Table 3.
Meta-analysis Results for the Odds of Suicide Attempts, Suicidal Ideation, and Low Life Satisfaction Among People Aged 50 and Older by Sexual Orientation
| Unadjusted | Adjusted | |||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Model | Studies (N) | LGB people (N) | Heterosexual people (N) | Odds ratio | Confidence interval | I 2 | τ 2 | Odds ratio | Confidence interval | I 2 | τ 2 | |
| Suicide attempts | Overall | 4a,b,c | 258 | 6,988 | 2.250 | 1.184–4.277 | 20.9% | 0.106 | 2.034 | 1.169–3.538 | 11.3% | 0.042 |
| Male | 4a,b,c | 124 | 3,109 | 3.418 | 1.708–6.840 | 0.0% | 0.000 | 2.290 | 1.187–4.420 | 0.0% | 0.000 | |
| Female | 4a,b,c | 134 | 3,879 | 0.936 | 0.794–4.717 | 14.2% | 0.149 | 1.881 | 0.969–3.651 | 0.0% | 0.000 | |
| Suicidal ideation | Overall | 1 | 189 | 3,532 | 1.445 | 0.919–2.274 | — | — | 1.310 | 0.825–2.079 | — | — |
| Male | 1 | 90 | 1,519 | 2.523 | 1.304–4.884 | — | — | 2.026 | 1.041–3.943 | — | — | |
| Female | 1 | 99 | 2,013 | 0.846 | 0.415–1.727 | — | — | 0.815 | 0.393–1.690 | — | — | |
| Low life satisfaction | Overall | 10 | 1,044 | 38,136 | 1.439 | 1.173–1.766 | 0.0% | 0.000 | 1.260 | 1.028–1.544 | 0.0% | 0.000 |
| Male | 10 | 525 | 17,050 | 1.630 | 1.230–2.162 | 0.0% | 0.000 | 1.329 | 1.006–1.756 | 0.0% | 0.000 | |
| Female | 10 | 519 | 21,086 | 1.325 | 0.999–1.757 | 0.0% | 0.000 | 1.239 | 0.943–1.628 | 0.0% | 0.000 | |
Note. LGB = lesbian, gay, and bisexual; LGBT = lesbian, gay, bisexual, and transgender.
aScottish Health Survey 2009 collected data on suicide attempts only from a subset of survey respondents and data are not included in the analysis due to very low number of LGBT respondents included in the subset which could not be disaggregated by gender.
bScottish Health Survey 2013 collected data on suicide attempts only from a subset of survey respondents and not included in the analysis due to very low number of LGBT respondents included in the subset.
cScottish Health Survey 2010 collected data on suicide attempts only from a subset of survey respondents and not included in the analysis due to very low number of LGBT respondents included in the subset.