- Email: [email protected]
A conjoint analysis of stated vaccine preferences in Shanghai, China
Vaccine xxx (xxxx) xxx
Contents lists available at ScienceDirect
Vaccine journal homepage: www.elsevier.com/locate/vaccine
A conjoint analysis of stated vaccine preferences in Shanghai, China Xiaodong Sun a, Abram L. Wagner b,⇑, Jiahui Ji b, Zhuoying Huang a, Brian J. Zikmund-Fisher c,d, Matthew L. Boulton b,e, Jia Ren a, Lisa A. Prosser f a
Department of Immunization Program, Shanghai Municipal Centers for Disease Control & Prevention, NO. 1380, West Zhongshan Road, 200336, Shanghai, China Department of Epidemiology, School of Public Health, University of Michigan, 1415 Washington Heights, Ann Arbor, MI 48109, USA c Department of Health Behavior & Health Education, School of Public Health, University of Michigan, 1415 Washington Heights, Ann Arbor, MI 48109, USA d Department of Internal Medicine, Division of General Medicine, University of Michigan Medical School, 1500 East Medical Center Drive, Ann Arbor, MI 48109, USA e Department of Internal Medicine, Division of Infectious Disease, University of Michigan Medical School, 1500 East Medical Center Drive, Ann Arbor, MI 48109, USA f Susan B. Meister Child Health Evaluation and Research Center, Department of Pediatrics, University of Michigan Medical School, Ann Arbor, MI, USA b
a r t i c l e
i n f o
Article history: Received 6 July 2019 Received in revised form 8 November 2019 Accepted 25 November 2019 Available online xxxx Keywords: China Conjoint analysis Discrete choice experiment Vaccine manufacturers
a b s t r a c t It is not clear what kind of preferences parents in China would have for vaccines that could be added to a future immunization schedule. This study’s aim was to assess Chinese parents’ preferences for attributes of vaccines. We surveyed parents of young infants 3 months of age at immunization clinics in Shanghai, China, in 2017. We used a discrete choice experiment (DCE) to present parents with choices between two hypothetical profiles of vaccines which were described using the following attributes: cost, risk of side effect, location of vaccine manufacturer, vaccine testing, vaccine effectiveness, severity of disease, disease prevalence. A logistic regression output estimates of preference utilities. In total, 599 caregivers completed the DCE. Parents expressed lower preference for vaccines with a 30% chance of fever as an adverse event vs a 10% chance (OR: 0.53, 95% CI: 0.44, 0.64), for vaccines only 85% effective vs those 95% effective (OR: 0.55, 95% CI: 0.48, 0.62), and for imported vaccines (OR: 0.74, 95% CI: 0.60, 0.92) and those not tested in Chinese children (OR: 0.45, 95% CI: 0.37, 0.53) compared to domestic vaccines. More affluent groups preferred more expensive vaccines whereas less affluent groups did not express cost-based preferences. Promotion of vaccines in China should focus on parents’ stated preferences, which include past testing done in Chinese children – which is, in fact, required of all licensed vaccines in China. Information about these trials could emphasize low risk of adverse events and high effectiveness. Ó 2019 Elsevier Ltd. All rights reserved.
1. Introduction Vaccines protect children against a number of diseases with the potential to cause serious illness and death. In China, the government-funded Expanded Program on Immunization (EPI) has been providing free and mandatory vaccines for tuberculosis, polio, measles, pertussis, diphtheria, tetanus, and others since 1978 [1]. All EPI vaccines in China are currently manufactured by pharmaceutical companies based in China [2]. The number of vaccines included on the EPI in China is fewer than recommended by the World Health Organization (WHO) [3], and is therefore likely to increase in the future.
⇑ Corresponding author at: 1415 Washington Heights, Ann Arbor, MI 48109, USA. E-mail addresses: [email protected] (X. Sun), [email protected] (A.L. Wagner), [email protected] (J. Ji), [email protected] (Z. Huang), [email protected] (B.J. Zikmund-Fisher), [email protected] (M.L. Boulton), [email protected] (J. Ren), [email protected] (L.A. Prosser).
In addition to free EPI vaccines, Chinese clinics offer some forfee non-EPI vaccines, which are recommended by the World Health Organization including Haemophilus influenzae type b vaccine (Hib), pneumococcal conjugate vaccine (PCV), and rotavirus vaccine. Those vaccines could help reduce childhood morbidity in China, and many analyses have posited that these vaccines would be cost-effective [4]. In China, the pneumococcal and Hib vaccines, in particular, are prime candidates for inclusion on the EPI schedule given the substantial burden of pneumonia in China [2]; 26.9% of children 2–8 years old in Beijing, China, have ever been diagnosed with pneumonia [5]. Nationwide, these two vaccines have the potential to prevent a significant portion of the 11,000 deaths due to pneumococcal pneumonia and meningitis [6] and 19,000 deaths from Hib pneumonia and meningitis in Chinese children under 5 years of age annually [7]. From 2003 to 2012, a total of 127,539 deaths from diarrhea were reported among Chinese children <5 years of age, of which an estimated 53,559 (42%) had illness attributable to rotavirus [8]. Although more widespread use of these vaccines could clearly reduce childhood morbidity
https://doi.org/10.1016/j.vaccine.2019.11.062 0264-410X/Ó 2019 Elsevier Ltd. All rights reserved.
Please cite this article as: X. Sun, A. L. Wagner, J. Ji et al., A conjoint analysis of stated vaccine preferences in Shanghai, China, Vaccine, https://doi.org/ 10.1016/j.vaccine.2019.11.062
2
X. Sun et al. / Vaccine xxx (xxxx) xxx
and mortality in China, the evidence base on parental vaccine decision-making needed to inform efforts to increase uptake of vaccines which are both optional and require a fee, is currently lacking. An additional consideration is that vaccines in China can be produced by domestic or foreign companies, and parents for some vaccines can make a choice based on manufacturer’s location. Although parental perceptions of these manufacturers may vary [9], how this affects their stated preferences is unclear. Vaccination behaviors in China have been shown to vary based on parental characteristics, like residency, urbanicity [10], and education [11]. Residency refers to where a family is officially registered (the specific city or village in the country, but also whether this is a rural or urban location). Some local governmental services are limited to individuals with official residence in that area [12], although EPI vaccines are provided to children regardless of whether they are local or non-locals (sometimes referred to as the ‘‘floating population”). Most (88%) of non-local children will attend government-run immunization clinics [13]. Residency differs from urbanicity, which is where an individual family currently lives. Shanghai only consists of urban and suburban areas, and the latter group can be divided into inner and outer suburbs based on their proximity to historically more economically developed areas. There are also external factors—attributes of vaccines and vaccination programs (like cost, safety of vaccine, and manufacturer characteristics), that are associated with people’s behaviors. However, highlighting vaccine attributes or certain programmatic characteristics could enhance vaccination behaviors. The government and health organizations could benefit from a greater understanding of those attributes which are favorably perceived by the public that would, in turn, permit tailoring of vaccines and vaccination programs to improve vaccine acceptance and use. Preferences for attributes can be measured through conjoint analysis such as discrete choice experiments (DCEs) [14]. DCE is a survey technique that has been used to understand parent preference toward vaccination in the US and other countries. DCEs, as a method of conjoint analysis, have been widely used to quantify values of public on health-related choices [14]. This approach has the advantage of allowing for several attributes to be evaluated simultaneously in the choice scenario. The respondents need to consider which profile provides more value to them based on all the attributes in a given choice set, which will necessarily involve trade-offs among attributes in the decision making process. For example, a previous study in the US looked at attributes like the risk of fever after vaccination, the cost of visit, community vaccination coverage, and doses required to complete the vaccine series [15]. The choice of attributes should be driven by local circumstances. In a previously published paper of this study population, over two-thirds expressed concerns about vaccine safety [16]. The objective of this study is to estimate preferences of Chinese parents for attributes of EPI and non-EPI pediatric vaccines, and to calculate their willingness to pay for certain attributes.
2. Methods 2.1. Study population This cross-sectional study enrolled a convenience sample of caregivers of infants <3 months at public immunization clinics in Shanghai, China, between May and September 2017. The sampling of clinics was tied to the population of the surrounding township, according to a probability-proportionate-to-size (PPS) method. The two inclusion criteria were as follows: study participants had to be (1) a parent or grandparent of an infant <3 months of age, and (2) at least 18 years old.
Participants completed paper questionnaires in a private room after their child had received a vaccine, and the answers to these were input into an electronic database. The questionnaire included non-invasive questions about the participants’ perceptions of vaccines, including their valuation of the attributes of vaccines in a conjoint analysis (in which parents were asked to choose between two hypothetical profiles of vaccines) (Fig. 1). Participants completed one of two versions of a questionnaire, one version, the focus of this manuscript, asked participants about vaccine attributes; the other asked about vaccination program attributes, which is addressed in a different paper. A rule of thumb for the sample size of conjoint analysis is the following equation [17]:
N>
500c ta
where N is the sample size, c is the largest number of attribute levels (4), t is the number of choice tasks (8, from the other discrete choice experiment), and a is the number of alternatives (2). This leads a sample size of 250 within a simple random sample. Accounting for a design effect of 1.6384 (estimated from an intracluster coefficient of 0.0456 from a previous study on vaccine attitudes [10,11], and with 15 individuals in each cluster), the required sample size would be 410. The dataset is publicly available at a figshare repository (https://doi.org/10.6084/m9.figshare.6463304). 2.2. Attributes and survey design The list of attributes is found in Table 1. Attributes included the following: cost per vaccine, risk of getting fever after vaccination, vaccine manufacturer, vaccine effectiveness, severity of disease, and how common the disease is among children without vaccination. These attributes were identified through a qualitative study undertaken in the summer of 2016 and published elsewhere [18]. The number of attributes we included (7) is similar to numbers included in past studies [19]. For ease of analysis, we combined vaccine manufacturer and vaccine testing into one variable with three categories: Chinesemade vaccine, foreign vaccine tested in Chinese children, and foreign vaccine not tested in Chinese children. All vaccines marketed in China must be tested in Chinese children, but we included this characteristic because it arose as a concern in qualitative interviews. We used a fractional factorial design, and gave participants a forced choice (i.e., respondents had to choose one of the two profiles with no opt-out). 2.3. Statistical analysis We used conditional multivariable logistic regression to measure all the discrete choices and evaluate each individual attribute. Observations with missing values were deleted from the analysis. The analysis used survey procedures clustered at the levels of individual and township immunization clinic. To further explore preferences related to cost, we ran several additional models, with only cost, a socioeconomic characteristic, and an interaction term between these two variables, as explanatory variables. Significance was assessed at an alpha = 0.05 level. All analyses were conducted in SAS version 9.4 (SAS Institute, Cary, NC, USA). 2.4. Ethical approval The research protocol was approved by the Health Sciences and Behavioral Sciences Institutional Review Board at the University of
Please cite this article as: X. Sun, A. L. Wagner, J. Ji et al., A conjoint analysis of stated vaccine preferences in Shanghai, China, Vaccine, https://doi.org/ 10.1016/j.vaccine.2019.11.062
X. Sun et al. / Vaccine xxx (xxxx) xxx
3
Fig. 1. Example choice task given to participants.
Michigan (#HUM00125379) and the ethical review committee at the Shanghai Centers for Disease Control and Prevention (#2017–2). All participants provided oral informed consent. After
completing the interview, participants were given a small gift (such as an umbrella or blanket) to compensate for time spend on completing the survey.
Table 1 Attributes and attribute levels in a discrete choice experiment of vaccine attributes.
3. Results
Attribute
Levels
Cost per vaccine
100 RMB 400 RMB 800 RMB 1200 RMB 10% 30% Chinese Foreign Tested in Chinese children Not tested in Chinese children 95% 85% Requires hospitalization for two days No hospitalization, child home sick for two days 10% 1%
Risk of getting fever after office visit Vaccine manufacturer Vaccine testing Vaccine effectiveness Severity of disease
How common is the disease among children without vaccination
In total, 1,285 individuals were asked to participate, and 1,188 (92.5%) agreed and consented to participate. Study participants were randomized into two groups with 599 in the vaccine attributes study presented in this manuscript. The demographic characteristics of study participants are found in Table 2. Most caregivers (408, 70%) were mothers, with fewer fathers (167, 29%) or grandparents (8, 1%). Caregivers were sampled from urban areas (30%), inner suburbs (50%), and outer suburbs (20%) of Shanghai (See Table 3). A total of 552 participants had complete data for the multivariable regression model of stated preferences. Caregivers expressed a strong preference for vaccines with lower risk of side effects (30% risk vs 10% risk OR: 0.53, 95% CI: 0.44, 0.64) and that were more effective (85% effective vs 95% effective OR: 0.55, 95% CI: 0.48, 0.62). Parents expressed a preference for a vaccine that protected against a more common disease, although this relationship was
Please cite this article as: X. Sun, A. L. Wagner, J. Ji et al., A conjoint analysis of stated vaccine preferences in Shanghai, China, Vaccine, https://doi.org/ 10.1016/j.vaccine.2019.11.062
4
X. Sun et al. / Vaccine xxx (xxxx) xxx
Table 2 Demographic distribution of caregivers in a study of vaccine preferences in Shanghai, China, 2017.
Table 4 Sensitivity analysis of cost preferences by demographic group, with an interaction term between demographic group and cost attribute.
Group
Category
Count (%)
Modela
Estimate
Caregiver’s relationship
Mother Father Grandparent
408 (70%) 167 (29%) 8 (1%)
Income
Urbanicity
Urban Inner suburbs Outer suburbs
182 (30%) 299 (50%) 118 (20%)
Cost Income (reference: low income [under 10,000 RMB each month]) Cost * income interaction
0.08
0.0494
Residency
Residency
Local Urban non-local Rural non-local
282 (47%) 111 (19%) 202 (34%)
Cost Urban non-locals (reference: locals) Rural non-locals (reference: locals) Cost * urban non-local interaction Cost * rural non-local interaction
0.02 0.71 0.09 0.06 0.00
0.4789 0.0516 0.7596 0.2379 0.9958
Education
Middle school High school Vocational College Graduate school
101 (17%) 87 (15%) 128 (22%) 226 (38%) 51 (9%)
Education
Cost Education (reference: low education [middle school or less]) Cost * education interaction
0.00 0.46
0.2330 0.2011
0.10
0.0593
<5,000 RMB 5,000–7,499 RMB 7,500–9,999 RMB 10,000–14,999 RMB 15,000 RMB
85 (15%) 103 (18%) 78 (13%) 132 (23%) 183 (31%)
Monthly family income
Note: sample sizes do not consistently sum to 599 if there were missing respondent data to specific questions.
Table 3 Results of discrete choice experiment of vaccine preferences among parents in Shanghai, China, 2017 (N = 552). Attributes
OR
95% CI
P-value
Risk of fever after vaccination (30% risk compared to 10% risk) Vaccine manufacturer and testing (foreign tested compared to Chinese) Vaccine manufacturer and testing (foreign not tested compared to Chinese) Vaccine effectiveness (85% compared to 95%)
0.53
<0.0001
Cost per vaccine (continuous 100 unit increase in RMB) Severity of disease (no hospitalization compared with required hospitalization) How common is the disease without vaccination (1% compared with 10%)
1.00
(0.44, 0.64) (0.60, 0.92) (0.37, 0.53) (0.48, 0.62) (0.98, 1.01) (1.76, 2.29) (0.77, 1.03)
0.74 0.45 0.55
2.01 0.89
0.0059 <0.0001 <0.0001 0.7974 <0.0001 0.1154
not significant (P = 0.1154). Contrary to our hypothesis, parents expressed a preference for a vaccine against a disease that was less serious (no hospitalization vs hospitalization OR: 2.01, 95% CI: 1.76, 2.29). The location of the manufacturer was a strong predictor of preferences. Parents preferred Chinese-made vaccines to those made by foreign companies (OR: 0.74, 95% CI: 0.60, 0.92) or made by foreign companies and not tested in Chinese children (OR: 0.45, 95% CI: 0.37, 0.53). Cost was not associated with stated preferences for a vaccine in this analysis. As an additional analysis (Table 4), we examined interactions between the cost attribute and sociodemographic characteristics in three separate models, each with one sociodemographic characteristic. There was a significant interaction term between cost and income, with individuals in families earning over 10,000 RMB each month expressing higher utilities for greater cost (P = 0.0494). The model with education did not show a significant result for those with a high school or greater education (P = 0.0593), and residency was likewise an insignificant interaction term.
Beta 0.01 0.37
P-value 0.7449 0.1657
a
Separate logistic regression models were fit for each demographic characteristic; no other independent variable (main effect or interaction) was entered into the model.
4. Discussion Expanding the use of these optional, for-fee vaccines can be an important component of disease control measures for LMICs, which like China, tend to have fewer vaccines on the EPI than in high income countries [20]. In this study of caregivers in Shanghai, China, parents expressed preferences for (EPI and non-EPI) vaccines with fewer side effects, which were more effective, and which were produced domestically or tested in Chinese children. Location of vaccine manufacturer was a strong predictor of utility – with parents in this study stating greater preferences for vaccines produced in China and tested in Chinese children. In a previous cross-sectional survey that we conducted among caregivers in Shanghai, we found that there was a clear divergence among those who perceived domestic vaccines as better, those who did not, and those who felt there were no differences [9]: most (56%) respondents thought that the effectiveness of domestic and imported vaccines were equivalent, whereas one-third stated domestic were more effective and 11% stated foreign vaccines were more effective. Two-thirds thought foreign and domestic vaccines had similar levels of safety, whereas 11% believed domestic vaccines were safer and 21% stated so for foreign vaccines [9]. In sum, that survey’s findings posited that people had a tendency to think Chinese-made vaccines were more effective but less safe. However, in this DCE study, we found that the overwhelming preference was for Chinese-made vaccines. Although vaccines included on the EPI in China are currently all Chinese-made [2], not all non-EPI vaccines are, and popularization of foreign-made vaccines should coincide with disseminating information on how these vaccines also undergo clinical trials in Chinese children prior to appearing on the market to insure uptake. The future of vaccine products in China made by foreign manufacturers is under question; China is one of the few countries that has been able to domestically produce all EPI vaccines [21], and is eyeing expansion into the global market [22]. However, vaccines like PCV and HPV vaccine are not produced by any domestic company yet, and so their use in the country is dependent on importing from foreign companies. Several previous studies have identified the importance of different vaccination attributes – with an obvious worldwide preference for safe and effective vaccines. A nationwide sample from the US from 2009 found that effectiveness of the vaccine and risk of side effects were the two most important attributes to consider
Please cite this article as: X. Sun, A. L. Wagner, J. Ji et al., A conjoint analysis of stated vaccine preferences in Shanghai, China, Vaccine, https://doi.org/ 10.1016/j.vaccine.2019.11.062
X. Sun et al. / Vaccine xxx (xxxx) xxx
when making a decision to get an influenza vaccine for their child, with 92% and 75% of the population, respectively, considering these attributes as important [23]. Gidengil et al. also found that parents in the US were willing to pay a large amount of money for vaccines with lower occurrence of side effects [15]. In a study of travel vaccines in Germany in 2015, disease severity was one of the strongest drivers of preference (along with risk of acquiring disease) [24]. And in a study from Thailand in 2008–2009, cost, vaccine efficacy, and absence of side effects were three of the strongest predictors of vaccine preference [25]. A study from Shandong province found that effectiveness and safety were strongly related to preferences for adult hepatitis B vaccination: individuals were willing to pay 68 RMB ($10) for a vaccine with 99% protection versus 79% protection, and they were willing to pay 38 RMB ($6) to avoid a 1/50,000 risk of side effects versus 1/150,000 risk of side effects [26]. Studies have also tried to tease out different conceptualizations of these attributes. For example, a DCE in Thailand had three attributes related to vaccine effectiveness: efficacy, duration of protection, and vaccine-induced seropositivity [25]. A study from Shandong province, China, also distinguished between effectiveness as a proportion of those vaccinated and duration of protection [26]. In regards to vaccine safety, a study from Australia distinguished between the attributes of ‘‘severe side effects” and ‘‘mild side effects” [27]. And disease risk could also be conceived of in different ways. A German DCE distinguished among the attributes of disease risk, transmissibility, and worry about infecting others [24]. These different conceptualizations could impact parent preferences, but that findings our relatively consistent across studies indicate that safety and effectiveness as abstract concepts are important, although how these concepts are perceived within a local population may be unique and be worth further study. 4.1. Price considerations We found that in the study population as a whole, cost was not a significant contributor to their preferences for a vaccine, although our evidence points to more affluent individuals (higher education or higher income) having a slight preference for more expensive vaccines and less affluent individuals having a null preference. This finding is a marked contrast to previous DCE studies worldwide and in China. Insensitivity to price for health care in China could be explained by several mechanisms. This could be a survey artifact – through a social desirability bias, participants may think that they should not be sensitive to, or this may be a difficult thought process for many given that EPI vaccines are given to parents for free. However, parents could perceive higher cost vaccines as having higher quality beyond what is explained in the other attributes (e.g., effectiveness, safety). This finding is somewhat in accordance with people’s preferences for name-brand (versus generic) drugs. A study conducted in Brazil found that poorer people preferred name-brand medicines because they thought these drugs had higher quality [28]. Another study of young and educated people found that about half of respondents were strongly sensitive to medicine prices, while the rest expressed a preference for a certain brand [29]. These studies’ findings suggest that it may be difficult to disentangle issues relating to cost and quality in stated preference surveys in China. 4.2. Strengths and limitations A strength of this conjoint analysis was that its list of attributes and attribute levels were derived from a prior qualitative study [18], and so should be relatively comprehensive of Shanghai parental concerns. Additionally, we included a representative sample of parents throughout different districts in the city through our sam-
5
pling scheme. However, this sample may be biased against individuals who do not regularly attend immunization clinics; we do not believe this is a major concern, however, given that EPI vaccines are free and mandatory for school entry in China. In a previous study of electronic immunization records in Shanghai, we found that coverage of diphtheria-tetanus-pertussis dose 1 (an EPI vaccine) was above 97% in the total population, regardless of residency and urbanicity, and that there was high levels of timely administration of this vaccine between 3 and 4 months of age [30]. Attendance of immunization clinics is high, although lower in some groups; a community survey of non-local families in Beijing found that 12% had not accessed the local immunization clinic [13]. Our survey therefore likely underrepresents the non-local population in Shanghai by a small degree. Future studies sampling directly from the community could explore whether the preferences for cost and other vaccine attributes vary from our findings. We purposefully selected caregivers of infants at a young age in order to limit the exposure parents had to pediatric vaccination or childhood diseases. However, we note that attitudes towards vaccination do change over time with parents being relatively more hesitant when their child is younger [31]. Our results may not be generalizable to parents of older children. As a sensitivity analysis, we looked at the interaction between preferences and some sociodemographic attributes, but others, like age and occupation could also be influential. An unexpected finding was that parents seemingly preferred vaccines for diseases that were not as severe. We measured severity of disease through a conceptualization of ‘‘hospitalization,” a term which was raised repeatedly in qualitative interviews that we conducted. Parents’ responses to this question could be based on a misunderstanding of the question. For example, parents may have believed that their child would be able to receive medical care for certain choice profiles but not for others, or they may have conflated severity of disease with side effects of the vaccine. Another possibility is that there are interactions that we were unable to detect in our study, in that at some levels of disease severity there is less preferences for other attributes like vaccine side effects. Future studies using stated preference methods in China should still consider measures of disease severity, but instead of dichotomizing this attribute based on hospitalization, it could instead be based on length of disease or clinical presentation. We also note that conjoint analyses should have attributes that are mutually exclusive, but it is possible that participants conflated vaccine safety and location of manufacturer. 5. Conclusions China will likely expand the number of vaccines in its publicly funded immunization schedule in the coming years, and beyond these mandatory vaccines, parents have the choice to purchase additional funds at the same immunization clinics. Understanding parents’ preferences can help public health officials and clinic workers improve acceptability and maximize uptake of certain vaccines. In this study of caregivers in Shanghai, China, parents expressed preferences for vaccines with fewer side effects, which were more effective, and which were produced domestically. Promotion of vaccines in China should focus on parents’ stated preferences, namely low risk of adverse events, high effectiveness, and information about vaccine production in China or past testing performed in Chinese children. CRediT authorship contribution statement Xiaodong Sun: Funding acquisition, Project administration, and Writing - review & editing. Abram L. Wagner: Conceptualization,
Please cite this article as: X. Sun, A. L. Wagner, J. Ji et al., A conjoint analysis of stated vaccine preferences in Shanghai, China, Vaccine, https://doi.org/ 10.1016/j.vaccine.2019.11.062
6
X. Sun et al. / Vaccine xxx (xxxx) xxx
Formal analysis, Writing - original draft. Jiahui Ji: Formal analysis and Writing - original draft. Zhuoying Huang: Investigation, Project administration, and Writing - review & editing. Brian J. Zikmund-Fisher: Funding acquisition, Supervision, Writing review & editing. Matthew L. Boulton: Funding acquisition, Supervision, Writing - review & editing. Jia Ren: Investigation, Project administration, and Writing - review & editing. Lisa A. Prosser: Funding acquisition, Supervision, Writing - review & editing. Declaration of Competing Interest The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper. Acknowledgements We are grateful to the vaccine providers who gave us access to their clinics to conduct these interviews. Funding This work was supported by The Fourth Round of Three-Year Public Health Action Plan of Shanghai, China (No. 15GWZK0101). This study was funded by grants from the Office of Global Public Health, University of Michigan, Ann Arbor, the Susan B. Meister Child Health Evaluation and Research Center, and from the MCubed program, University of Michigan, Ann Arbor. ALW has received funding as a Health Outcomes Post Doctoral Fellow from the PhRMA Foundation. Author contributions XS contributed to design of study, acquisition of data, and revised manuscript critically for important intellectual content. ALW contributed to design of study, acquisition of data, data analysis, data interpretation, and drafted the article. JJ contributed to data analysis and drafted the article. ZH and JR contributed to acquisition of data and revised manuscript critically for important intellectual content. BJZ, MLB, and LAP contributed to funding, design of study, data interpretation, and revised manuscript critically for important intellectual content. All authors gave final approval of the version to be published. References [1] Zheng J, Zhou Y, Wang H, Liang X. The role of the china experts advisory committee on immunization program. Vaccine 2010;28S:A84–7. https://doi. org/10.1016/j.vaccine.2010.02.039. [2] Yu H, Yang W, Varma JK. To save children’s lives, China should adopt an initiative to speed introduction of pneumonia vaccines. Heal Aff. 2012;31 (11):2545–53. https://doi.org/10.1377/hlthaff.2011.1272. [3] World Health Organization. Summary of WHO Position Papers Recommendations for Routine Immunization.. Published 2017. Accessed September 7, 2017. http://www.who.int/immunization/policy/ Immunization_routine_table1.pdf. [4] Pan X-F, Griffiths UK, Pennington M, Yu H, Jit M. Systematic review of economic evaluations of vaccination programs in mainland China: Are they sufficient to inform decision making?. Vaccine 2015;33:6164–72. https://doi. org/10.1016/j.vaccine.2015.09.081. [5] Qu F, Weschler LB, Sun Y, Sundell J. High pneumonia lifetime-ever incidence in Beijing children compared with locations in other countries, and implications for national PCV and Hib vaccination. PLoS ONE 2017;12(2):e0171438. https:// doi.org/10.1371/journal.pone.0171438. [6] Chen Y, Deng W, Wang S-M, et al. Burden of pneumonia and meningitis caused by Streptococcus pneumoniae in China among children under 5 years of age: a systematic literature review. PLoS ONE 2011;6(11):e27333. https://doi.org/ 10.1371/journal.pone.0027333.
[7] Watt JP, Wolfson LJ, O’Brien KL, et al. Burden of disease caused by Haemophilus influenzae type b in children younger than 5 years: global estimates. Lancet 2009;374(9693):903–11. https://doi.org/10.1016/S0140-6736(09)61203-4. [8] Zhang J, Duan Z, Payne DC, et al. Rotavirus-specific and overall diarrhea mortality in Chinese children younger than 5 years 2003 to 2012. Pediatr Infect Dis J. 2015;34(10):e233–7. https://doi.org/10.1097/INF.0000000000000799. [9] Huang Z, Sun X, Wagner AL, et al. Parent and caregiver perceptions about the safety and effectiveness of foreign and domestic vaccines in Shanghai, China. PLoS ONE 2018;13(5):e0197437. https://doi.org/10.1371/journal. pone.0197437. [10] Wagner AL, Boulton ML, Sun X, et al. Perceptions of measles, pneumonia, and meningitis vaccines among caregivers in Shanghai, China, and the Health Belief Model: a cross-sectional study. BMC Pediatr. 2017;17:143. https://doi. org/10.1186/s12887-017-0900-2. [11] Wagner AL, Boulton ML, Sun X, et al. Parents’ concerns about vaccine scheduling in Shanghai China. Vaccine 2017;35(34):4362–7. https://doi.org/ 10.1016/j.vaccine.2017.06.077. [12] Gong P, Liang S, Carlton EJ, et al. Urbanisation and health in China. Lancet 2012;379(9818):843–52. https://doi.org/10.1016/S0140-6736(11)61878-3. [13] Sun M, Ma R, Zeng Y, Luo F, Zhang J, Hou W. Immunization status and risk factors of migrant children in densely populated areas of Beijing China. Vaccine 2010;28(5):1264–74. https://doi.org/10.1016/j.vaccine.2009.11.015. [14] Bridges JFP, Hauber AB, Marshall D, et al. Conjoint analysis applications in health - A checklist: A report of the ISPOR good research practices for conjoint analysis task force. Value Heal. 2011;14(4):403–13. https://doi.org/10.1016/j. jval.2010.11.013. [15] Gidengil C, Lieu TA, Payne K, Rusinak D, Messonnier M, Prosser LA. Parental and societal values for the risks and benefits of childhood combination vaccines. Vaccine 2012;30(23):3445–52. https://doi.org/10.1016/ j.vaccine.2012.03.022. [16] Ren J, Wagner AL, Zheng A, et al. The demographics of vaccine hesitancy in Shanghai, China. PLoS ONE 2018;13(12):e0209117. https://doi.org/10.1371/ journal.pone.0209117. [17] Orme BK. Sample Size Issues for Conjoint Analysis. In: Getting Started with Conjoint Analysis: Strategies for Product Design and Pricing Research. Madison, WI: Research Publishers LLC; 2010. p. 57–66. [18] Sun X, Huang Z, Wagner AL, et al. The role of severity perceptions and beliefs in natural infections in Shanghai parents’ vaccine decision-making: a qualitative study. BMC Public Health. 2018;18:813. https://doi.org/10.1186/s12889-0185734-9. [19] Marshall D, Bridges JFP, Hauber B, et al. Conjoint analysis applications in health - how are studies being designed and reported?: An update on current practice in the published literature between 2005 and 2008. Patient 2010;3(4):249–56. https://doi.org/10.2165/11539650-000000000-00000. [20] Makinen M, Kaddar M, Molldrem V, Wilson L. New vaccine adoption in lowermiddle-income countries. Heal Policy Plan 2012;27:ii39–49. https://doi.org/ 10.1093/heapol/czs036. [21] Xu M, Liang Z, Xu Y, Wang J. Chinese vaccine products go global: Vaccine development and quality control. Expert Rev Vaccines 2015;14(5):763–73. https://doi.org/10.1586/14760584.2015.1012503. [22] Stevenson MA. Geneva-Seattle collaboration in support of developing country vaccine manufacturing. Glob Public Health. 2016;1692:1–16. https://doi.org/ 10.1080/17441692.2016.1245349. [23] Flood EM, Ryan KJ, Rousculp MD, et al. Parent preferences for pediatric influenza vaccine attributes. Clin Pediatr (Phila). 2011;50(4):338–47. https:// doi.org/10.1177/0009922810391247. [24] Poulos C, Curran D, Anastassopoulou A, De Moerlooze L. German travelers’ preferences for travel vaccines assessed by a discrete choice experiment. Vaccine 2018;36(7):969–78. https://doi.org/10.1016/j.vaccine.2018.01.004. [25] Cameron MP, Newman PA, Roungprakhon S, Scarpa R. The marginal willingness-to-pay for attributes of a hypothetical HIV vaccine. Vaccine 2013;31(36):3712–7. https://doi.org/10.1016/j.vaccine.2013.05.089. [26] Guo N, Zhang G, Zhu D, Wang J, Shi L. The effects of convenience and quality on the demand for vaccination: Results from a discrete choice experiment. Vaccine 2017;35(21):2848–54. https://doi.org/10.1016/j.vaccine.2017.04.006. [27] Hall J, Kenny P, King M, Louviere J, Viney R, Yeoh A. Using stated preference discrete choice modelling to evaluate the introduction of varicella vaccination. Health Econ. 2002;11:457–65. https://doi.org/10.1002/hec.694. [28] Guttier MC, Silveira MPT, Luiza VL, Bertoldi AD. Factors influencing the preference for purchasing generic drugs in a Southern Brazilian city. Rev Saude Publica. 2017;51:59. https://doi.org/10.1590/S1518-8787.2017051006786. [29] Halme M, Linden K, Kääria K. Patients’ preferences for generic and branded over-the-counter medicines: An adaptive conjoint analysis approach. Patient 2009;2(4):243–55. https://doi.org/10.2165/11314130-000000000-00000. [30] Wagner AL, Sun X, Montgomery JP, Huang Z, Boulton ML. The Impact of residency and urbanicity on haemophilus influenzae type b and pneumococcal immunization in shanghai children: a retrospective cohort study. PLoS ONE 2014;9(5):e97800. https://doi.org/10.1371/journal.pone.0097800. [31] Henrikson NB, Anderson ML, Opel DJ, Dunn J, Marcuse EK, Grossman DC. Longitudinal trends in vaccine hesitancy in a cohort of mothers surveyed in Washington State, 2013–2015. Public Health Rep. 2017;132(4):451–4. https:// doi.org/10.1177/0033354917711175.
Please cite this article as: X. Sun, A. L. Wagner, J. Ji et al., A conjoint analysis of stated vaccine preferences in Shanghai, China, Vaccine, https://doi.org/ 10.1016/j.vaccine.2019.11.062
Contents lists available at ScienceDirect
Vaccine journal homepage: www.elsevier.com/locate/vaccine
A conjoint analysis of stated vaccine preferences in Shanghai, China Xiaodong Sun a, Abram L. Wagner b,⇑, Jiahui Ji b, Zhuoying Huang a, Brian J. Zikmund-Fisher c,d, Matthew L. Boulton b,e, Jia Ren a, Lisa A. Prosser f a
Department of Immunization Program, Shanghai Municipal Centers for Disease Control & Prevention, NO. 1380, West Zhongshan Road, 200336, Shanghai, China Department of Epidemiology, School of Public Health, University of Michigan, 1415 Washington Heights, Ann Arbor, MI 48109, USA c Department of Health Behavior & Health Education, School of Public Health, University of Michigan, 1415 Washington Heights, Ann Arbor, MI 48109, USA d Department of Internal Medicine, Division of General Medicine, University of Michigan Medical School, 1500 East Medical Center Drive, Ann Arbor, MI 48109, USA e Department of Internal Medicine, Division of Infectious Disease, University of Michigan Medical School, 1500 East Medical Center Drive, Ann Arbor, MI 48109, USA f Susan B. Meister Child Health Evaluation and Research Center, Department of Pediatrics, University of Michigan Medical School, Ann Arbor, MI, USA b
a r t i c l e
i n f o
Article history: Received 6 July 2019 Received in revised form 8 November 2019 Accepted 25 November 2019 Available online xxxx Keywords: China Conjoint analysis Discrete choice experiment Vaccine manufacturers
a b s t r a c t It is not clear what kind of preferences parents in China would have for vaccines that could be added to a future immunization schedule. This study’s aim was to assess Chinese parents’ preferences for attributes of vaccines. We surveyed parents of young infants 3 months of age at immunization clinics in Shanghai, China, in 2017. We used a discrete choice experiment (DCE) to present parents with choices between two hypothetical profiles of vaccines which were described using the following attributes: cost, risk of side effect, location of vaccine manufacturer, vaccine testing, vaccine effectiveness, severity of disease, disease prevalence. A logistic regression output estimates of preference utilities. In total, 599 caregivers completed the DCE. Parents expressed lower preference for vaccines with a 30% chance of fever as an adverse event vs a 10% chance (OR: 0.53, 95% CI: 0.44, 0.64), for vaccines only 85% effective vs those 95% effective (OR: 0.55, 95% CI: 0.48, 0.62), and for imported vaccines (OR: 0.74, 95% CI: 0.60, 0.92) and those not tested in Chinese children (OR: 0.45, 95% CI: 0.37, 0.53) compared to domestic vaccines. More affluent groups preferred more expensive vaccines whereas less affluent groups did not express cost-based preferences. Promotion of vaccines in China should focus on parents’ stated preferences, which include past testing done in Chinese children – which is, in fact, required of all licensed vaccines in China. Information about these trials could emphasize low risk of adverse events and high effectiveness. Ó 2019 Elsevier Ltd. All rights reserved.
1. Introduction Vaccines protect children against a number of diseases with the potential to cause serious illness and death. In China, the government-funded Expanded Program on Immunization (EPI) has been providing free and mandatory vaccines for tuberculosis, polio, measles, pertussis, diphtheria, tetanus, and others since 1978 [1]. All EPI vaccines in China are currently manufactured by pharmaceutical companies based in China [2]. The number of vaccines included on the EPI in China is fewer than recommended by the World Health Organization (WHO) [3], and is therefore likely to increase in the future.
⇑ Corresponding author at: 1415 Washington Heights, Ann Arbor, MI 48109, USA. E-mail addresses: [email protected] (X. Sun), [email protected] (A.L. Wagner), [email protected] (J. Ji), [email protected] (Z. Huang), [email protected] (B.J. Zikmund-Fisher), [email protected] (M.L. Boulton), [email protected] (J. Ren), [email protected] (L.A. Prosser).
In addition to free EPI vaccines, Chinese clinics offer some forfee non-EPI vaccines, which are recommended by the World Health Organization including Haemophilus influenzae type b vaccine (Hib), pneumococcal conjugate vaccine (PCV), and rotavirus vaccine. Those vaccines could help reduce childhood morbidity in China, and many analyses have posited that these vaccines would be cost-effective [4]. In China, the pneumococcal and Hib vaccines, in particular, are prime candidates for inclusion on the EPI schedule given the substantial burden of pneumonia in China [2]; 26.9% of children 2–8 years old in Beijing, China, have ever been diagnosed with pneumonia [5]. Nationwide, these two vaccines have the potential to prevent a significant portion of the 11,000 deaths due to pneumococcal pneumonia and meningitis [6] and 19,000 deaths from Hib pneumonia and meningitis in Chinese children under 5 years of age annually [7]. From 2003 to 2012, a total of 127,539 deaths from diarrhea were reported among Chinese children <5 years of age, of which an estimated 53,559 (42%) had illness attributable to rotavirus [8]. Although more widespread use of these vaccines could clearly reduce childhood morbidity
https://doi.org/10.1016/j.vaccine.2019.11.062 0264-410X/Ó 2019 Elsevier Ltd. All rights reserved.
Please cite this article as: X. Sun, A. L. Wagner, J. Ji et al., A conjoint analysis of stated vaccine preferences in Shanghai, China, Vaccine, https://doi.org/ 10.1016/j.vaccine.2019.11.062
2
X. Sun et al. / Vaccine xxx (xxxx) xxx
and mortality in China, the evidence base on parental vaccine decision-making needed to inform efforts to increase uptake of vaccines which are both optional and require a fee, is currently lacking. An additional consideration is that vaccines in China can be produced by domestic or foreign companies, and parents for some vaccines can make a choice based on manufacturer’s location. Although parental perceptions of these manufacturers may vary [9], how this affects their stated preferences is unclear. Vaccination behaviors in China have been shown to vary based on parental characteristics, like residency, urbanicity [10], and education [11]. Residency refers to where a family is officially registered (the specific city or village in the country, but also whether this is a rural or urban location). Some local governmental services are limited to individuals with official residence in that area [12], although EPI vaccines are provided to children regardless of whether they are local or non-locals (sometimes referred to as the ‘‘floating population”). Most (88%) of non-local children will attend government-run immunization clinics [13]. Residency differs from urbanicity, which is where an individual family currently lives. Shanghai only consists of urban and suburban areas, and the latter group can be divided into inner and outer suburbs based on their proximity to historically more economically developed areas. There are also external factors—attributes of vaccines and vaccination programs (like cost, safety of vaccine, and manufacturer characteristics), that are associated with people’s behaviors. However, highlighting vaccine attributes or certain programmatic characteristics could enhance vaccination behaviors. The government and health organizations could benefit from a greater understanding of those attributes which are favorably perceived by the public that would, in turn, permit tailoring of vaccines and vaccination programs to improve vaccine acceptance and use. Preferences for attributes can be measured through conjoint analysis such as discrete choice experiments (DCEs) [14]. DCE is a survey technique that has been used to understand parent preference toward vaccination in the US and other countries. DCEs, as a method of conjoint analysis, have been widely used to quantify values of public on health-related choices [14]. This approach has the advantage of allowing for several attributes to be evaluated simultaneously in the choice scenario. The respondents need to consider which profile provides more value to them based on all the attributes in a given choice set, which will necessarily involve trade-offs among attributes in the decision making process. For example, a previous study in the US looked at attributes like the risk of fever after vaccination, the cost of visit, community vaccination coverage, and doses required to complete the vaccine series [15]. The choice of attributes should be driven by local circumstances. In a previously published paper of this study population, over two-thirds expressed concerns about vaccine safety [16]. The objective of this study is to estimate preferences of Chinese parents for attributes of EPI and non-EPI pediatric vaccines, and to calculate their willingness to pay for certain attributes.
2. Methods 2.1. Study population This cross-sectional study enrolled a convenience sample of caregivers of infants <3 months at public immunization clinics in Shanghai, China, between May and September 2017. The sampling of clinics was tied to the population of the surrounding township, according to a probability-proportionate-to-size (PPS) method. The two inclusion criteria were as follows: study participants had to be (1) a parent or grandparent of an infant <3 months of age, and (2) at least 18 years old.
Participants completed paper questionnaires in a private room after their child had received a vaccine, and the answers to these were input into an electronic database. The questionnaire included non-invasive questions about the participants’ perceptions of vaccines, including their valuation of the attributes of vaccines in a conjoint analysis (in which parents were asked to choose between two hypothetical profiles of vaccines) (Fig. 1). Participants completed one of two versions of a questionnaire, one version, the focus of this manuscript, asked participants about vaccine attributes; the other asked about vaccination program attributes, which is addressed in a different paper. A rule of thumb for the sample size of conjoint analysis is the following equation [17]:
N>
500c ta
where N is the sample size, c is the largest number of attribute levels (4), t is the number of choice tasks (8, from the other discrete choice experiment), and a is the number of alternatives (2). This leads a sample size of 250 within a simple random sample. Accounting for a design effect of 1.6384 (estimated from an intracluster coefficient of 0.0456 from a previous study on vaccine attitudes [10,11], and with 15 individuals in each cluster), the required sample size would be 410. The dataset is publicly available at a figshare repository (https://doi.org/10.6084/m9.figshare.6463304). 2.2. Attributes and survey design The list of attributes is found in Table 1. Attributes included the following: cost per vaccine, risk of getting fever after vaccination, vaccine manufacturer, vaccine effectiveness, severity of disease, and how common the disease is among children without vaccination. These attributes were identified through a qualitative study undertaken in the summer of 2016 and published elsewhere [18]. The number of attributes we included (7) is similar to numbers included in past studies [19]. For ease of analysis, we combined vaccine manufacturer and vaccine testing into one variable with three categories: Chinesemade vaccine, foreign vaccine tested in Chinese children, and foreign vaccine not tested in Chinese children. All vaccines marketed in China must be tested in Chinese children, but we included this characteristic because it arose as a concern in qualitative interviews. We used a fractional factorial design, and gave participants a forced choice (i.e., respondents had to choose one of the two profiles with no opt-out). 2.3. Statistical analysis We used conditional multivariable logistic regression to measure all the discrete choices and evaluate each individual attribute. Observations with missing values were deleted from the analysis. The analysis used survey procedures clustered at the levels of individual and township immunization clinic. To further explore preferences related to cost, we ran several additional models, with only cost, a socioeconomic characteristic, and an interaction term between these two variables, as explanatory variables. Significance was assessed at an alpha = 0.05 level. All analyses were conducted in SAS version 9.4 (SAS Institute, Cary, NC, USA). 2.4. Ethical approval The research protocol was approved by the Health Sciences and Behavioral Sciences Institutional Review Board at the University of
Please cite this article as: X. Sun, A. L. Wagner, J. Ji et al., A conjoint analysis of stated vaccine preferences in Shanghai, China, Vaccine, https://doi.org/ 10.1016/j.vaccine.2019.11.062
X. Sun et al. / Vaccine xxx (xxxx) xxx
3
Fig. 1. Example choice task given to participants.
Michigan (#HUM00125379) and the ethical review committee at the Shanghai Centers for Disease Control and Prevention (#2017–2). All participants provided oral informed consent. After
completing the interview, participants were given a small gift (such as an umbrella or blanket) to compensate for time spend on completing the survey.
Table 1 Attributes and attribute levels in a discrete choice experiment of vaccine attributes.
3. Results
Attribute
Levels
Cost per vaccine
100 RMB 400 RMB 800 RMB 1200 RMB 10% 30% Chinese Foreign Tested in Chinese children Not tested in Chinese children 95% 85% Requires hospitalization for two days No hospitalization, child home sick for two days 10% 1%
Risk of getting fever after office visit Vaccine manufacturer Vaccine testing Vaccine effectiveness Severity of disease
How common is the disease among children without vaccination
In total, 1,285 individuals were asked to participate, and 1,188 (92.5%) agreed and consented to participate. Study participants were randomized into two groups with 599 in the vaccine attributes study presented in this manuscript. The demographic characteristics of study participants are found in Table 2. Most caregivers (408, 70%) were mothers, with fewer fathers (167, 29%) or grandparents (8, 1%). Caregivers were sampled from urban areas (30%), inner suburbs (50%), and outer suburbs (20%) of Shanghai (See Table 3). A total of 552 participants had complete data for the multivariable regression model of stated preferences. Caregivers expressed a strong preference for vaccines with lower risk of side effects (30% risk vs 10% risk OR: 0.53, 95% CI: 0.44, 0.64) and that were more effective (85% effective vs 95% effective OR: 0.55, 95% CI: 0.48, 0.62). Parents expressed a preference for a vaccine that protected against a more common disease, although this relationship was
Please cite this article as: X. Sun, A. L. Wagner, J. Ji et al., A conjoint analysis of stated vaccine preferences in Shanghai, China, Vaccine, https://doi.org/ 10.1016/j.vaccine.2019.11.062
4
X. Sun et al. / Vaccine xxx (xxxx) xxx
Table 2 Demographic distribution of caregivers in a study of vaccine preferences in Shanghai, China, 2017.
Table 4 Sensitivity analysis of cost preferences by demographic group, with an interaction term between demographic group and cost attribute.
Group
Category
Count (%)
Modela
Estimate
Caregiver’s relationship
Mother Father Grandparent
408 (70%) 167 (29%) 8 (1%)
Income
Urbanicity
Urban Inner suburbs Outer suburbs
182 (30%) 299 (50%) 118 (20%)
Cost Income (reference: low income [under 10,000 RMB each month]) Cost * income interaction
0.08
0.0494
Residency
Residency
Local Urban non-local Rural non-local
282 (47%) 111 (19%) 202 (34%)
Cost Urban non-locals (reference: locals) Rural non-locals (reference: locals) Cost * urban non-local interaction Cost * rural non-local interaction
0.02 0.71 0.09 0.06 0.00
0.4789 0.0516 0.7596 0.2379 0.9958
Education
Middle school High school Vocational College Graduate school
101 (17%) 87 (15%) 128 (22%) 226 (38%) 51 (9%)
Education
Cost Education (reference: low education [middle school or less]) Cost * education interaction
0.00 0.46
0.2330 0.2011
0.10
0.0593
<5,000 RMB 5,000–7,499 RMB 7,500–9,999 RMB 10,000–14,999 RMB 15,000 RMB
85 (15%) 103 (18%) 78 (13%) 132 (23%) 183 (31%)
Monthly family income
Note: sample sizes do not consistently sum to 599 if there were missing respondent data to specific questions.
Table 3 Results of discrete choice experiment of vaccine preferences among parents in Shanghai, China, 2017 (N = 552). Attributes
OR
95% CI
P-value
Risk of fever after vaccination (30% risk compared to 10% risk) Vaccine manufacturer and testing (foreign tested compared to Chinese) Vaccine manufacturer and testing (foreign not tested compared to Chinese) Vaccine effectiveness (85% compared to 95%)
0.53
<0.0001
Cost per vaccine (continuous 100 unit increase in RMB) Severity of disease (no hospitalization compared with required hospitalization) How common is the disease without vaccination (1% compared with 10%)
1.00
(0.44, 0.64) (0.60, 0.92) (0.37, 0.53) (0.48, 0.62) (0.98, 1.01) (1.76, 2.29) (0.77, 1.03)
0.74 0.45 0.55
2.01 0.89
0.0059 <0.0001 <0.0001 0.7974 <0.0001 0.1154
not significant (P = 0.1154). Contrary to our hypothesis, parents expressed a preference for a vaccine against a disease that was less serious (no hospitalization vs hospitalization OR: 2.01, 95% CI: 1.76, 2.29). The location of the manufacturer was a strong predictor of preferences. Parents preferred Chinese-made vaccines to those made by foreign companies (OR: 0.74, 95% CI: 0.60, 0.92) or made by foreign companies and not tested in Chinese children (OR: 0.45, 95% CI: 0.37, 0.53). Cost was not associated with stated preferences for a vaccine in this analysis. As an additional analysis (Table 4), we examined interactions between the cost attribute and sociodemographic characteristics in three separate models, each with one sociodemographic characteristic. There was a significant interaction term between cost and income, with individuals in families earning over 10,000 RMB each month expressing higher utilities for greater cost (P = 0.0494). The model with education did not show a significant result for those with a high school or greater education (P = 0.0593), and residency was likewise an insignificant interaction term.
Beta 0.01 0.37
P-value 0.7449 0.1657
a
Separate logistic regression models were fit for each demographic characteristic; no other independent variable (main effect or interaction) was entered into the model.
4. Discussion Expanding the use of these optional, for-fee vaccines can be an important component of disease control measures for LMICs, which like China, tend to have fewer vaccines on the EPI than in high income countries [20]. In this study of caregivers in Shanghai, China, parents expressed preferences for (EPI and non-EPI) vaccines with fewer side effects, which were more effective, and which were produced domestically or tested in Chinese children. Location of vaccine manufacturer was a strong predictor of utility – with parents in this study stating greater preferences for vaccines produced in China and tested in Chinese children. In a previous cross-sectional survey that we conducted among caregivers in Shanghai, we found that there was a clear divergence among those who perceived domestic vaccines as better, those who did not, and those who felt there were no differences [9]: most (56%) respondents thought that the effectiveness of domestic and imported vaccines were equivalent, whereas one-third stated domestic were more effective and 11% stated foreign vaccines were more effective. Two-thirds thought foreign and domestic vaccines had similar levels of safety, whereas 11% believed domestic vaccines were safer and 21% stated so for foreign vaccines [9]. In sum, that survey’s findings posited that people had a tendency to think Chinese-made vaccines were more effective but less safe. However, in this DCE study, we found that the overwhelming preference was for Chinese-made vaccines. Although vaccines included on the EPI in China are currently all Chinese-made [2], not all non-EPI vaccines are, and popularization of foreign-made vaccines should coincide with disseminating information on how these vaccines also undergo clinical trials in Chinese children prior to appearing on the market to insure uptake. The future of vaccine products in China made by foreign manufacturers is under question; China is one of the few countries that has been able to domestically produce all EPI vaccines [21], and is eyeing expansion into the global market [22]. However, vaccines like PCV and HPV vaccine are not produced by any domestic company yet, and so their use in the country is dependent on importing from foreign companies. Several previous studies have identified the importance of different vaccination attributes – with an obvious worldwide preference for safe and effective vaccines. A nationwide sample from the US from 2009 found that effectiveness of the vaccine and risk of side effects were the two most important attributes to consider
Please cite this article as: X. Sun, A. L. Wagner, J. Ji et al., A conjoint analysis of stated vaccine preferences in Shanghai, China, Vaccine, https://doi.org/ 10.1016/j.vaccine.2019.11.062
X. Sun et al. / Vaccine xxx (xxxx) xxx
when making a decision to get an influenza vaccine for their child, with 92% and 75% of the population, respectively, considering these attributes as important [23]. Gidengil et al. also found that parents in the US were willing to pay a large amount of money for vaccines with lower occurrence of side effects [15]. In a study of travel vaccines in Germany in 2015, disease severity was one of the strongest drivers of preference (along with risk of acquiring disease) [24]. And in a study from Thailand in 2008–2009, cost, vaccine efficacy, and absence of side effects were three of the strongest predictors of vaccine preference [25]. A study from Shandong province found that effectiveness and safety were strongly related to preferences for adult hepatitis B vaccination: individuals were willing to pay 68 RMB ($10) for a vaccine with 99% protection versus 79% protection, and they were willing to pay 38 RMB ($6) to avoid a 1/50,000 risk of side effects versus 1/150,000 risk of side effects [26]. Studies have also tried to tease out different conceptualizations of these attributes. For example, a DCE in Thailand had three attributes related to vaccine effectiveness: efficacy, duration of protection, and vaccine-induced seropositivity [25]. A study from Shandong province, China, also distinguished between effectiveness as a proportion of those vaccinated and duration of protection [26]. In regards to vaccine safety, a study from Australia distinguished between the attributes of ‘‘severe side effects” and ‘‘mild side effects” [27]. And disease risk could also be conceived of in different ways. A German DCE distinguished among the attributes of disease risk, transmissibility, and worry about infecting others [24]. These different conceptualizations could impact parent preferences, but that findings our relatively consistent across studies indicate that safety and effectiveness as abstract concepts are important, although how these concepts are perceived within a local population may be unique and be worth further study. 4.1. Price considerations We found that in the study population as a whole, cost was not a significant contributor to their preferences for a vaccine, although our evidence points to more affluent individuals (higher education or higher income) having a slight preference for more expensive vaccines and less affluent individuals having a null preference. This finding is a marked contrast to previous DCE studies worldwide and in China. Insensitivity to price for health care in China could be explained by several mechanisms. This could be a survey artifact – through a social desirability bias, participants may think that they should not be sensitive to, or this may be a difficult thought process for many given that EPI vaccines are given to parents for free. However, parents could perceive higher cost vaccines as having higher quality beyond what is explained in the other attributes (e.g., effectiveness, safety). This finding is somewhat in accordance with people’s preferences for name-brand (versus generic) drugs. A study conducted in Brazil found that poorer people preferred name-brand medicines because they thought these drugs had higher quality [28]. Another study of young and educated people found that about half of respondents were strongly sensitive to medicine prices, while the rest expressed a preference for a certain brand [29]. These studies’ findings suggest that it may be difficult to disentangle issues relating to cost and quality in stated preference surveys in China. 4.2. Strengths and limitations A strength of this conjoint analysis was that its list of attributes and attribute levels were derived from a prior qualitative study [18], and so should be relatively comprehensive of Shanghai parental concerns. Additionally, we included a representative sample of parents throughout different districts in the city through our sam-
5
pling scheme. However, this sample may be biased against individuals who do not regularly attend immunization clinics; we do not believe this is a major concern, however, given that EPI vaccines are free and mandatory for school entry in China. In a previous study of electronic immunization records in Shanghai, we found that coverage of diphtheria-tetanus-pertussis dose 1 (an EPI vaccine) was above 97% in the total population, regardless of residency and urbanicity, and that there was high levels of timely administration of this vaccine between 3 and 4 months of age [30]. Attendance of immunization clinics is high, although lower in some groups; a community survey of non-local families in Beijing found that 12% had not accessed the local immunization clinic [13]. Our survey therefore likely underrepresents the non-local population in Shanghai by a small degree. Future studies sampling directly from the community could explore whether the preferences for cost and other vaccine attributes vary from our findings. We purposefully selected caregivers of infants at a young age in order to limit the exposure parents had to pediatric vaccination or childhood diseases. However, we note that attitudes towards vaccination do change over time with parents being relatively more hesitant when their child is younger [31]. Our results may not be generalizable to parents of older children. As a sensitivity analysis, we looked at the interaction between preferences and some sociodemographic attributes, but others, like age and occupation could also be influential. An unexpected finding was that parents seemingly preferred vaccines for diseases that were not as severe. We measured severity of disease through a conceptualization of ‘‘hospitalization,” a term which was raised repeatedly in qualitative interviews that we conducted. Parents’ responses to this question could be based on a misunderstanding of the question. For example, parents may have believed that their child would be able to receive medical care for certain choice profiles but not for others, or they may have conflated severity of disease with side effects of the vaccine. Another possibility is that there are interactions that we were unable to detect in our study, in that at some levels of disease severity there is less preferences for other attributes like vaccine side effects. Future studies using stated preference methods in China should still consider measures of disease severity, but instead of dichotomizing this attribute based on hospitalization, it could instead be based on length of disease or clinical presentation. We also note that conjoint analyses should have attributes that are mutually exclusive, but it is possible that participants conflated vaccine safety and location of manufacturer. 5. Conclusions China will likely expand the number of vaccines in its publicly funded immunization schedule in the coming years, and beyond these mandatory vaccines, parents have the choice to purchase additional funds at the same immunization clinics. Understanding parents’ preferences can help public health officials and clinic workers improve acceptability and maximize uptake of certain vaccines. In this study of caregivers in Shanghai, China, parents expressed preferences for vaccines with fewer side effects, which were more effective, and which were produced domestically. Promotion of vaccines in China should focus on parents’ stated preferences, namely low risk of adverse events, high effectiveness, and information about vaccine production in China or past testing performed in Chinese children. CRediT authorship contribution statement Xiaodong Sun: Funding acquisition, Project administration, and Writing - review & editing. Abram L. Wagner: Conceptualization,
Please cite this article as: X. Sun, A. L. Wagner, J. Ji et al., A conjoint analysis of stated vaccine preferences in Shanghai, China, Vaccine, https://doi.org/ 10.1016/j.vaccine.2019.11.062
6
X. Sun et al. / Vaccine xxx (xxxx) xxx
Formal analysis, Writing - original draft. Jiahui Ji: Formal analysis and Writing - original draft. Zhuoying Huang: Investigation, Project administration, and Writing - review & editing. Brian J. Zikmund-Fisher: Funding acquisition, Supervision, Writing review & editing. Matthew L. Boulton: Funding acquisition, Supervision, Writing - review & editing. Jia Ren: Investigation, Project administration, and Writing - review & editing. Lisa A. Prosser: Funding acquisition, Supervision, Writing - review & editing. Declaration of Competing Interest The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper. Acknowledgements We are grateful to the vaccine providers who gave us access to their clinics to conduct these interviews. Funding This work was supported by The Fourth Round of Three-Year Public Health Action Plan of Shanghai, China (No. 15GWZK0101). This study was funded by grants from the Office of Global Public Health, University of Michigan, Ann Arbor, the Susan B. Meister Child Health Evaluation and Research Center, and from the MCubed program, University of Michigan, Ann Arbor. ALW has received funding as a Health Outcomes Post Doctoral Fellow from the PhRMA Foundation. Author contributions XS contributed to design of study, acquisition of data, and revised manuscript critically for important intellectual content. ALW contributed to design of study, acquisition of data, data analysis, data interpretation, and drafted the article. JJ contributed to data analysis and drafted the article. ZH and JR contributed to acquisition of data and revised manuscript critically for important intellectual content. BJZ, MLB, and LAP contributed to funding, design of study, data interpretation, and revised manuscript critically for important intellectual content. All authors gave final approval of the version to be published. References [1] Zheng J, Zhou Y, Wang H, Liang X. The role of the china experts advisory committee on immunization program. Vaccine 2010;28S:A84–7. https://doi. org/10.1016/j.vaccine.2010.02.039. [2] Yu H, Yang W, Varma JK. To save children’s lives, China should adopt an initiative to speed introduction of pneumonia vaccines. Heal Aff. 2012;31 (11):2545–53. https://doi.org/10.1377/hlthaff.2011.1272. [3] World Health Organization. Summary of WHO Position Papers Recommendations for Routine Immunization.. Published 2017. Accessed September 7, 2017. http://www.who.int/immunization/policy/ Immunization_routine_table1.pdf. [4] Pan X-F, Griffiths UK, Pennington M, Yu H, Jit M. Systematic review of economic evaluations of vaccination programs in mainland China: Are they sufficient to inform decision making?. Vaccine 2015;33:6164–72. https://doi. org/10.1016/j.vaccine.2015.09.081. [5] Qu F, Weschler LB, Sun Y, Sundell J. High pneumonia lifetime-ever incidence in Beijing children compared with locations in other countries, and implications for national PCV and Hib vaccination. PLoS ONE 2017;12(2):e0171438. https:// doi.org/10.1371/journal.pone.0171438. [6] Chen Y, Deng W, Wang S-M, et al. Burden of pneumonia and meningitis caused by Streptococcus pneumoniae in China among children under 5 years of age: a systematic literature review. PLoS ONE 2011;6(11):e27333. https://doi.org/ 10.1371/journal.pone.0027333.
[7] Watt JP, Wolfson LJ, O’Brien KL, et al. Burden of disease caused by Haemophilus influenzae type b in children younger than 5 years: global estimates. Lancet 2009;374(9693):903–11. https://doi.org/10.1016/S0140-6736(09)61203-4. [8] Zhang J, Duan Z, Payne DC, et al. Rotavirus-specific and overall diarrhea mortality in Chinese children younger than 5 years 2003 to 2012. Pediatr Infect Dis J. 2015;34(10):e233–7. https://doi.org/10.1097/INF.0000000000000799. [9] Huang Z, Sun X, Wagner AL, et al. Parent and caregiver perceptions about the safety and effectiveness of foreign and domestic vaccines in Shanghai, China. PLoS ONE 2018;13(5):e0197437. https://doi.org/10.1371/journal. pone.0197437. [10] Wagner AL, Boulton ML, Sun X, et al. Perceptions of measles, pneumonia, and meningitis vaccines among caregivers in Shanghai, China, and the Health Belief Model: a cross-sectional study. BMC Pediatr. 2017;17:143. https://doi. org/10.1186/s12887-017-0900-2. [11] Wagner AL, Boulton ML, Sun X, et al. Parents’ concerns about vaccine scheduling in Shanghai China. Vaccine 2017;35(34):4362–7. https://doi.org/ 10.1016/j.vaccine.2017.06.077. [12] Gong P, Liang S, Carlton EJ, et al. Urbanisation and health in China. Lancet 2012;379(9818):843–52. https://doi.org/10.1016/S0140-6736(11)61878-3. [13] Sun M, Ma R, Zeng Y, Luo F, Zhang J, Hou W. Immunization status and risk factors of migrant children in densely populated areas of Beijing China. Vaccine 2010;28(5):1264–74. https://doi.org/10.1016/j.vaccine.2009.11.015. [14] Bridges JFP, Hauber AB, Marshall D, et al. Conjoint analysis applications in health - A checklist: A report of the ISPOR good research practices for conjoint analysis task force. Value Heal. 2011;14(4):403–13. https://doi.org/10.1016/j. jval.2010.11.013. [15] Gidengil C, Lieu TA, Payne K, Rusinak D, Messonnier M, Prosser LA. Parental and societal values for the risks and benefits of childhood combination vaccines. Vaccine 2012;30(23):3445–52. https://doi.org/10.1016/ j.vaccine.2012.03.022. [16] Ren J, Wagner AL, Zheng A, et al. The demographics of vaccine hesitancy in Shanghai, China. PLoS ONE 2018;13(12):e0209117. https://doi.org/10.1371/ journal.pone.0209117. [17] Orme BK. Sample Size Issues for Conjoint Analysis. In: Getting Started with Conjoint Analysis: Strategies for Product Design and Pricing Research. Madison, WI: Research Publishers LLC; 2010. p. 57–66. [18] Sun X, Huang Z, Wagner AL, et al. The role of severity perceptions and beliefs in natural infections in Shanghai parents’ vaccine decision-making: a qualitative study. BMC Public Health. 2018;18:813. https://doi.org/10.1186/s12889-0185734-9. [19] Marshall D, Bridges JFP, Hauber B, et al. Conjoint analysis applications in health - how are studies being designed and reported?: An update on current practice in the published literature between 2005 and 2008. Patient 2010;3(4):249–56. https://doi.org/10.2165/11539650-000000000-00000. [20] Makinen M, Kaddar M, Molldrem V, Wilson L. New vaccine adoption in lowermiddle-income countries. Heal Policy Plan 2012;27:ii39–49. https://doi.org/ 10.1093/heapol/czs036. [21] Xu M, Liang Z, Xu Y, Wang J. Chinese vaccine products go global: Vaccine development and quality control. Expert Rev Vaccines 2015;14(5):763–73. https://doi.org/10.1586/14760584.2015.1012503. [22] Stevenson MA. Geneva-Seattle collaboration in support of developing country vaccine manufacturing. Glob Public Health. 2016;1692:1–16. https://doi.org/ 10.1080/17441692.2016.1245349. [23] Flood EM, Ryan KJ, Rousculp MD, et al. Parent preferences for pediatric influenza vaccine attributes. Clin Pediatr (Phila). 2011;50(4):338–47. https:// doi.org/10.1177/0009922810391247. [24] Poulos C, Curran D, Anastassopoulou A, De Moerlooze L. German travelers’ preferences for travel vaccines assessed by a discrete choice experiment. Vaccine 2018;36(7):969–78. https://doi.org/10.1016/j.vaccine.2018.01.004. [25] Cameron MP, Newman PA, Roungprakhon S, Scarpa R. The marginal willingness-to-pay for attributes of a hypothetical HIV vaccine. Vaccine 2013;31(36):3712–7. https://doi.org/10.1016/j.vaccine.2013.05.089. [26] Guo N, Zhang G, Zhu D, Wang J, Shi L. The effects of convenience and quality on the demand for vaccination: Results from a discrete choice experiment. Vaccine 2017;35(21):2848–54. https://doi.org/10.1016/j.vaccine.2017.04.006. [27] Hall J, Kenny P, King M, Louviere J, Viney R, Yeoh A. Using stated preference discrete choice modelling to evaluate the introduction of varicella vaccination. Health Econ. 2002;11:457–65. https://doi.org/10.1002/hec.694. [28] Guttier MC, Silveira MPT, Luiza VL, Bertoldi AD. Factors influencing the preference for purchasing generic drugs in a Southern Brazilian city. Rev Saude Publica. 2017;51:59. https://doi.org/10.1590/S1518-8787.2017051006786. [29] Halme M, Linden K, Kääria K. Patients’ preferences for generic and branded over-the-counter medicines: An adaptive conjoint analysis approach. Patient 2009;2(4):243–55. https://doi.org/10.2165/11314130-000000000-00000. [30] Wagner AL, Sun X, Montgomery JP, Huang Z, Boulton ML. The Impact of residency and urbanicity on haemophilus influenzae type b and pneumococcal immunization in shanghai children: a retrospective cohort study. PLoS ONE 2014;9(5):e97800. https://doi.org/10.1371/journal.pone.0097800. [31] Henrikson NB, Anderson ML, Opel DJ, Dunn J, Marcuse EK, Grossman DC. Longitudinal trends in vaccine hesitancy in a cohort of mothers surveyed in Washington State, 2013–2015. Public Health Rep. 2017;132(4):451–4. https:// doi.org/10.1177/0033354917711175.
Please cite this article as: X. Sun, A. L. Wagner, J. Ji et al., A conjoint analysis of stated vaccine preferences in Shanghai, China, Vaccine, https://doi.org/ 10.1016/j.vaccine.2019.11.062