Table 4.
Model-standardized risks for intervenable variables on suicide attempts among trans people who have seriously considered suicide, Ontario, Canada (n = 110)
| Potential intervention factor | number | RF a | Model-standardized RRd | Counterfactual population attributable risk (RF - R1)f | Counterfactual population attributable risk proportion | |
|---|---|---|---|---|---|---|
| R1 b | ||||||
| R0 c | RR | 95 % CIe | ||||
| Social inclusion | ||||||
| Social support | 109 | 0.30 | ||||
| 90th percentile (4.895) | 0.08 | 0.18 | (0.04, 0.73) | 0.22 | 73 % | |
| 10th percentile (1.947) | 0.47 | 1.00 | -- | |||
| Any strong parental support for genderh | 99 | 0.31 | ||||
| Yes | 0.10 | 0.29 | (0.07, 1.24) | |||
| No | 0.35 | 1.00 | -- | |||
| Any strong family support for genderh | 106 | 0.30 | ||||
| Yes | 0.27 | 0.81 | (0.41, 1.62) | |||
| No | 0.33 | 1.00 | -- | |||
| Any strong support from leadersh | 72 | 0.34 | ||||
| Yes | 0.58 | 5.24 | (2.20, 12.46) | −0.24 | n/a | |
| No | 0.11 | 1.00 | -- | |||
| Any strong support from peersh | 106 | 0.30 | ||||
| Yes | 0.32 | 1.34 | (0.43, 4.23) | |||
| No | 0.24 | 1.00 | -- | |||
| Religiosity or spirituality | 109 | 0.30 | ||||
| 90th percentile (5) | 0.20 | 0.52 | (0.12, 2.18) | |||
| 10th percentile (1) | 0.39 | 1.00 | -- | |||
| Having ≥1 ID concordant with lived genderi | 55 | 0.37 | ||||
| Yes | 0.14 | 0.26 | (0.11, 0.62) | 0.23 | 62 % | |
| No | 0.54 | 1.00 | -- | |||
| Transphobia | ||||||
| Transphobia scale | 108 | 0.31 | ||||
| 10th percentile (5) | 0.11 | 0.24 | (0.07, 0.82) | 0.20 | 65 % | |
| 90th percentile (23) | 0.45 | 1.00 | -- | |||
| Transphobic harassment and violence | 110 | 0.32 | ||||
| None | 0.18 | 0.30 | (0.08, 1.16) | 0.14 | 44 % | |
| Verbal harassment or threats | 0.18 | 0.31 | (0.11, 0.83) | |||
| Physical or sexual assault | 0.59 | 1.00 | -- | |||
| Transition | ||||||
| Medical transition statusj | 100 | 0.35 | ||||
| Completed | 0.11 | 0.51 | (0.07, 3.74) | 0.24 | 69 % | |
| In process | 0.65 | 2.91 | (1.47, 5.76) | |||
| Planning, but not begun | 0.22 | 1.00 | -- | |||
| Current hormone usej | 98 | 0.36 | ||||
| Yes | 0.30 | 0.76 | (0.41, 1.39) | |||
| No | 0.40 | 1.00 | -- | |||
| Social transition status | 109 | 0.31 | ||||
| Full-time | 0.45 | 5.30 | (0.66, 42.68) | |||
| Part-time | 0.21 | 2.53 | (0.34, 18.60) | |||
| Not living in core gender | 0.08 | 1.00 | -- | |||
| Being socially seen as cisgenderi | 56 | 0.36 | ||||
| Almost always or always | 0.48 | 0.98 | (0.58, 1.64) | |||
| About half time or often | 0.14 | 0.28 | (0.02, 3.26) | |||
| Rarely or never | 0.49 | 1.00 | -- | |||
Risks are standardized to sociodemographic factors, childhood trauma factors and major health-related background factors. These included: age, gender spectrum, gender fluidity, ethno-racial group, immigration history, sexual orientation, region of province, disability status, parental status, religious upbringing, childhood sexual or physical abuse, diagnosis with major mental health disorder (excluding depression or anxiety), and chronic illness or pain
aRF = estimated risk of past-year suicidal ideation in the factual trans population of Ontario
bR1 = model-standardized estimated risk of past-year suicidal ideation in the trans population of Ontario, under the same distribution of background factors, but where all members have a positive value of the potential intervention (e.g., high support, no transphobic violence)
cR0 = model-standardized estimated risk of past-year suicidal ideation in the trans population of Ontario, under the same distribution of background factors, but where all members have a negative value of the potential intervention (e.g., low support, exposure to transphobic violence)
dWill vary slightly from R1/R0 as are calculated as average of individual marginal risk ratios, rather than the ratio of model-standardized risks
e95 % confidence intervals from Taylor series linearization methods in SUDAAN
fCounterfactual population attributable risk = potential proportion of population protected from suicidal ideation by a hypothetical move from population levels of this factor to an intervention level
gCounterfactual population attributable risk proportion = potential proportion of outcomes that could be averted under a change in exposure frequency
hSupport variables include either indication of – or expectation of – strong support
iamong those living full-time in a non-fluid gender (n = 57)
jamong those reporting need to medically transition sex (n = 100)