Table 5.
Multivariable analysis to examine longitudinal associations with mental health outcomes and effect-modification with uptake of counselling
| Final adjusted OR (95% CI) | p valuea | Did not attend counselling | Attended counselling | p valuea | ||
|---|---|---|---|---|---|---|
| Baseline demographicsb | ||||||
| Age (years) at baselineb | <25 | 1.0 | 1.0 | 1.0 | ||
| 25–34 | 2.68 (1.40–5.12) | 2.13 (0.72–6.24) | 3.49 (1.19–10.2) | |||
| 35+ | 2.29 (1.19–4.42) | 0.01 | 1.46 (0.51–4.21 | 3.84 (1.33–11.06 | 0.2 | |
| Religion | Catholic | 1.0 | 1.0 | 1.0 | ||
| Protestant | 0.79 (0.54–1.15) | 0.81 (0.47–1.40) | 0.69 (0.40–1.19 | |||
| Muslim | 2.74 (1.43–5.24) | 3.22 (1.13–9.17) | 1.55 (0.54–4.41) | |||
| Other/none | 1.83 (0.83–4.00) | <0.001 | 1.35 (0.48–3.84 | 2.70 (0.95–7.68) | 0.5 | |
| Ever married | Yes | 2.07 (1.20–3.56) | 0.009 | 2.01 (0.84–4.79) | 2.65 (1.11–6.31) | 0.6 |
| Financial stressors, sex work characteristics, sexual risk behaviours, violence and alcohol/substance use (time variant exposures)c | ||||||
| Recent hunger – missed a meal in the last 7 days | Yes | 1.99 (1.37–2.88) | <0.001 | 1.77 (1.03–3.03) | 2.23 (1.40–3.82) | 0.5 |
| Clients per week | <5 | 1.0 | 1.0 | 1.0 | ||
| 5–9 | 0.68 (0.40–1.16) | 0.63 (0.28–1.44) | 0.76 (0.34–1.73) | |||
| 10+ | 0.45 (0.20–1.00) | 0.07 | 0.23 (0.09–0.8) | 0.78 (0.27–2.25) | 0.4 | |
| Any recent sexual and/or physical non-IP violence | Yes | 2.23 (1.55–3.19) | <0.0001 | 2.27 (1.36–3.80) | 2.26 (1.35–3.79) | 1.0 |
a
Adjusted Wald test.
b
Adjusted for all a priori potential confounders (age at baseline, religion, SES, literacy, ever married) and SWOP clinic.
c
Adjusted for all a priori potential confounders (age at baseline, religion, SES, literacy, ever married), clinic, violence from non-IPs, hunger and clients per week.