 0.05).
- Bivariate Analysis: ANOVA/t-tests for mental health days; chi-square for categorical variables.
- Multivariable Analysis: GLM for mental health days; logistic regression for depressive disorder, adjusted for covariates.
- Effect Modification: Stratified analysis by age; tested age-sexual orientation interactions.
- Measures of Association: Effect estimates (95% CI) for continuous; odds ratios (95% CI) for binary outcomes.
- Diagnostics: Checked linearity, normality, homoscedasticity.
| Characteristic | Summary (n=106,285) |
|---|---|
| Sexual Orientation | |
| - Lesbian or Gay | 2,213 (2.08%) |
| - Straight/not gay | 99,446 (93.57%) |
| - Bisexual | 3,280 (3.09%) |
| - Something else | 1,346 (1.27%) |
| Sex | |
| - Male | 44,662 (42.02%) |
| - Female | 61,623 (57.98%) |
| Race | |
| - White, Non-Hispanic | 82,143 (77.29%) |
| - Black, Non-Hispanic | 9,560 (8.99%) |
| - Asian, Non-Hispanic | 2,419 (2.28%) |
| - American Indian/Alaskan Native, Non-Hispanic | 1,725 (1.62%) |
| - Hispanic | 5,967 (5.61%) |
| - Other race, Non-Hispanic | 4,471 (4.21%) |
| Age | |
| - 18-29 | 12,492 (11.75%) |
| - 30-39 | 12,433 (11.70%) |
| - 40-59 | 34,620 (32.57%) |
| - 60 and older | 46,740 (43.98%) |
| Income Category | |
| - Less than $50,000 | 78,257 (73.63%) |
| - $50,000 or more | 28,028 (26.37%) |
| BMI Category | |
| - Underweight | 2,008 (1.89%) |
| - Normal weight | 30,656 (28.84%) |
| - Overweight | 35,022 (32.95%) |
| - Obese | 38,599 (36.32%) |
| Mental Health and Well-being | |
| - Poor mental health days (mean, SD) | 7.081 (9.943) |
| - Depressive Disorder | |
| - Yes | 32,670 (30.74%) |
| - No | 73,615 (69.26%) |
| - Poor health days (mean, SD) | 5.449 (9.462) |
| General Health | |
| - Excellent | 10,210 (9.61%) |
| - Very good | 29,529 (27.78%) |
| - Good | 34,422 (32.39%) |
| - Fair | 21,873 (20.58%) |
| - Poor | 10,251 (9.64%) |
| Health Insurance | |
| - Yes | 98,032 (92.24%) |
| - No | 8,253 (7.76%) |
| Smoking Status | |
| - Current smoker - daily | 14,384 (13.53%) |
| - Current smoker - some days | 5,458 (5.14%) |
| - Former smoker | 31,329 (29.48%) |
| - Never smoked | 55,114 (51.85%) |
The analytic sample comprised 106,285 adults from the 2018 BRFSS, representing a broad demographic range. Most respondents identified as straight or not gay (93.6%), with bisexual (3.1%), lesbian or gay (2.1%), and “something else” (1.3%) groups also represented. Females made up 58% of the sample, and the majority were White, non-Hispanic (77%). Nearly 44% were aged 60 years or older, and about 74% reported annual incomes below $50,000. On average, participants reported 7 poor mental-health days and 5 poor physical-health days in the past month; 31% had been diagnosed with a depressive disorder. Most respondents had health-insurance coverage (92%), and smoking habits were mixed, with 52% never having smoked.
| Characteristic | Straight (n=99,446) | Lesbian/Gay (n=2,213) | Bisexual (n=3,280) | Something else (n=1,346) |
|---|---|---|---|---|
| Sex | ||||
| - Male | 41,846 (42.08%) | 1,263 (57.07%) | 1,008 (30.73%) | 545 (40.49%) |
| - Female | 57,600 (57.92%) | 950 (42.93%) | 2,272 (69.27%) | 801 (59.51%) |
| Race | ||||
| - White, Non-Hispanic | 77,195 (77.63%) | 1,664 (75.19%) | 2,380 (72.56%) | 904 (67.16%) |
| - Black, Non-Hispanic | 9,031 (9.08%) | 172 (7.77%) | 257 (7.84%) | 100 (7.43%) |
| - Asian, Non-Hispanic | 2,261 (2.27%) | 42 (1.90%) | 68 (2.07%) | 48 (3.57%) |
| - American Indian/Alaskan Native, Non-Hispanic | 1,618 (1.63%) | 29 (1.31%) | 50 (1.52%) | 28 (2.08%) |
| - Hispanic | 5,332 (5.36%) | 174 (7.86%) | 299 (9.12%) | 162 (12.04%) |
| - Other race, Non-Hispanic | 4,009 (4.03%) | 132 (5.96%) | 226 (6.89%) | 104 (7.73%) |
| Age Category | ||||
| - 18–29 | 10,414 (10.47%) | 475 (21.46%) | 1,288 (39.27%) | 315 (23.40%) |
| - 30–39 | 11,272 (11.33%) | 317 (14.32%) | 675 (20.58%) | 169 (12.56%) |
| - 40–59 | 32,784 (32.97%) | 765 (34.57%) | 775 (23.63%) | 296 (21.99%) |
| - 60 and older | 44,976 (45.23%) | 656 (29.64%) | 542 (16.52%) | 566 (42.05%) |
| Income Category | ||||
| - Less than $50,000 | 46,513 (53.73%) | 1,084 (53.64%) | 1,751 (63.10%) | 786 (71.00%) |
| - $50,000 or more | 40,060 (46.27%) | 937 (46.36%) | 1,024 (36.90%) | 321 (29.00%) |
| BMI Category | ||||
| - Underweight | 1,810 (1.82%) | 53 (2.39%) | 88 (2.68%) | 57 (4.23%) |
| - Normal weight | 28,444 (28.60%) | 704 (31.81%) | 1,086 (33.11%) | 422 (31.35%) |
| - Overweight | 33,017 (33.20%) | 707 (31.95%) | 881 (26.86%) | 417 (30.98%) |
| - Obese | 36,175 (36.38%) | 749 (33.85%) | 1,225 (37.35%) | 450 (33.43%) |
| Mental Health Metrics | ||||
| - Poor mental health days (mean, SD) | 6.87 (9.85) | 8.61 (10.34) | 11.05 (10.86) | 10.23 (11.31) |
| Depressive Disorder | ||||
| - Yes | 29,404 (29.57%) | 943 (42.61%) | 1,731 (52.77%) | 592 (43.98%) |
| - No | 70,042 (70.43%) | 1,270 (57.39%) | 1,549 (47.23%) | 754 (56.02%) |
| - Poor health days (mean, SD) | 5.39 (9.46) | 5.79 (9.27) | 6.25 (9.21) | 6.90 (10.11) |
| General Health | ||||
| - Excellent | 9,566 (9.62%) | 236 (10.66%) | 295 (8.99%) | 113 (8.40%) |
| - Very good | 27,696 (27.85%) | 624 (28.20%) | 930 (28.35%) | 279 (20.73%) |
| - Good | 32,139 (32.32%) | 745 (33.66%) | 1,097 (33.45%) | 441 (32.76%) |
| - Fair | 20,404 (20.52%) | 445 (20.11%) | 691 (21.07%) | 333 (24.74%) |
| - Poor | 9,641 (9.69%) | 163 (7.37%) | 267 (8.14%) | 180 (13.37%) |
| Health Insurance | ||||
| - Yes | 91,948 (92.46%) | 2,033 (91.87%) | 2,877 (87.71%) | 1,174 (87.22%) |
| - No | 7,498 (7.54%) | 180 (8.13%) | 403 (12.29%) | 172 (12.78%) |
| Smoking Status | ||||
| - Current smoker – daily | 13,201 (13.27%) | 402 (18.17%) | 596 (18.17%) | 185 (13.74%) |
| - Current smoker – some days | 4,943 (4.97%) | 156 (7.05%) | 269 (8.20%) | 90 (6.69%) |
| - Former smoker | 29,594 (29.76%) | 632 (28.56%) | 779 (23.75%) | 324 (24.07%) |
| - Never smoked | 51,708 (52.00%) | 1,023 (46.23%) | 1,636 (49.88%) | 747 (55.50%) |
-
Sex distribution differed across sexual orientation groups, with males comprising 57.07% of lesbian/gay participants compared with 42.08% among straight participants, while females accounted for 69.27% of bisexual participants.
-
Younger age groups were overrepresented among sexual minorities, particularly among bisexual individuals, where 39.27% were aged 18–29, compared with 10.47% among straight participants. Individuals identifying as “something else” also showed a higher concentration in this age group (23.40%).
-
Lower household income was more common among sexual minority groups, with 63.10% of bisexual participants and 71.00% of those identifying as “something else” reporting incomes below $50,000, compared with 53.73% among straight participants.
-
BMI category distributions varied by sexual orientation, with obesity prevalence ranging from 33.43% among those identifying as “something else” to 37.35% among bisexual participants.
-
Depressive disorders were substantially more prevalent among sexual minorities, affecting 52.77% of bisexual individuals, 43.98% of those identifying as “something else,” and 42.61% of lesbian/gay participants, compared with 29.57% among straight participants.
-
Average poor mental health burden was higher among sexual minorities, with bisexual participants reporting a mean of 11.05 days of poor mental health in the past 30 days, compared with 6.87 days among straight participants.
| Characteristic | Measure of Effect (β, 95% CI) | P-value |
|---|---|---|
| Sexual Orientation | ||
| Straight/not gay (ref) | 0.00 | — |
| Lesbian/Gay | 1.1790 (0.7713, 1.5866) | <0.0001 |
| Bisexual | 2.4543 (2.1134, 2.7952) | <0.0001 |
| Something else | 2.6512 (2.1275, 3.1749) | <0.0001 |
| Sex | ||
| Male (ref) | 0.00 | — |
| Female | 1.0825 (0.9624, 1.2026) | <0.0001 |
| Race/Ethnicity | ||
| White, Non-Hispanic (ref) | 0.00 | — |
| Black, Non-Hispanic | −0.3177 (−0.5258, −0.1097) | 0.0028 |
| Asian, Non-Hispanic | −1.2726 (−1.6666, −0.8785) | <0.0001 |
| American Indian/Alaskan Native, Non-Hispanic | 1.1250 (0.6614, 1.5887) | <0.0001 |
| Hispanic | −0.2293 (−0.4903, 0.0317) | 0.0851 |
| Other race, Non-Hispanic | 0.7220 (0.4292, 1.0149) | <0.0001 |
| Age Category | ||
| 18–29 (ref) | 0.00 | — |
| 30–39 | −0.9891 (−1.2323, −0.7460) | <0.0001 |
| 40–59 | −1.1419 (−1.3476, −0.9362) | <0.0001 |
| 60 and older | −3.6062 (−3.8089, −3.4035) | <0.0001 |
| Income Category | ||
| Less than $50,000 (ref) | 0.00 | — |
| $50,000 or more | −2.5905 (−2.7159, −2.4652) | <0.0001 |
| BMI Category | ||
| Normal weight (ref) | 0.00 | — |
| Underweight | 1.2235 (0.7842, 1.6628) | <0.0001 |
| Overweight | 0.1950 (0.0448, 0.3451) | 0.0109 |
| Obese | 1.0718 (0.9243, 1.2193) | <0.0001 |
| Health Insurance | ||
| No (ref) | 0.00 | — |
| Yes | −0.6492 (−0.8751, −0.4233) | <0.0001 |
| Smoking Status | ||
| Never smoked (ref) | 0.00 | — |
| Current smoker – daily | 4.2994 (4.1159, 4.4828) | <0.0001 |
| Current smoker – some days | 3.2445 (2.9733, 3.5158) | <0.0001 |
| Former smoker | 1.1688 (1.0307, 1.3068) | <0.0001 |
-
Sexual orientation was independently associated with poor mental health days. Compared with straight participants, adjusted mean days were higher among lesbian/gay (β = 1.18), bisexual (β = 2.45), and “something else” participants (β = 2.65) (all p < 0.0001).
-
Female sex was associated with a modest increase in poor mental health days compared with males (β = 1.08; p < 0.0001).
-
Older age was strongly protective, with participants aged 60 years or older reporting 3.61 fewer poor mental health days compared with those aged 18–29 (p < 0.0001).
-
Higher income was associated with fewer poor mental health days (β = −2.59; p < 0.0001).
-
Smoking status showed the largest effect sizes, with daily smokers reporting 4.30 additional poor mental health days compared with never smokers (p < 0.0001).
Multivariable logistic regression was used to assess the association between sexual orientation and self-reported depressive disorder, adjusting for sex, race/ethnicity, age category, income category, BMI category, health insurance status, and smoking status. Note: Odds ratios in this table correspond to the outcome “no depressive disorder” modeled as the event of interest.
| Sexual Orientation | Adjusted Odds Ratio (95% CI) | P-value |
|---|---|---|
| Straight/not gay (ref) | 1.00 | — |
| Lesbian/Gay | 0.546 (0.499, 0.597) | <0.0001 |
| Bisexual | 0.446 (0.414, 0.481) | <0.0001 |
| Something else | 0.536 (0.479, 0.600) | <0.0001 |
-
After adjustment for demographic, socioeconomic, and health-related covariates, sexual orientation was significantly associated with odds of reporting a depressive disorder.
-
Compared with straight participants, individuals identifying as lesbian/gay, bisexual, or “something else” had lower adjusted odds of reporting a depressive disorder (all p < 0.0001).
-
These associations persisted after controlling for sex, age, race/ethnicity, income, BMI category, health insurance status, and smoking status.
Age-stratified multivariable models were fit to evaluate whether age modified the association between sexual orientation and mental health outcomes. Models were adjusted for sex, race/ethnicity, income category, BMI category, health insurance status, and smoking status. Note: For age-stratified logistic regression models, odds ratios correspond to the outcome “ Yes depressive disorder” modeled as the event of interest.
| Sexual Orientation (vs. Straight/not gay) | 18–29 years | 30–39 years | 40–59 years | ≥60 years |
|---|---|---|---|---|
| Lesbian/Gay | 2.61 (1.73, 3.48) | 0.74 (0.34, 1.81) | 0.86 (0.14, 1.59) | 0.84 (0.12, 1.57) |
| Bisexual | 3.48 (2.93, 4.04) | 2.81 (2.06, 3.56) | 1.67 (0.95, 2.38) | 0.87 (0.08, 1.67) |
| Something else | 5.48 (4.40, 6.55) | 3.18 (1.73, 4.64) | 1.20 (0.03, 2.38) | 1.85 (1.07, 2.62) |
Interaction p-value: <0.0001
| Sexual Orientation (vs. Straight/not gay) | 18–29 years | 30–39 years | 40–59 years | ≥60 years |
|---|---|---|---|---|
| Lesbian/Gay | 2.31 (1.90, 2.81) | 2.01 (1.58, 2.55) | 1.77 (1.52, 2.06) | 1.63 (1.38, 1.92) |
| Bisexual | 2.88 (2.54, 3.26) | 2.30 (1.95, 2.72) | 1.98 (1.71, 2.31) | 1.58 (1.32, 1.89) |
| Something else | 4.10 (3.22, 5.22) | 2.14 (1.55, 2.95) | 1.60 (1.26, 2.04) | 1.30 (1.08, 1.55) |
Interaction p-value: <0.0001
-
Age significantly modified the association between sexual orientation and both poor mental health days and depressive disorder (interaction p < 0.0001).
-
Across all age groups, sexual minority participants reported higher poor mental health burden and higher odds of depressive disorder compared with straight participants.
-
The magnitude of association was strongest among adults aged 18–29 and consistently attenuated with increasing age.
-
Bisexual individuals and those identifying as “something else” showed the largest effect sizes across age strata for both outcomes.
This analysis of the 2018 BRFSS data identifies meaningful disparities in mental health outcomes by sexual orientation. Several key patterns emerge from the descriptive, multivariable, and age-stratified analyses.
- Sexual minority groups consistently experienced poorer mental health outcomes compared with straight individuals.
- Bisexual individuals demonstrated the highest burden of poor mental health across multiple measures.
- Individuals identifying as “something else” also showed elevated mental health challenges, particularly in younger age groups.
- Sexual minority groups were disproportionately represented in younger age categories.
- Gender distributions differed by sexual orientation, with a higher proportion of males among lesbian/gay respondents and a higher proportion of females among bisexual respondents.
- Lower income levels were more common among bisexual individuals and those identifying as “something else,” suggesting potential socioeconomic vulnerability.
- The association between sexual orientation and mental health outcomes was strongest among younger adults.
- Sexual minorities aged 18–29 exhibited the largest disparities in both poor mental health days and depressive disorder.
- Although the magnitude of association decreased with age, disparities persisted across all age categories.
- Female sex was independently associated with worse mental health outcomes, particularly for depressive disorder.
- Lower income was strongly associated with increased mental health burden.
- Racial and ethnic associations varied by outcome, indicating complex and non-uniform patterns across groups.
00_run_all.sas: Master script to execute the full analytic pipeline01_import_and_formats.sas: Data import, formatting, and preprocessing02_clean_and_label.sas: Data cleaning, recoding, and variable labeling03_descriptives_bivariate.sas: Descriptive statistics and bivariate analyses04_model_menthlth.sas: Adjusted linear regression for poor mental health days05_model_depression.sas: Adjusted logistic regression for depressive disorder06_age_stratified_models.sas: Age-stratified models and interaction analyses
All analyses were conducted using SAS statistical software.