Social Medical Insurances, Choices of Medical Institutions and the ‘Siphon Effect’ in the Health Service Market: Evidence from 2021 Yangtze River Delta Region of China

Figures & data

Figure 1 The framework of China’s social medical insurance systems.

Figure 2 The flowchart of sample selection.

Table 1 The Values of Distance-First Visit Choice Variable

The First Choice The Nearest Hospital	Primary Hospitals	General Hospitals
Primary hospitals	0 was valued (4518)	1 was valued (6390)
General hospitals	0 was valued (180)	0 was valued (3137)

Note: Data source: HLSSYRD.

Table 2 Definition and Descriptive Results of Variables

Variables	Definitions	Frequency/Mean	Percent/SE
Choice of medical institution	Primary hospital = 0	4698	33.03%
	General hospital = 1	9527	66.97%
Siphon effect	No = 0	7835	55.08%
	Yes = 1	6390	44.92%
Social medical insurance	BMISURR = 0	3730	26.22%
	BMISUE = 1	10,495	73.78%
Age	15–29 = 1	2845	20.00%
	30–44 = 2	4963	34.89%
	45–59 = 3	3437	24.16%
	≥60 = 4	2980	20.95%
Gender	Female = 0	7060	49.63%
	Male = 1	7165	50.37%
Marital status	Unmarried = 0	2347	16.50%
	Married = 1	11,878	83.50%
Education level	Middle school and below = 1	1419	9.98%
	High school/Vocational school = 2	4687	32.95%
	Three-year college = 3	4987	35.06%
	Four-year college and above = 4	3132	22.02%
Working status	No work = 0	4765	33.50%
	Having a work = 1	9460	66.50%
Monthly household income	0–10,000 CNY = 1	2700	18.98%
	10,001–20,000 CNY = 2	7915	55.64%
	20,001–40,000 CNY = 3	3189	22.42%
	40,001 CNY and above = 4	421	2.96%
Chronic disease	No = 0	10,763	75.66%
	Yes = 1	3462	24.34%
Self-rated health	Poor = 0	3667	25.78%
	Good = 1	10,558	74.22%

Note: Data source: HLSSYRD.

Table 3 The Results of the Chi-Square Test

	Social Medical Insurance		Chi-Square Test	P-value
	BMISURR	BMISUE
Choice of medical institution			2.0e+03	<0.001
Primary hospital	2343 (49.87)	2355 (50.13)
General hospital	1387 (14.56)	8140 (85.44)
Siphon effect			922.53	<0.001
No	2847 (36.34)	4988 (63.67)
Yes	883 (13.82)	5507 (86.18)
Age			461.69	<0.001
15–29	1121 (39.40)	1724 (60.60)
30–44	951 (19.16)	4012 (80.84)
45–59	733 (21.33)	2704 (78.67)
≥ 60	925 (31.04)	2055 (68.96)
Gender			0.03	0.856
Female	1856 (26.29)	5204 (73.71)
Male	1874 (26.15)	5291 (73.85)
Marital status			140.24	<0.001
Unmarried	846 (36.05)	1501 (63.95)
Married	2884 (24.28)	8994 (75.72)
Education level			648.81	<0.001
Middle school and below	756 (53.28)	663 (46.72)
High school/Vocational school	1258 (26.84)	3429 (73.16)
Three-year college	1084 (21.74)	3903 (78.26)
Four-year college and above	632 (20.18)	2500 (79.82)
Working status			527.29	<0.001
No	1818 (38.15)	2947 (61.85)
Yes	1912 (20.21)	7548 (79.79)
Monthly household income			451.89	<0.001
0–10,000	1133 (41.96)	1567 (58.04)
10,001–20,000	1891 (23.89)	6024 (76.11)
20,001–40,000	632 (19.82)	2557 (80.18)
40,001 and above	74 (17.58)	347 (82.42)
Chronic disease			3.35	0.067
No	2781 (25.84)	7982 (74.16)
Yes	949 (27.41)	2513 (72.59)
Self-rated health			735.86	<0.001
No	1584 (43.20)	2083 (56.80)
Yes	2146 (20.33)	8412 (79.67)

Note: Percent were reported in brackets; Data source: HLSSYRD.

Table 4 The Regression Results for the Association Between Health Insurance and Choices of Medical Institution

Variables	(1)		(2)
	Choice of Medical Institutions		Siphon Effect
BMISUE (Ref: BMISURR)	5.377***	(4.887, 5.915)	3.240***	(2.945, 3.565)
Age (Ref: 15–29)	1.008***	(1.003, 1.013)	0.998	(0.993, 1.002)
30–44	1.182*	(1.030, 1.357)	0.881*	(0.778, 0.999)
45–59	1.197*	(1.017, 1.409)	0.744***	(0.649, 0.865)
60 and above	1.339**	(1.076, 1.666)	0.656***	(0.538, 0.799)
Gender (Ref: Female)
Male	1.003	(0.922, 1.085)	1.057	(0.984, 1.135)
Martial status (Ref: Unmarried)
Married	2.315***	(1.967, 2.724)	0.966	(0.850, 1.097)
Education level (Ref: Middle school and below)
High school/Vocational school	1.002	(0.854, 1.171)	1.145	(0.988, 1.327)
Three-year college	1.211*	(1.012, 1.450)	1.336***	(1.132, 1.578)
Four-year college and above	1.203	(0.986, 1.469)	1.454***	(1.213, 1.743)
Work status (Ref: no work)
Having a job	0.789***	(0.693, 0.899)	0.679***	(0.606, 0.763)
Monthly family income (Ref: 0–10,000)
10,001–20,000	4.465***	(3.998, 4.987)	2.625***	(2.363, 2.916)
20,001–40,000	3.495***	(3.066, 3.983)	1.440***	(1.271, 1.632)
40,001 and above	2.628***	(2.039, 3.385)	0.810	(0.633, 1.036)
Chronic disease (Ref: No)
Yes	1.141*	(1.003, 1.298)	1.370***	(1.219, 1.540)
Self-rated health (Ref: Poor)
Good	0.512*	(0.237, 1.102)	0.237***	(0.124, 0.453)
Constant	0.077***	(0.032, 0.188)	0.422***	(0.199, 0.897)
Observations	14,225		14,225
Pseudo R2	0.182		0.093

Notes: 1) OR, odds ratio; 2) CI, confidence interval; 3) *P < 0.05; **P < 0.01; ***P < 0.001; 4) Data source: HLSSYRD.

Table 5 PSM Results for the Association Between Social Medical Insurance and the Choice of Medical Institutions

		Types of matching	Treated	Controlled	ATT	S. E	T-value
(1)	Choice of medical institution	K-nearest neighbor matching	0.776	0.446	0.329	0.016	20.72***
		Caliper matching	0.776	0.443	0.332	0.011	28.93***
		Kernel matching	0.776	0.440	0.335	0.011	29.23***
(2)	Siphon effect	K-nearest neighbor matching	0.525	0.278	0.247	0.015	16.91***
		Caliper matching	0.525	0.282	0.243	0.011	22.81***
		Kernel matching	0.525	0.280	0.244	0.011	22.99***

Note: ***P < 0.001; Data source: HLSSYRD.

Table 6 The Regression Results of Heterogeneity Analysis

Variables	(1)	(2)	(3)	(4)	(5)
	Age Differences		Household Income Differences
	15–59 years	≥ 60 years	0–10,000	10,001–20,000	> 20,000
BMISUE	3.543*** (3.165, 3.967)	2.462*** (2.030, 2.986)	2.019*** (1.565, 2.604)	3.123*** (2.494, 3.918)	4.033*** (3.570, 4.556)
Other variables	Yes	Yes	Yes	Yes	Yes
Observations	11,245	2980	2700	7915	3610
Pseudo R2	0.093	0.112	0.039	0.091	0.056

Notes: 1) ***P < 0.001; 2) 95% confidence interval (CI) were reported in brackets; 3) Data source: HLSSYRD.

Table 7 Reimbursement Rates in Different Hospital Levels in YRDC

Medical Institution/Province		Zhejiang (Huzhou City)		Anhui (Hefei City)
		BMISUE	BMISURR	BMISUE	BMISURR
Primary care facilities	Outpatient	60%	55%	80%	60%
	Inpatient	90%	95%	94%	90%
Secondary hospitals	Outpatient	50%	30%	70%	60%
	Inpatient	85%	75%	92%	85%
Tertiary hospitals	Outpatient	50%	20%	60%	60%
	Inpatient	80%	60%	80%	75–80%

Notes: Data source: Huzhou healthcare Security Administration and Hefei healthcare Security Administration.