Table 4.

Sensitivity analysis of the models in Table 2

Variable	M/N	(P25, P75)/%	Total medical expenses (reference = low)				Actual medical insurance payment level (reference = high)				The group with high out-of-pocket expenses
			Middle		High		Low		Middle
			P	AOR (95% CI)	P	AOR (95% CI)	P	AOR (95% CI)	P	AOR (95% CI)	P	AOR (95% CI)
Length of stay	5	(2, 11)	< 0.001	1.251* (1.238, 1.263)	< 0.001	1.506* (1.490,1.523)	< 0.001	0.988* (0.985, 0.992)	< 0.001	1.006* (1.003, 1.009)	< 0.001	1.106* (1.101, 1.110)
Age
< 60	9819	39.6	0.010	1.097 (1.022, 1.176)	< 0.001	1.412* (1.295, 1.540)	< 0.001	1.619* (1.508, 1.737)	< 0.001	1.584* (1.480, 1.695)	< 0.001	1.244* (1.127, 1.374)
≥ 60 (reference)	14,984	60.4
Insurance type
Urban and rural residents	8281	33.4	0.006	0.902* (0.839, 0.970)	< 0.001	0.801* (0.731, 0.877)	< 0.001	6.183* (5.713, 6.692)	< 0.001	2.649* (2.446, 2.869)	< 0.001	1.440* (1.301, 1.593)
Urban workers (reference)	16,522	66.6
Economic level
High	15,776	63.6	< 0.001	2.608* (2.357, 2.887)	< 0.001	15.509* (13.274, 18.120)	< 0.001	2.551* (2.286, 2.847)	< 0.001	1.279* (1.166, 1.404)	< 0.001	8.375* (6.576, 10.667)
Middle	5974	24.1	0.070	0.905 (0.813, 1.008)	< 0.001	2.129* (1.087, 2.510)	< 0.001	3.109* (2.752, 3.513)	< 0.001	1.713* (1.540, 1.905)	< 0.001	5.159* (3.994, 6.662)
Low (reference)	3053	12.3

This table contained the results of three models after adding the length of stay variable and “*” indicated significant at the 0.01 level