TABLE 3.

Multivariate linear regression analysis results of the proportionate OOP expenditure for lung cancer-related treatment with medicines of interest.

Variable	Coefficient	SE	95% CI		t	p-value
			Lower	Upper
Age (ref.<50 years old)
50–60 years old	0.028	0.022	−0.014	0.071	1.323	0.186
60–70 years old	0.013	0.022	−0.031	0.056	0.578	0.563
>70 years old	0.016	0.024	−0.031	0.063	0.679	0.498
Male (ref. Female)	−0.022	0.012	−0.045	0.002	−1.778	0.076
Local medical patient (ref. Non-local)	−0.002	0.041	−0.082	0.079	−0.037	0.970
Type of health insurance (ref. Formal employee program enrollee entitled to government-funded supplementary health insurance coverage)
Formal employee program enrollee not entitled to government-funded supplementary health insurance coverage	0.183	0.022	0.141	0.226	8.506	<0.001
Resident program enrollee	0.267	0.029	0.210	0.324	9.211	<0.001
Status of employment (ref. formal employed)
Retired	-0.085	0.027	-0.137	-0.033	-3.198	0.001
Flexible employed	-0.011	0.048	-0.106	0.083	-0.235	0.814
Medication (ref. Icotinib)	0.104	0.013	0.079	0.129	8.054	<0.001

Note: Bold implies statistically significant. Coefficient means the expected absolute changes of the proportionate OOP expenditure for lung cancer-related treatment with medicines of interest when the categorical variable shifted from the reference category to respective category, holding all the other variables constant, e.g., if the proportionate OOP of patients treated with icotinib was 30%, that of patients treated with gefitinib was 40.4%.

SE, standard error; CI, confidence interval; OOP, out of pocket.