Table 3.

Estimates of multivariable multinomial logistic regression model of latent cluster allocation (modelling probability of being in Cluster 1).

Variable	Characteristic	Cluster 2		Cluster 3		Type III P-Value
		Odds Ratio (95% CI)	Pairwise P-Value	Odds Ratio (95% CI)	Pairwise P-Value
Age	35–54 vs. 18–34	1.273 (0.720, 2.251)	0.4059	1.749 (0.995, 3.073)	0.0519	0.0125*
	55 + vs. 18–34	1.229 (0.509, 2.970)	0.6461	3.180 (1.392, 7.264)	0.0061*
Gendera	Female vs. Male	1.381 (0.815, 2.338)	0.2303	2.907 (1.727, 4.891)	<0.0001*	<0.0001*
Marital	Married or living together in a relationship vs. Never married	1.300 (0.687, 2.460)	0.4199	1.544 (0.832, 2.862)	0.1683	0.1711
	Other vs. Never married	1.749 (0.881, 3.471)	0.1100	1.146 (0.586, 2.241)	0.6899
Education	Completed HSC vs. Some high school or less	1.492 (0.687, 3.239)	0.3114	1.590 (0.734, 3.444)	0.2394	0.0446*
	TAFE certificate or diploma vs. Some high school or less	0.952 (0.534, 1.700)	0.8692	0.815 (0.461, 1.440)	0.4804
	University degree or higher vs. Some high school or less	0.308 (0.127, 0.747)	0.0092*	0.392 (0.175, 0.877)	0.0227*
Employment	Other vs. Employed	1.142 (0.524, 2.488)	0.7385	1.160 (0.535, 2.514)	0.7074	0.0572
	Unable to work due to health reasons vs. Employed	1.518 (0.769, 2.993)	0.2287	2.766 (1.419, 5.391)	0.0028*
	Unemployed vs. Employed	1.293 (0.643, 2.600)	0.4710	1.528 (0.758, 3.079)	0.2360
Depression	Yes vs. No	0.987 (0.593, 1.643)	0.9595	1.536 (0.930, 2.536)	0.0933	0.0703
Personality disorder	Yes vs. No	0.874 (0.437, 1.747)	0.7025	0.365 (0.177, 0.751)	0.0062*	0.0036*
Substance use disorder	Yes vs. No	11.953 (1.563, 91.423)	0.0168*	7.504 (0.954, 59.014)	0.0555	0.0381*

*Asterix denotes (P=<0.05).

^aGender included male and female, where transgender or gender non-conforming was excluded from the respective logistic regression model due to low numbers (n = 2).