Trends in silicosis prevalence and the healthy worker effect among gold miners in South Africa: a prevalence study with follow up of employment status

Table 4 Survival analysis (Cox regression) of time to exiting the workforce (for any reason) during follow up (N = 11 529)^a

Model	Disease	Presence	Number	Left workforce	Cumulative proportion leaving workforce by 2 years	HR (95 % CI)	P value (Wald)	Adjusted HR (95 % CI)	P value (Wald)
1	Silicosis (grade 1/1+)	No	11,053	2473	0.175	1.0	0.001	1.0	0.12
	Silicosis (grade 1/1+)	Yes	476	174	0.252	1.69 (1.34, 2.11)^b		1.18 (0.96, 1.47)^c
	TB^d	No	8512	1623	0.150	1.0	<0.001	1.0	<0.001
	TB^d	Yes	3017	1024	0.260	1.97 (1.77, 2.19)^b		1.64 (1.47, 1.84)^c
	Interaction between silicosis and active/previous TB
2	Silicosis (grade 1/1+) and TB^d	No silicosis, no TB	8230	1525	0.147	1.0	<0.001	1.0	<0.001
		Silicosis, no TB	282	98	0.233	1.97 (1.51, 2.57)^b		1.54 (1.17, 2.04)^e
		No silicosis, TB	2823	948	0.258	2.01 (1.79, 2.26)^b		1.71 (1.51, 1.94)^e
		Silicosis plus TB*	194	76	0.281	2.26 (1.84, 2.78)^b		1.54 (1.20, 1.96)^e

HR hazard ratio, CI confidence interval, TB tuberculosis
^aBased on reader I’s readings
^bAdjusted for clustering using robust standard errors
^cAdjusted for age category, years since first employed, country of origin, time spent underground, occupational level, silicosis/TB (active or previous) as appropriate, and clustering using robust standard errors. (n = 11 371)
^dTB defined as either previous or active TB at baseline by medical history or CXR by reader I
^eAdjusted for age category, years since first employed, country of origin, time spent underground, occupational level, and clustering using robust standard errors. (n = 11 371)
*p-value for interaction between active/previous TB and silicosis <0.001 adjusting for clustering only, and p-value <0.001 from the fully adjusted model

ISSN: 1471-2458