Table 3.

Multivariate multinomial logistic regression model of return to same rural region of work as rural region of training experienced during basic medical training

Ref: work in metropolitan	Same rural region of work (RRR, 95% CI)	Different rural region of work (RRR, 95% CI)
Career stage (ref = Pre-registrar)
Non-GP: registrar	0.24 (0.14–0.39)***	0.76 (0.53–1.09)
Non-GP: independent	0.63 (0.44–0.90)*	1.08 (0.80–1.46)
GP: registrar	2.70 (1.67–4.38)***	4.37 (2.87–6.64)***
GP: independent	2.09 (1.51–2.89)***	3.53 (2.64–4.72)***
Female (ref = Male)	0.75 (0.59–0.94)*	0.79 (0.66–0.94)**
Rural region of training (ref =  < 12 weeks)
3–12 month training	1.42 (1.08–1.88)*	0.99 (0.81–1.21)
> 1 year training	5.22 (3.95–6.89)***	1.54 (1.19–1.98)**
Rural region of origin (ref = no)a	3.24 (2.54–4.12)***	2.32 (1.91–2.81)***

N = 4097 (Excluded: nil rural training); RRR: relative risk ratio; *p < 0.05, **p < 0.01, ***p < 0.001

^aAnalysis was repeated with rural origin separated into 6–11 years and 12–18 years; however, this revealed similar effect sizes of same rural region of work: 6–11 years (RRR 3.16, 2.08–4.81)***, 12–18 years (RRR 3.26, 2.51–4.24)***