Preferences of pregnant women toward a future maternal Group B Streptococcus vaccine in China: A cross-sectional survey with a discrete choice experiment

ABSTRACT

Group B streptococcus (GBS) is a major cause of neonatal death worldwide. A GBS vaccine for pregnant women is under development and is expected to be available in the near future. The perceptions and preferences of pregnant women in China of GBS vaccines has not been investigated, and this study aimed to investigate pregnant women‘s awareness of GBS and their potential preferences for the GBS vaccine. A discrete choice experiment was conducted among pregnant women in hospitals from Shaanxi, Hunan, and Zhejiang provinces located in Western, Central, and Eastern China, respectively. A conditional logit model was used to analyze the data and calculate willingness to pay values and choice probabilities of different GBS vaccine programs. A total of 354 pregnant women were included in the final analysis, 45.8% of whom were willing to receive a GBS vaccine if it were licensed. Vaccine safety was the most important attribute of a future vaccine, while cost was the least important attribute. Compared with no vaccination, pregnant women had a strong preference for future GBS vaccination (ASC = 1.267, p < .001). Pregnant women’s decisions were highly influenced by those of other pregnant women. Improving the safety, efficacy, and vaccination rate of the GBS vaccine in China is of great significance for future GBS vaccine development and vaccination. Compared to other variable options, the cost of a GBS vaccine was of the least importance among pregnant women in mainland China. These findings can inform public health policy decisions related to GBS vaccination in China.

KEYWORDS:

• Group B Streptococcus
• vaccine
• preference
• pregnant women
• discrete choice experiment

Introduction

Group B streptococcus (GBS) is a leading cause of neonatal death, causing meningitis, sepsis, and pneumonia, which may lead to long-term neurological sequelae, contributing to a large disease burden worldwide.^1,2 GBS disease in infants is divided into two types: early-onset GBS (EOGBS), which occurs within the first 7 days of life, and late-onset GBS (LOGBS), which occurs between 7 and 89 days of life.² Globally, GBS contributes to around 9,700 infant deaths annually¹ and the overall global incidence is 0.49 per 1,000 live births.² Our previous national multicenter study found that in China, the incidence of GBS disease was about 0.31 per 1,000 live births.³ The absolute number of Chinese infants suffering from GBS disease is 25,000 (UR 0–59,000), which ranks second in the world, following India.⁴

The current approach to preventing GBS involves identifying high-risk women (maternal fever during labor ≥38°C, unavoidable premature delivery, preterm premature rupture of membranes (PPROM), rupture of membranes ≥18 hours) during pregnancy and administering antibiotics during labor, also known as intrapartum antibiotic prophylaxis (IAP) use.⁴ The use of IAP in USA has reduced the incidence of GBS EOD, from 0.37 to 0.23 per 1,000 live births (37.8%),⁵ and the Perinatal Medicine Branch of Chinese Medical Association has recommended IAP since 2021.⁶ However, IAP cannot prevent LOGBS, and the use of antibiotics may influence the infant’s bacterial flora, leading to drug resistance, allergy, and other risks.⁷ Moreover, absence of adequate laboratory testing and identification resources poses a challenge for GBS screening and of implementation of the widespread use of IAP in China’s secondary and primary level healthcare facilities.⁸

Due to the fact that GBS screening and IAP cannot cover all pregnant women, maternal vaccination against GBS is crucial. The 6-valent GBS vaccine under development, which has been granted breakthrough therapy certification by the Food and Drug Administration, Priority Medicines by the European Medicine Administration,⁹ is expected to reduce drug resistance¹⁰ and protect newborns from both LOGBS and EOGBS.^11,12

It can be challenging for pregnant women to accept GBS vaccines which are offered during pregnancy, and the refusal rate among pregnant women can be as high as 32%.¹³ Pregnant Chinese women’s knowledge of GBS and their preferences for future GBS vaccination are unknown. Recent studies in the United States and Ireland suggest that pregnant women in countries with different incidence and preventive interventions may have different preferences for GBS vaccination.¹⁴ Gaining insights into how various attributes of vaccines influence individual preferences regarding vaccination can provide valuable information to public health authorities.

Discrete Choice Experiments（DCE）have been widely used in the field of vaccination preference research.¹⁵ DCE relies on individuals’ personal knowledge and perceptions to assess their preferences,¹⁶ considering trade-offs between different options. Compared to other survey methods, DCE is argued to better simulate real-world decision-making processes.¹⁷ Many DCE studies of vaccines which benefit children have been conducted by interviewing parents and doctors, and existing DCE vaccine studies include those for human papillomavirus (HPV),¹⁸ influenza,¹⁹ haemophilus influenzae type b,²⁰ and meningitis.²¹ The GBS vaccine, which would be administered to mothers during pregnancy, would confers benefits to newborns.^12,22

As such, in this study, we applied a DCE to evaluate pregnant Chinese women’s preferences regarding the attributes and characteristics of an upcoming GBS vaccine. By examining these preferences, we aimed to identify the factors that influence their vaccination choices and determine which attributes should be considered for future information and awareness campaigns, and in the integration of the GBS vaccine into the national immunization program.

Methods

We applied the DCE method following the research practice guideline established by the International Society for Pharmacoeconomics and Outcomes Research (ISPOR).^23–25

Attributes, levels, and DCE development

In DCE studies on vaccine preference, it is necessary to identify the attributes that have an impact on vaccination preference. A literature review of vaccine-related DCE was conducted using keywords including “discrete choice experiment,” “vaccine,” and “preference,” through Pubmed/Medline, Web of Science, Cochrane, CNKI, and WanFang databases, and nine attributes were extracted from the literature to be included in this study: vaccine efficacy, vaccine safety, out-of-pocket cost, proximity of vaccination location to home, whether the vaccine is imported, vaccine duration of protection, vaccine recommenders, proportion of the pregnant women vaccinated. A single retrieved article explored future GBS vaccination preferences, in both Philadelphia, USA and Dublin, Ireland.¹⁴ A review by Marshall et al.²⁶ estimated that 70% of studies use between 3 and 7 attributes, and a review by Marilyn et al.²⁷ found that literature review and focus interviews were the most common methods for identifying attributes.

The final attributes and levels of the survey were determined through online one-on-one semi-structured focus interviews with a stakeholder group consisting of 3 obstetricians, 2 gynecologists, 2 pediatricians, 6 pregnant women, and 11 vaccine developers. We asked respondents to rate the importance of the 9 attributes extracted from the literature, with the most important being 9 and the least important being 1. We summarized the scores of each attribute and selected the 6 attributes with the highest scores for inclusion in the study. Six attributes were selected which were the most important of the 9 discussed with participants: efficacy, safety (Adverse Vaccine Event (AVE)), duration of protection, out-of-pocket cost of single dose, total number of doses, and proportion of pregnant women who have been vaccinated (Table 1). The maximum level (or different categories) of cost was determined using the price of the ACYW135 meningococcal polysaccharide conjugate vaccine (CRM197 vector)²⁸ because of their similarities in R&D principles. The level of other attributes was determined by referring to previous DCE studies, the Guiding Principles of the National Medical Products Administration of China on the Issuance of Grading Standards for Adverse Events in Clinical Trials of Vaccines for Prophylactic Use,²⁹ and the ISPOR Task Force^23–25 on DCE.

Table 1. Attributes and levels for GBS vaccine discrete choice experiment.

Vaccine attributes^a	Levels
Safety (Adverse Vaccine Events)	Minor AVE, Moderate AVE, Severe AVE
Efficacy^b	65%, 80%, 95%
Duration of effect^c	1 year, 10 years, 30 years
Total number of doses	1 dose, 2 doses, 3 doses
Proporation of pregnant women vaccinated	30%, 60%, 90%
Out-of-pocket cost	0 USD^d, 31.16 USD, 62.31USD

^aSee Appendix 1 for complete description of attributes and levels.

^bBased on WHO Preferred Product Characteristics for Group B Streptococcus Vaccines.

^cBased on assumptions since the vaccine is not licensed.

^dUSD: US dollars.

Our selection of attributes and levels resulted in a total of 729 hypothetical scenarios (calculated by multiplying 3 options for each of the 6 attributes). Additionally, there were 265,356 potential paired choices (calculated using the formula (729 × 728)/2) that could be generated from these scenarios. We used SAS software (version 9.4; Cary, NC) D-optimal algorithm to generate choice sets, and 18 was the minimum number of choice sets to complete a 100% efficient design. To reduce the cognitive burden on respondents, 18 choice sets were divided into two versions, each consisting of 9 unique choices (example listed in Table 2). Within the survey, each set of choice questions was randomly assigned to half of the sample for distribution.

Table 2. An example of choice sets.

	Program 1	Program 2	Opt-out
Safety (AVE)	Moderate	Minor	Neither
Efficacy	80%	65%
Duration of effect	10 years	30 years
Total number of doses	1 dose	3 doses
Proporation of pregnant women vaccinated	90%	30%
Out-of-pocket cost	0 USD^a	62.31 USD
Pregnant women’s choice

^aUSD: US dollars.

Study sample and data collection

Using stratified random sampling, based on the regional divisions outlined in the China Health Statistics Yearbook,³⁰ 1 city was selected from each of the eastern (Hangzhou, Zhejiang), western (Xi’an, Shannxi), and central regions (Changsha, Hunan) of China. According to the empirical formula proposed by Johnson and Orme, the number of required samples can be calculated using the equation: n > 500c/ta.³¹

When only the main effect exists, c represents the maximum level value in the attribute. When the second-order interaction effect between attributes is considered, c represents the maximum number of combinations of any two attributes; t represents the number of questions in the questionnaire; and a represents the number of options in each question (excluding the opt-out). The final sample size was determined to be at least n > 500 * 3 * 3/9 * 2 = 250.

The inclusion criteria for subjects in this study were as follows: pregnant women who go to the hospital for routine prenatal checkups, and willing to undergo a discrete choice experiment. Exclusion criteria for the study included: unwillingness to participate in the preference study, incomplete informed consent, and inability to understand the content of the study. From March to April 2023, the survey was conducted anonymously in quiet fetal monitor rooms in Xi ‘an Fourth Hospital, Xi ‘an International Medical Center, Zhejiang University Maternity Hospital, and Changsha Maternal and Child Health Hospital. The questionnaire contained demographic information, GBS vaccination intention question (whether the pregnant women are willing to receive GBS vaccine), GBS knowledge questions (based on the subjective feelings of pregnant women), relative importance ranking of GBS vaccine attributes, and 9 choice sets per participant. Each choice set required subjects to choose among 2 vaccine programs or 1 opt-out option. In the discrete choice experiment, 2 trained researchers from Xi ‘an Jiaotong University carried computer tablets to the site to complete a one-on-one interview. Researchers introduced GBS, the concept of a GBS vaccine, and the meaning of attributes (see Appendix 1), before they started the interview. In order to control the quality of the study, one trap question (repeated question) was set for each version of the questionnaire, and the quality of the questionnaire was controlled through the consistency of the answers. All data from participants who provided inconsistent responses to trap questions was removed from the study analysis.

Data analyses

EpiData (EpiData Association, Odense, Denmark) was used for double data entry, and except for the price level, which was treated as a continuous variable,^32,33 the remaining attributes were encoded using dummy variables. Descriptive analyses were used to examine the demographic characteristics of the participants in the study. To determine the influence of each attribute, a conditional logit model was employed in R (See Equation (1)(1) $Un=Vn+\epsilon =ASC+\beta 1\times Moderat{e}_{ADE}+\beta 2\times Sever{e}_{ADE}+\beta 3\times Efficac{y}_{80}\mathrm{\%}+\beta 4\times Efficac{y}_{95}\mathrm{\%}+\beta 5\times Protectio{n}_{duration{10}_{years}}+\beta 6\times Protectio{n}_{duration{30}_{years}}+\beta 7\times dose2+\beta 8\times dose3+\beta 9\times coverage60\mathrm{\%}+\beta 10\times coverage90\mathrm{\%}+\beta 11\times cost+\epsilon$(1) ).

(1) $Un=Vn+\epsilon =ASC+\beta 1\times Moderat{e}_{ADE}+\beta 2\times Sever{e}_{ADE}+\beta 3\times Efficac{y}_{80}\mathrm{\%}+\beta 4\times Efficac{y}_{95}\mathrm{\%}+\beta 5\times Protectio{n}_{duration{10}_{years}}+\beta 6\times Protectio{n}_{duration{30}_{years}}+\beta 7\times dose2+\beta 8\times dose3+\beta 9\times coverage60\mathrm{\%}+\beta 10\times coverage90\mathrm{\%}+\beta 11\times cost+\epsilon$(1)

The utility ($Un$) of different GBS vaccination programs is determined by the deterministic utility of 6 observed vaccine attributes ($Vn$), along with a random error term ($\epsilon$). The Alternative Specific Constant (ASC) is a randomly determined value that reflects the preference individuals have for vaccination options when compared to the choice of not getting vaccinated (opt-out). Preferences $\beta 1$ to $\beta 11$ represent the preference co-efficiency of each attribute level compared to the reference level, with the sign of coefficients indicating the positive or negative effect on utility. The goodness of fit of the final model was assessed using the Akaike Information Criterion (AIC) and the Bayesian Information Criterion (BIC).

Policy analysis

We assessed the participants’ willingness to pay (WTP) for variations in attribute levels by calculating the ratio between the coefficients of categorical variables and the coefficient associated with the cost of vaccination. To illustrate, when all other attribute levels remained constant, the ratio β1/β11 indicated the parents’ WTP for enhancing vaccine effectiveness from 65% to 85%.³³

We performed policy simulations aimed at determining the likelihood of individuals favoring a specific vaccination program with certain characteristics. This estimation was derived by employing the following equation:

(2) $P_i=\frac{e^{U_i}}{{\sum }_j{e}^{U_j}}$(2)

$P_i$ represents the likelihood of selecting a specific vaccination strategy (alternative) among various options available, $U_i$ refers to the measure of overall satisfaction, benefit, or value derived from choosing alternative “i” in the vaccination program, and $U_j$ represents a combination of both receiving vaccinations and not receiving them.

Results

Socio-demographic characteristics of the respondents

Table 3 shows the demographics of the study respondents. A total of 420 pregnant women participated in the survey, including 140 from Zhejiang province, 150 from Shaanxi province, and 130 from Hunan Province. Excluding the pregnant women with inconsistent choices in the trap questions, a total of 354 samples were collected (116 from Zhejiang province, 130 from Shaanxi Province, and 108 from Hunan Province). The average age of the respondents was 30.9 (SD = 3.77) years old, the average gestational weeks of the respondents were 35.86 (SD = 2.42) weeks. 99.72% of the respondents were married, 89.83% of them have a bachelor or above degree, and 42.37% of them are enterprise employees. 7.91% of the respondents had an occupation related to health care, 88.7% of them are from urban areas, and 45.76% of the respondents were willing to be vaccinated against GBS (this question was asked prior to the DCE portion of the survey). Only 22.32% of the respondents were very familiar with GBS, and 46.9% of respondents reported having a prior pregnancy. 75.99% of the respondents have an annual household income of more than 100,000 CNY.

Table 3. Demographic characteristics.

	Number of people who have passed the conformance test (N = 354）
Marital status	N	%
Married	353	99.72
Unmarried	1	0.28
Divorced	0	0
Education
Master’s degree or above	64	18.08
Bachelor’s degree	254	71.75
High school	31	8.76
Middle school	5	1.41
Occupation
Enterprise employee	150	42.37
Employee of public institution	64	18.08
Unemployed	35	9.89
Self-employed	22	6.21
Civil servant	9	2.54
Student	2	0.56
Other	72	20.34
Medical-related career
Yes	28	7.91
No	326	92.09
Place of residence
Urban	314	88.7
Rural	40	11.3
Annual household income
<50,000 CNY^a	18	5.08
50,000–100,000 CNY	67	18.93
100,000–150,000 CNY	104	29.38
150,000–200,000 CNY	70	19.77
>200,000 CNY	95	26.84
GBS knowledge
Very familiar with	8	2.26
Moderately familiar	56	15.82
Heard of but not familiar with	211	59.6
Completely unheard of	79	22.32
Pregnancy history
Yes	166	46.89
No	188	53.11
Number of children already born
0	236	66.67
1	106	29.94
2	11	3.11
3	1	0.28
Vaccination willingness
Yes	162	45.76
No	18	5.08
Uncertain	174	49.15
Geographical location
Eastern	130	36.72
Central	108	30.51
Western	116	32.77
Age (mean ± SD)	30.90 ± 3.77
Gestational weeks (mean ± SD)	35.86 ± 2.42

^aCNY: Chinese Yuan (Exchange rate:6.74 according to World Bank 2022).

Preferences of pregnant women toward GBS vaccination and subgroup analysis

The outcomes of the conditional logit model for two different scenarios are presented in Table 4. Model 1 displays the results for the main effects only, while Model 2 includes the main effects along with interaction effects. Significant differences (p < .001) were observed in participants’ preference for GBS vaccination based on vaccine versus no vaccine, minor versus moderate/severe adverse events following vaccination, efficacy of 95% versus 65%, number of doses (1 versus 3), and vaccine coverage of 90% versus 30%. The preference coefficients show that pregnant women prefer GBS vaccines with low risk of adverse events, high efficacy, 1 dose, and high pregnant population coverage. Pregnant women surveyed showed a slight positive preference for the price of GBS vaccine: 0.007/10USD, but this preference was not statistically significant. Compared with no vaccination, pregnant women had a strong preference for future GBS vaccination (coefficient of preference = 1.267, p < .001). The safety of a vaccine (avoiding severe AVE) was the most important attribute (coefficient of preference = −1.707, p < .001), the second most important attribute is the vaccination rate among pregnant women (90%) (coefficient of preference = 0.249, p < .001), while vaccination cost was the least significant attribute (coefficient of preference = 0.007, p = .503).

Table 4. Results of conditional logit model with main effects and interactions.

Attributes and levels	Regression coefficient	exp(coef)/OR	SE	Regression coefficient	exp(coef)/OR	SE
	Model 1			Model2
ASC	1.267***	3.552	0.079	1.100***	3.004	0.074
Safety (Reference level: minor AVE)
Moderate AVE	−0.976***	0.377	0.066	−0.469***	0.626	0.064
Severe AVE	−1.707***	0.181	0.066	−0.943***	0.390	0.063
Efficacy (Reference level: 65%)
80%	0.120	1.127	0.063	−0.061	0.941	0.061
95%	0.191**	1.211	0.062	0.020	1.020	0.060
Protection duration (Reference level: 1 year)
10 years	0.126*	1.134	0.064	0.010	1.010	0.062
30 years	−0.027	0.973	0.065	−0.096	0.908	0.063
The number of doses (Reference level: 1 dose)
2 doses	−0.242***	0.785	0.061	−0.340***	0.711	0.059
3 doses	−0.191**	0.826	0.061	−0.296***	0.744	0.059
Pregnant women vaccination rate (Reference level: 30%)
60%	−0.164*	0.849	0.067	−0.048	0.954	0.065
90%	0.249***	1.283	0.064	0.261***	1.298	0.062
Out-of-pocket Cost (Unit: 10 USD)	0.007	1.001	0.001	0.025*	1.003	0.001
ASC × bachelor_degree_or_above				−0.106	0.899	0.062
Moderate_AVE × bachelor_degree_or_above				−0.399***	0.671	0.065
Severe_AVE × bachelor_degree_or_above				−0.533***	0.587	0.065
80% × bachelor_degree_or_above				0.162*	1.176	0.063
95% × bachelor_degree_or_above				0.135*	1.145	0.061
10 year × bachelor_degree_or_above				0.102	1.107	0.063
30 year × bachelor_degree_or_above				0.071	1.073	0.064
2 dose × bachelor_degree_or_above				0.159**	1.173	0.061
3 dose × bachelor_degree_or_above				0.159**	1.172	0.061
60% × bachelor_degree_or_above				−0.096	0.909	0.066
90% × bachelor_degree_or_above				−0.065	0.937	0.064
Out-of-pocket Cost (Unit: 10 USD) × bachelor_degree_or_above				−0.021*	0.998	0.001
Logarithmic likelihood value	−2884.973			−3044.802
AIC value	5793.945			6139.604
BIC value	5864.977			6287.588
Sample size	354			354
Observed value	6372			6372

*p < .05, **p < .01, ***p < .001, ASC: Alternative specific constant, SE: standard error, AIC: Akaike Information Criterion, BIC: Bayesian Information Criterion, ASC: Alternative Specific Constant. ×interaction terms.

When considering demographic interactions, it becomes evident that pregnant women from varying household income backgrounds exhibit no difference in their preferences for GBS vaccines. Women with higher levels of education exhibited a greater capacity to endure minor adverse events after receiving immunizations and demonstrated a greater acceptance of receiving multiple doses of the vaccine compared to those with lower levels of education. Additionally, more educated women expressed a dispreference regarding the increased cost associated with the vaccine.

Policy analysis

Trade-offs among attributes

The marginal willingness to pay (WTP) values for each attribute associated with GBS vaccination are presented in Table 5. Pregnant women exhibited a pronounced inclination toward greater vaccine efficacy and demonstrated a willingness to allocate 257.26 USD for a GBS vaccine with 95% efficacy as opposed to 65%. Moreover, pregnant women indicated a monetary valuation of 2,293.71 USD in order to mitigate the occurrence of severe adverse drug events (AVEs).

Table 5. Willingness to pay.

Base Level	Comparison Level	Estimated WTP（USD）
No vaccine	vaccine	1703.55
Minor AVE	Moderate AVE	1311.94
	Severe AVE	2293.71
65% efficacy	80% efficacy	160.64
	95% efficacy	257.26
1 year protection	10 years protection	169.33
	30 years protection	36.13
1 dose	2 dose	325.13
	3 dose	257.38
0% coverage	60% coverage	220.46
	90% coverage	334.95

In terms of the relative importance, the most crucial attribute is vaccine efficacy, with a score of 27.29%, whereas total number of doses ranked as the least important attribute at 9.94%. There were no notable disparities in the relative importance of vaccine attributes among women residing in different regions of China. The results also indicate that the preferences observed in the direct relative importance ranking are distinct from those observed in the DCE, implying a disparity between the two methods in terms of exhibited preferences.

Choice probability of vaccine program

The base case vaccine option has Moderate AVE, costs 31.16 USD and has an efficacy of 65%. It provides protection for 1 year with a single dose and has a coverage rate of 30%. The choice probability for this vaccine is 27.8%. In Figure 1, we analyze the change in choice probability compared to the base case when modifying vaccine attributes. When the vaccine’s AVE changes from moderate to minor, the choice probability increases by 22.8%, while increasing efficacy from 65% to 95% leads to a 4.8% increase in choice probability. For the vaccination programs with the following attributes: minor AVE, 31.16 USD, 95% efficacy, 10-year protection, 1 dose, and 90% coverage, the choice probability was found to be 64.4%.

Figure 1. Probability of choice compared to base case (moderate AVE, 31.16USD, 65% 263 efficacy, 1 year protection, 1 dose, 30% coverage).

Discussion

To the best of our knowledge, this research represents the inaugural attempt in mainland China to examine pregnant women’s preferences regarding possible GBS vaccination through DCE methodology. Although a GBS vaccine is not yet available on the market, substantial progress has been made in its research and development. It is crucial to evaluate the preferences for GBS vaccination among pregnant women to optimize vaccine marketing strategies and vaccination campaigns, and thereby safeguarding newborns against GBS infection.

Our primary findings reveal that the most critical factor considered by pregnant women in mainland China was the vaccine’s safety, specifically in preventing severe adverse effects (AVE), the cost of vaccination was the attribute with the least impact on their decision-making, and pregnant women residing in mainland China exhibit a preference for a GBS vaccines that possess the following attributes: minor risk of adverse events, high efficacy, 1 dose, and high pregnant population coverage. In previous studies, respondents have shown a negative preference for vaccine price increases.^32,34,35 In our study, however, the coefficient of preference for vaccine price is positive, which means that pregnant women are willing to take the vaccine at a higher price (though this relationship was not statistically significant), which is consistent with a previous study of influenza vaccine on the elderly.³⁶ The reason for this finding may be that pregnant women and the elderly are more concerned about the benefits and risks of vaccination and are relatively insensitive to price, but more research is needed to confirm this. Notably, in our study we found that pregnant women’s decisions were highly influenced by those of other pregnant women. When other pregnant women decided to get a GBS vaccine (assuming 90% coverage), pregnant women showed a clear preference for receiving vaccine. This may be explained by the fact that pregnant women believe that vaccines with high coverage tend to reflect public confidence in the good quality of vaccines, which is consistent with previous studies.^37,38 Furthermore, the results suggest that there are variations in preferences regarding GBS vaccination among pregnant women with different levels of education, with more highly educated women preferring minor adverse events following immunization and demonstrated a greater acceptance of receiving multiple doses. The vaccination intention rate among the cohort of pregnant women in this study was approximately 45.76%, which is comparatively low compared with the rates observed in developed countries like the United States (68%),³⁹ Australia (74%),¹³ and the United Kingdom (68.4%).^40,41 We also found that the results of the DCE were different from relative importance ranking: coverage of GBS vaccine is considered to be the least important attribute in questions of relative importance ranking, but in DCE results, whether other pregnant women have received GBS vaccine is considered to be the second important attribute. These insights are valuable in informing the development and introduction of future GBS vaccines in China. By considering the preferences and priorities of pregnant women, researchers and policymakers can focus on developing vaccines that meet these criteria and develop vaccination campaigns which highlight key attributes. This information can aid in guiding public health initiatives and policy decisions related to GBS vaccination in China, ultimately contributing to the reduction of GBS-related complications and improving maternal and neonatal health outcomes.

This study had several notable strengths. First, it involved conducting face-to-face interviews by trained researchers with pregnant women. When the interviewees are unable to comprehend the DCE questions, the researchers can clarify the questions. Second, the inclusion of an opt-out option within the choice set allowed participants to refrain from making a challenging decision, thereby enhancing the validity of their responses. The third strength was the incorporation of repeated “trap questions” within the choice set, enabling the identification of pregnant women who may not have fully grasped the purpose of the DCE design, with that data being removed from analysis. However, it is important to acknowledge that this approach resulted in a reduction in the sample size.

Our study had some limitations. Screening for GBS in China is more well established in tertiary hospitals, whereas secondary and primary healthcare facilities may face limitations due to inadequate laboratory resources.⁸ Consequently, pregnant women who receive obstetric care in tertiary hospitals are more likely to be informed about the risks associated with GBS. As a result, we specifically chose pregnant women from tertiary hospitals in major cities as our sample, and this sample may not accurately represent the preferences of pregnant women residing in rural areas. As well, since a GBS vaccine has not yet been approved, our study was conducted with limited information regarding potential vaccine adverse events. Furthermore, since participants may opt out of answering some questions, not only to genuinely express their real-life preferences but also to evade challenging trade-offs at attribute levels, the accuracy of parameter estimates may have been impacted.^42,43

Conclusions

In conclusion, the findings of this study indicate that pregnant women are more inclined toward receiving the GBS vaccination compared to opting for no vaccination, and they strongly favor a future GBS vaccine with certain key attributes: a low risk of adverse events, high efficacy, 1 dose administration, and the adoption of the vaccine by a significant portion of the pregnant population. Women with a higher educational attainment demonstrated an increased tolerance in minor adverse events following immunization and exhibited a greater willingness to receive multiple doses of the GBS vaccine, in comparison to pregnant women with lower levels of education. Moreover, pregnant women with higher educational backgrounds expressed an elevated level of apprehension regarding the increased cost of the vaccine.

Author contributions

W.J, J.D, and Y.F conceived the original idea and design the study. J.D conducted the DCE design, DCE interview, data analysis and manuscript drafting. H.Z helped with DCE interview. Y.Z, J.C and D.M provided critical review and edits. W.J obtained the funding. All authors have review and approved the final manuscript.

Ethical approval

As this study is a cross-sectional survey which does not involve patient privacy, poses no harm to patient well-being, and ensures complete anonymity, it has been exempted from ethical review. Informed consent was obtained from the participants of this study.

Acknowledgments

We thank the pregnant women and researchers from Xi ‘an Fourth Hospital, Xi ‘an International Medical Center, Zhejiang University Maternity Hospital, and Changsha Maternal and Child Health Hospital for their cooperation and support in this study.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Data availability statement

Datasets are available from the corresponding author on reasonable request.

Supplementary data

Supplemental data for this article can be accessed on the publisher’s website at https://doi.org/10.1080/21645515.2023.2281713.

Additional information

Funding

The work was funded by the National Natural Science Foundation of China [grant number 82204109], China Postdoctoral Science Foundation [grant number 2020M673427] and the Key Research and Development Plan of Shaanxi Province [grant number 2022SF-106]. All funders have no role in the study design, data collection and analysis, manuscript writing, and the decision to submit the manuscript for publication.

Appendix

Before DCE begins, we showed the interviewees the specific meanings of the different attributes in DCE.

Table1. Attributes affecting the vaccination of pregnant women with Group B Streptococcus vaccine and their definitions.

Attributes	Attributes meaning	Level	Level meaning
Safety (AVE)	The severity of adverse events following vaccination, such as localized pain at the injection site, fatigue, and headache.	mild	Mild: Brief (less than 48 hours) or slight discomfort, does not affect daily activities, no treatment needed.
		moderate	Moderate: Mild to moderate activity limitation, may require medical consultation, may not require treatment, or may only require mild treatment.
		severe	Severe: Significant impairment of activity, requiring medical consultation and treatment, possibly necessitating hospitalization.
Efficacy	The protection rate of the vaccine against neonatal Group B Streptococcus infection	65%	After pregnant women receive GBS vaccine, 65% of newborns will no longer be infected with GBS
		80%	After pregnant women receive GBS vaccine, 80% of newborns will no longer be infected with GBS
		95%	After pregnant women receive GBS vaccine, 95% of newborns will no longer be infected with GBS
Duration of effect	Duration of vaccine protection	1 year	\
		10 years
		30 years
Out-of-pocket cost	Out-of-pocket cost of 1 single dose	0 USD
		31.16 USD
		62.31 USD
Total number of doses	\	1 dose
		2 doses
		3 doses
Proportion of pregnant women vaccinated	The proportion of pregnant women who have received the Group B Streptococcus vaccine	25%
		50%
		100%