Preferences for cervical cancer screening service attributes in rural China: a discrete choice experiment

Figures & data

Table 1 DCE attributes and levels

Attribute 1	The percentage of cervical cancer-related death reduction
Level 1	Reduced 20%
Level 2	Reduced 50%
Level 3	Reduced 80%
Attribute 2	Screening interval
Level 1	Every three years
Level 2	Every two years
Level 3	Every year
Attribute 3	Screening location
Level 1	County
Level 2	Town
Level 3	Village
Attribute 4	Screening pain
Level 1	Moderate
Level 2	Mild
Level 3	None
Attribute 5	Waiting time for screening results
Level 1	Three months to half year
Level 2	Two weeks to three months
Level 3	Within two weeks
Attribute 6	Out-of-pocket costs
Level 1	300 CNY
Level 2	50 CNY
Level 3	0 CNY

Notes: The average annual exchange rate between U$ and CNY in 2015 was: US$1= CNY 6.227.

Table 2 An example discrete choice task.  Please consider the following two cervical cancer screening scenarios and let us know your preference by ticking the corresponding box. Will you actually participate in the screening program you chose if it was offered to you? 

Attribute	Screening scenario 1	Screening scenario 2
The percentage of cervical cancer-related death reduction	Reduced 80%	Reduced 20%
Screening interval	Every three years	Every year
Screening location	County	Village
Screening pain	None	Mild
Waiting time for results	Within 2 weeks	2 weeks-3 months
Out-of-pocket costs	300 CNY	0 CNY
Your preference?	☐	☐
Will you actually participate in the screening program you chose if it was offered to you?	☐Yes	☐No

Table 3 Demographic characteristics of participants

Characteristics of participants	Full sample N=405		Analysis sample (who passed the consistency test) N=348		Excluded sample (who failed the consistency test) N=57		χ2 (P-value)
Age (year), N (%)							0.940 (0.6)
Mean ± SD	49.2	8.30	49.1	8.33	50.1	8.10
30–44	110	27.2%	97	27.9%	13	22.8%
45–54	185	45.6%	159	45.7%	26	45.6%
55–65	110	27.2%	92	26.4%	18	31.6%
Marital status, N (%)							2.170 (0.3)
Married	390	96.3%	336	96.6%	54	94.7%
Divorced or widowed	15	3.7%	12	3.4%	3	5.3%
Education level, N (%)							4.892 (0.3)
Did not complete primary school	111	27.4%	93	26.7%	18	31.6%
Completed Primary school	124	30.6%	102	29.3%	22	38.6%
Completed Middle school	134	33.1%	120	34.5%	14	24.6%
Completed High school or above	36	8.9%	33	9.5%	3	5.3%
Family income (CNY), N (%)							0.966 (0.8)
<10,000	104	25.7%	88	25.3%	16	28.1%
10,000–20,000	104	25.7%	89	25.6%	15	26.3%
20,000–30,000	91	22.4%	81	23.3%	10	17.5%
>30,000	106	26.2%	90	25.9%	16	28.1%

Table 4 Main effects model for DCE (n=348)

Attribute levels	Mean (SE)	SD (SE)	WTP (CNY)	95% Confidence interval
No screening	−4.309** (0.599)	5.532** (0.526)
Reduced 50%	0.877** (0.113)	0.055 (0.163)	66.829	49.333	85.489
Reduced 80%	1.404** (0.127)	0.503 (0.274)	106.942	87.483	129.774
Every 2 years	0.992** (0.122)	0.299 (0.305)	75.606	57.220	96.445
Every year	1.260** (0.131)	1.174** (0.168)	96.000	75.892	119.296
Town	0.678** (0.128)	0.963** (0.193)	51.680	32.708	72.570
Village	1.448** (0.124)	0.684** (0.191)	110.270	91.514	131.890
Mild pain	−0.219 (0.114)	0.074 (0.193)	−16.667	−34.738	0.147
No pain	0.270** (0.094)	0.119 (0.260)	20.582	6.628	34.994
2 weeks −3 months	0.581** (0.118)	0.131 (0.232)	44.256	26.627	63.301
<2weeks	0.802** (0.118)	0.799** (0.161)	61.103	43.738	80.772
Cost	−0.013** (0.001)	0.009** (0.001)
Log likelihood	−1899.8619 348 9,396
Participants
Observations

Notes: Mixed logit estimates reported. *P<0.05; ** P<0.01. Dummy coding used except for the cost variable.

Abbreviations: SD, Standard Deviation estimates reflect preference heterogeneity in the participants; SE, Standard Error.

Figure 1 Simulated probability for cervical cancer screening under various potential policy scenarios. Baseline cervical cancer screening program: the percentage of cervical cancer-related death reduction “20%”; screening location “county”; screening interval “every three years”; moderate screening pain; waiting time for screening results “3 months – 6 months”; out-of-pocket cost “300 CNY”.