Table 4.

Cox regression on associations between risk factors and incident drinking

Potential factors	Case of incident drinking	Person-year (PY) of observation	IR (1000 PYs)	HR (95% CI)*	aHR (95% CI)†
Total	1107	36 137.23	30.63
Gender
Female	442	24 564.26	17.99	1.00	1.00
Male	665	11 572.97	57.46	3.10 (2.75 to 3.50)	3.46 (3.02 to 3.96)
Age, years
18–29	273	7396.68	36.91	1.00	1.00
30–39	267	7862.44	33.96	0.98 (0.83 to 1.16)	1.01 (0.84 to 1.20)
40–49	262	8860.09	29.57	0.92 (0.77 to 1.09)	1.04 (0.86 to 1.25)
50–59	167	5851.17	28.54	0.94 (0.78 to 1.14)	1.00 (0.82 to 1.23)
≥60	138	6166.85	22.38	0.71 (0.58 to 0.87)	0.72 (0.58 to 0.90)
Urban
No	703	24 932.37	28.20	1.00	1.00
Yes	404	11 204.86	36.06	2.30 (2.10 to 2.70)	3.50 (3.01 to 4.08)
Han Chinese
Yes	638	21 887.41	29.15	1.00	1.00
No	469	14 249.82	32.91	1.20 (1.10 to 1.40)	1.65 (1.44 to 1.88)
Education
>9 years	151	4082.76	36.98	1.00	1.00
≤9 years	956	32 054.47	29.82	0.82 (0.69 to 0.98)	1.14 (0.94 to 1.37)
Employed
No	159	6457.90	24.62	1.00	1.00
Yes	948	29 679.33	31.94	1.30 (1.10 to 1.50)	1.33 (1.16 to 1.52)
Married
Yes	856	29 149.81	29.37	1.00	1.00
No	251	6987.42	35.92	1.15 (1.01 to 1.33)	1.06 (0.91 to 1.24)
Current smokers
No	853	31 167.35	27.37	1.00	1.00
Yes	254	4969.88	51.11	1.80 (1.60 to 2.10)	1.17 (1.00 to 1.38)
History of chronic diseases‡
Yes	686	23 647.71	29.01	1.00	1.00
No	421	12 489.52	33.71	1.31 (1.16 to 1.48)	1.19 (1.04 to 1.35)

*Univariate Cox regression.

†Multivariate Cox regression adjusted by all covariates excluding itself.

‡Chronic diseases include diabetes mellitus, hypertension and dyslipidaemia at baseline.

aHR, adjusted HR; IR, incidence rate.