Table 5. Confounder adjusted association between costs due to TB diagnosis and various predictor variables using linear regression models in the study of Bao'an district, Shenzhen City, China, 2013 (n = 514).

Predictors in the model	Beta coefficient	(95% CI)	P value
Number of times visiting health-care facilities
< = 2	-0.95	-1.18,-0.73	< 0.001
2~6	-0.46	-0.65,-0.27	< 0.001
> = 7	Ref	Ref	Ref
Whether in hospital due to TB diagnosis
Yes	Ref	Ref	Ref
No	-1.13	-1.14,-0.95	< 0.001
Education
Primary/illiterate	Ref	Ref	Ref
Junior high school	0.10	-0.03,0.24	0.139
Senior high school	0.11	-0.04,0.25	0.146
College or above	0.24	0.07,0.41	0.007
Occupation
Workers	Ref	Ref	Ref
Individual business	0.026	-0.12,0.17	0.731
Others national civil servant/services /retired staff etc.)	0.12	0.002,0.24	0.046
Unemployed	-0.07	-0.22,0.8	0.341

Linear regression was done after conversion to log scale.

Only significant variables were presented.

Ref-reference.