Table 2.
Estimated Percentages of U.S. Adults Endorsing At Least One Symptom From Past-Year DSM-5 Alcohol Use Disorder Criteria Groupings Based on Sexual Orientation: Overall Sample
| No AUD criteria (no criteria) % | Impaired Control (at least one symptom; criteria #1 - #4) % | Social Impairment (at least one symptom; criteria #5 - #7) % | Risky Use (at least one symptom; criteria #8 - #9) % | Pharmacologic (at least one symptom; criteria #10 - #11) % | |
|---|---|---|---|---|---|
| Men | |||||
| All Men (n = 15,544) | 71.9% | 22.1% | 7.0% | 14.5% | 12.5% |
| Sexual Orientation Subgroups | |||||
| Heterosexual-identified, no same-sex attraction or behavior (n = 14,228) | 72.2% | 21.7% | 6.9% | 14.5% | 12.3% |
| Heterosexual-identified, same-sex attraction or behavior (n = 782) | 73.4% | 21.3% | 7.0% | 10.9% | 12.1% |
| Gay-identified (n = 321) | 60.4% | 33.0% | 12.2% | 20.8% | 16.7% |
| Bisexual-identified (n = 144) | 59.0% | 38.4% | 14.1% | 20.2% | 21.5% |
| Not sure (n = 69) | 67.9% | 27.9% | 10.3% | 13.1% | 21.6% |
| Differences1 | P < 0.01 | P < 0.01 | P = 0.01 | P < 0.01 | P = 0.02 |
| Women | |||||
| All Women (n = 19,956) | 82.0% | 14.0% | 3.6% | 7.7% | 7.7% |
| Sexual Orientation Subgroups | |||||
| Heterosexual-identified, no same-sex attraction or behavior (n = 17,845) | 83.5% | 12.7% | 3.0% | 6.6% | 6.8% |
| Heterosexual-identified, same-sex attraction or behavior (n = 1,294) | 72.2% | 22.1% | 6.5% | 15.1% | 12.9% |
| Lesbian-identified (n = 265) | 68.6% | 27.5% | 10.1% | 18.9% | 18.2% |
| Bisexual-identified (n = 422) | 57.1% | 35.3% | 15.0% | 22.4% | 21.6% |
| Not sure (n = 130) | 64.8% | 32.0% | 13.1% | 21.5% | 22.9% |
| Differences1 | P < 0.01 | P < 0.01 | P < 0.01 | P < 0.01 | P < 0.01 |
Notes:
1
Differences are based on Rao-Scott chi-square tests. Past-year non-drinkers originally had missing values on the 11 DSM-5 criteria in the NESARC-III data, had the missing values overwritten with responses of “no” prior to variable creation and analysis. Small counts of unknown responses (9) were recoded to 0 (no) for each analysis; results did not change when treating unknown responses as missing.