Investigate Non-EPI Vaccination Recommendation Practice from a Socio-Ecological Perspective: A Mixed-Methods Study in China

The uptake of non-EPI vaccines, such as influenza and pneumonia vaccines, are very low in China compared to other countries. In China, immunization services are provided by dedicated vaccination service providers (VSPs), and their recommendation is the key to improve vaccine uptake. This study explores VSP recommendation practices for non-EPI vaccines from a socio-ecological perspective. A mixed-methods study, combining a questionnaire survey and key informant interviews, was conducted in Anhui, Shaanxi, and Guangdong provinces. 555 VSPs completed the valid questionnaire, and 49 VSPs participated in in-depth interviews. Among the surveyed VSPs, 51.54% stated that they always or often recommended non-EPI vaccines in work, and the remaining half reported that they sometimes or never recommended non-EPI vaccines. Most VSPs interviewed communicated about non-EPI vaccines with the public in an informed style, not a presumptive one, and provided the public with all the decision-making latitude. The infrequent recommendation of non-EPI vaccines was widely prevalent among Chinese VSPs regardless of their individual characteristics, and was mainly driven by the interpersonal relationship, institutional arrangement, and public policy. Firstly, the VSPs were concerned about conflicts arising from the recommendation of self-paid vaccines and the risk of adverse reactions following vaccination. Secondly, high workloads left them insufficient time to communicate about non-EPI vaccines. Thirdly, there was no performance assessment or financial incentive for VSPs to recommend non-EPI vaccination, and their main responsibility was around EPI vaccination. Therefore, multi-level socio-ecological systems around non-EPI vaccination should be improved to optimize the communication between VSPs and the public, which include a better system of legal redress to resolve potential misunderstandings between the VSPs and the public, more effective workload management through whole-process health information system and strengthening public health workforce, and the introduction of performance assessment and appropriate incentives on non-EPI vaccination.


Introduction
Immunization has proven to be one of the most cost-effective health interventions [1,2]. China initiated its national expanded program on immunization (EPI) in 1978. Currently, framework that incorporates the influences from other layers, such as institution regulation and policy.
The social-ecological model provides a conceptual framework to direct attention to both behavior and its individual and environmental determinants [20,21]. This model presents behavior as a product of the interdependence between the individual and subsystems of the ecosystem (e.g., family, community, culture, physical and social environment) [20]. In this model, patterned behavior is the outcome of interest and is viewed as being determined by five sub-ecosystems, which are intrapersonal, interpersonal, institutional, community, and policy. It has been used as a framework for studying medical services, such as non-prescription antibiotic dispensing [22]. It can also help to investigate HCWs' recommendation behaviors for vaccines in a comprehensive manner. This study aims to frame the potential determinants of HCWs' recommendation for non-EPI vaccines in China from a socio-ecological perspective. A mixed-methods study combining a crosssectional survey and key informant interviews was adopted for this purpose. Our target population was VSPs since they are dedicated to deliver vaccination services in China, instead of general HCWs.

Study Design
We conducted a mixed-methods cross-sectional study in January 2019 in Shenzhen megacity, Anhui province, and Shaanxi province, covering the East, Middle and West of China, respectively. One urban district and one rural county were selected separately in the Anhui and Shaanxi provinces, and one urban district was selected in Shenzhen megacity. In total, this study was conducted in five districts/counties in China. The

Survey of Vaccination Service Providers
To estimate the recommendation practice of non-EPI vaccines, a cross-sectional survey was conducted for all VSPs in the sampled districts and counties. A multi-stage sampling process was used to ensure the representativeness of the sample. Guangdong Province, Anhui Province and Shaanxi Province were selected to represent higher, median and lower social-economic tiers, respectively. At the provincial level, one urban district and one rural county were included in Shaanxi and Anhui provinces, and one urban district was included in Shenzhen megacity, Guangdong province. All VSPs (600) working in the sampled areas were invited to participate in a mobile-phone-based questionnaire survey by scanning a QR code. The self-administered questionnaire was distributed and managed using the online platform Wenjuanxing (https://www.wjx.cn/ (accessed on 9 June 2021)).

Interview
To understand the determinants of recommendation practice of non-EPI vaccines in depth, semi-structured interview was conducted following the questionnaire survey. In each sampled district/county, we interviewed one immunization program manager from CDC, and VSPs from vaccination clinics in three selected community health centers. These three community health centers were selected to represent low, medium, and high socioeconomic tiers within each district/county. Generally, there are 3-5 VSPs at a vaccination clinic, who are the director in charge of the clinic, vaccinators for vaccination service delivery and consultation, and a pediatrician for medical pre-screening and adverse reaction response. In each vaccination clinic, we invited one VSP from each job responsibility to participate in an interview.

Questionnaire
The questionnaire was piloted for 10 VSPs in two non-study communities in Shanghai (Appendix B). The content of the questionnaire included the (a) study site, rural or urban residence, gender, age, education level, and profession (doctors, nurses or public health workers); (b) recommendation frequency of non-EPI vaccines, measured using the following question-How often do you recommend non-EPI vaccines to the public?. There were four response options-"always", "often", "sometimes" and "never". Response options were further grouped into two categories for the analysis: often (including "always" and "often"), and not often (including "sometimes" and "never"). The question were linked to previous studies in the fields of HCWs' recommendation practice of vaccines [16].

Interview Guides
We developed interview guides according to five sub-ecosystems of the social-ecological model (Appendix C) [20]. First, we asked interviewees about their communication and recommendation of vaccines to the public in their daily work. In terms of intrapersonal sub-ecosystem, we focused on VSPs' knowledge, perception, and confidence in vaccines and vaccination services. For interpersonal sub-ecosystem, we asked about the quality of doctor-patient relationships and relationships with other colleagues. For the institutional sub-ecosystem, we asked the VSPs about their routine work, self-evaluation of workload, and the potential impact of both on the recommendation practice of non-EPI vaccines. For community sub-ecosystem, we enquired about the supply of non-EPI vaccines and whether any shortage of non-EPI vaccines had ever occurred. For public policy sub-ecosystem, we investigated the influence of financial incentive policy on the recommendation practice of non-EPI vaccines and assessment from superiors (CDC).
All participants were informed of the purpose of the study. They were also informed that participation was voluntary and that they could withdraw at any time. All participants were assured of the confidentiality of the interviews. Each interview lasted between 30 and 60 min and were audio-recorded after obtaining written informed consent.

Statistical Analysis for Survey Data
The recommendation practice of non-EPI vaccines was measured by the proportion of VSPs, who often recommend non-EPI vaccines among the total sample. Univariate analyses were performed to compare the VSPs' recommendation practice of non-EPI vaccines by their socio-demographic characteristics using Chi-square tests. A multivariable logistic regression analysis was further conducted to examine the factors associated with the VSPs' recommendation practice of non-EPI vaccines. Odds ratios with 95% confidence intervals were presented. All survey data were analyzed using STATA, version 14.0 (Stata Corp, College Station, TX, USA).

Data Analysis for Interviews
All interviews were transcribed verbatim and checked by another investigator. We conducted a thematic analysis using a combination of deductive and inductive coding to analyze the transcripts of the interviews [23]. We first identified detailed sub-themes via deductive, iterative coding of the data. Subsequently, exemplary data extracts were selected from the key sub-themes for inclusion as quotations. The interview transcripts were independently coded by two investigators, and any discrepancies were then discussed until a consensus was reached. All qualitative analysis were conducted using NVivo, version 11 (QSR International Inc., Burlington, MA, USA).

Quantitative Results
Surveyed VSPs' characteristics and recommendation practices for non-EPI vaccines are summarized in Table 1. Respondents who completed the questionnaire in less than 2 min or left more than 50% of the questionnaire incomplete, 45 in all, were excluded. In total, 555 of 600 VSPs completed the valid questionnaire. Of the 555 respondents, 15.32% and 36.22% stated that they always or often recommended non-EPI vaccines to patients in work, whereas 36.4% and 12.07% of respondents reported that they sometimes or never recommended these vaccines, respectively. Results from multivariate logistic regression (Table 1) suggested that respondents living in Anhui province were significantly more likely to recommend non-EPI vaccines than those in Shenzhen city (OR = 1.52, 95%CI: 1.04-2.20). VSPs older than 45 years old were significantly more likely to recommend non-EPI vaccines than those younger than 25 years old (OR = 2.50, 95%CI: 1.42-4.39). However, rural or urban residence, gender, education level, and professions had no significant association with recommendation practices for non-EPI vaccines.

Qualitative Results on Health Education and Recommendation Practices for Non-EPI Vaccines
In total, we conducted 43 interviews with VSPs and six interviews with immunization program managers (Table 2). Communication about non-EPI vaccines between VSPs and the public covers health education and recommendation practices.

Health Education on Immunization
Most participants said that health education on immunization (including education to parents of newborn babies) was provided routinely in their workplace. The content mainly covered the importance of vaccination and the introduction of the EPI in China. As one VSP noted: "First, we will give a general explanation of the components of the vaccine. Then, patients could wonder, some vaccines are free, and the other are not, why? Any difference between those two types of vaccines? We will tell, every vaccine is of the same importance. We also want parents to make sufficient preparation before vaccination. We will tell them to focus on five things: wearing the right clothes to keep warm; [ . . . ]. We need to popularize these for parents. The main thing is to get them to understand the importance of vaccination, the safety, right? And vaccines are very cost-effective." (VSP 4, male, Dongzhi county, Anhui province).

Recommendation Practices for Non-EPI Vaccines
Most VSPs said that they informed parents about age-appropriate vaccines for their children and asked about their intention to be vaccinated (mainly non-EPI vaccines) after the completion of EPI vaccinations. However, they did not actively recommend non-EPI vaccines. Almost all VSPs said that the purpose of this notification was to remind parents of the availability of non-EPI vaccines, and at the same time, honor parents' decision-making autonomy on non-EPI vaccinations for their children. As one VSP noted: "Definitely no recommendation, but every time after finishing one free (EPI) vaccination, I would talk to them. It's like, before the next free (EPI) vaccine, there are other vaccines available, they are voluntary and not free. Then I would tell them, if you want to get it, I can make another appointment for you. If you don't, we won't force you to get vaccinated. It's voluntary, basically. They would ask, didn't you say vaccination was free? Then I say this is non-EPI vaccine, you can choose to get it or not [ . . . ]." (VSP 3, female, Dongzhi county, Anhui province).

Interpersonal Sub-Ecosystem
Many VSPs indicated that they were concerned about adverse reactions following vaccination, which could cause conflicts between parents and themselves if they recommended non-EPI vaccines to parents. As one participant said: "One problem is. In one hospital, there was a case of adverse reaction related to non-EPI vaccination, and the dispute is very tricky. I remember that they compensated for it. They (the hospital) make so little money on vaccination, but finally have to pay so much compensation. They can't even carry out routine work at that time. Later, because of this, they almost gave up the inoculation of non-EPI vaccine. We just don't want to do it. The dean thought this was so tricky and he did not want to get involved in non-EPI vaccine. This case really hit him hard." (VSP 39, female, Jingyang County, Shaanxi province).
Some VSPs also stated that parents resent being recommended paid medical services (including vaccines). Therefore, recommending non-EPI vaccines may lead parents to perceive that the healthcare providers are profit-seeking and may further reduce parents' trust in them. As one participant said: "We don't recommend it, only inform them (with the age-appropriate vaccines). Why? They will be unsatisfied. For example, we will tell him that there are two kinds of Hepatitis A vaccines, one is imported, the other is domestic, and we let parents choose on their own. They would ask which one is better? Go online for information, we just tell you we have the vaccine." (VSP 15, female, Nanshan District, Shenzhen city).

Institutional Sub-Ecosystem
Many VSPs said that heavy workloads leave them insufficient time to communicate to parents about vaccines. As one VSP said: "[ . . . ]., I need to vaccinate more than 100 people a day. I remember a training I received before, it goes like, vaccination service provider should not vaccinate more than 50 people per day, otherwise, his/her working status will be negatively affected, and he/she may make mistakes, or not be able to communication well with parents, so the satisfaction of parents will decrease [ . . . ]." (VSP 26, female, Shushan District, Anhui province).
In addition to the heavy workload of vaccinating itself, vaccinators often mentioned two other reasons contributing to their increased workload. Firstly, since many vaccination clinics are not equipped with electronic information system, all the work, including the reminders for children's vaccination appointments and entry of vaccination information, needs to be done manually. Secondly, some vaccinators said that, in addition to their vaccination work, they are also given other public health responsibilities within their respective jurisdictions, such as a health check-up. As two participants said: "There is too much work to serve so many people. Now the requirements are so strict, and more and more detailed, right? Registration work, for example, can take you a whole morning if you write it by hand. If there is a set of electronic information system, first, it could alleviate the shortage of workforce, then avoid some mistakes [ . . . ]." (VSP 4, female, Dongzhi County, Anhui province).
"What I've been thinking is how to fulfill the annual work plan, I think a lot, but the plan just couldn't catch up with change. Our VSPs don't work only on vaccination, but also other types of work, such as poverty alleviation in rural area. Then scheduled work, such as professional improvement, will be disrupted. We also have to carry out physical examination for the elderly every year. It basically takes two months to complete the physical examination for the elderly in the whole town, and we work every day in two months." (VSP 10, female, Jingyang County, Shaanxi province).

Community Sub-Ecosystem
Many VSPs said that the cost of non-EPI vaccines is too expensive for local residents. High costs make them feel hesitant to recommend it to parents. As a director of a vaccination clinic described: "Especially Pentaxim, its price is very high, 500 or 600 Chinese yuan. I do not advocate this vaccine, because we are in rural areas, here residents' affordability is limited, right? Its demand is not large." (VSP 1, male, Dongzhi County, Anhui province).
Many VSPs also indicated that there was a shortage of non-EPI vaccines, such as flu vaccine. They said that they could not recommend it to parents if they did not have it in stock. One VSP commented: "For EPI vaccines, it's the leprosy vaccine, for non-EPI vaccines, it's Pentaxim, both vaccines are often out of stock. It was really difficult to conduct vaccination work at that time." (VSP 17, female, Nanshan District, Shenzhen city).

Public Policy Sub-Ecosystem
All VSPs indicated that the superior unit (District/County CDC) has clear assessment criteria for EPI vaccination rates but not for non-EPI vaccines. A director of a vaccination clinic described it as follows: "We will count how many children need to be vaccinated, how many children have been vaccinated. County CDC's assessment criteria is that the vaccination rate of EPI vaccines should be at least 95%. Depending on the percentage you reach, you reach 80% and you get 80% merit pay, if 90% and then 90% merit pay. There is no assessment for non-EPI vaccines." (VSP 42, male, Jingyang County, Shaanxi province).
All VSPs said that a small service fee can be charged for non-EPI vaccinations. However, their income was fixed and not related to the number of non-EPI vaccines they administer. Two vaccinators described: "We are paid a fixed salary. It has nothing to do with the number of non-EPI vaccine used." (VSP 24, female, Qingdu District, Shaanxi province).
"Non-EPI vaccines have no impact on our performance salary. Our work performance is generally assessed by the dean. It just depends on the working hours . . . Our performance income has nothing to do with the amount of EPI and non-EPI vaccination services. It's all arranged by the hospital [ . . . ]." (VSP 43, female, Jingyang County, Shaanxi province).

Discussion
This study used a mixed-method design to investigate the patterns and determinants of VSPs' communication and recommendation for non-EPI vaccines in the Chinese context. Only half (51.54%) of the VSPs often recommended non-EPI vaccines, and the low frequency of recommendation was independent of their individual characteristics. The VSPs routinely conducted health education about vaccination for the public. Most VSPs recommended non-EPI vaccines in an informed style, not a presumptive one, and provided the public with all decision-making latitude.
Recommendation from HCWs is regarded as one of the most consistent correlates of vaccination [24]. In our study, nearly half of the VSPs never or only sometimes recommended non-EPI vaccines, although they are full-time designated staffs in charge of vaccination services in China. The low level of recommendation practice is consistent with the previous surveys in China [16,25], but much lower than that in US and European countries [26][27][28][29]. Meanwhile, as for the style of VSPs' communication practice, the qualitative analysis showed that most VSPs did not recommend but instead informed parents about non-EPI childhood vaccines to honor the parents' decision-making autonomy. That is only information flows from VSPs to parents, but deliberation and decision on a vaccination option are delegated to parents, according to the framework of patient-provider interactions proposed by Charles et al. [17]. It has been shown that provider-driven communication through the shared and presumptive styles was highly effective for encouraging vaccination than the informed style [18]. Therefore, it is necessary to identify the factors associated with VSPs' communication practice.
The infrequent recommendation of non-EPI vaccines was widely prevalent among Chinese VSPs in this study sample, no matter their individual characteristics. This indicated that recommendation practices were possibly not influenced by individual characteristics. Among the five sub-ecosystems in the social-ecological model, interpersonal relationship, institutional arrangement, and public policy mainly contribute to the widely infrequent recommendation of non-EPI vaccines in an informed style instead of a presumptive one in China.
Firstly, at the interpersonal sub-ecosystem, some VSPs were concerned about potential conflicts arising from recommending the self-paid non-EPI vaccines and adverse reactions after administering these vaccines. The recommendation of paid medical services may lead to the patients considering doctors as retailers pursuing profits and reducing their adherence to the doctors' recommendation [30]. Discontentment from patients and doctors can even lead to the occurrence of adverse events [31]. In China, doctor-patient relationships has deteriorated during the past decade [32]. The tense doctor-patient relationship may be rooted in the Chinese health system with the long history of profit-pursuing medical behaviors and unaffordable medical services before the 2009 healthcare reform [33]. In addition, most parents have the low awareness on vaccine-preventable diseases due to the preventative nature of vaccines, which may make the public more adverse to being recommended vaccines than clinical services. To reduce the concerns of VSPs, it is necessary to address the tense doctor-patient relationship and improve the compensation mechanism for adverse reactions following vaccination.
Secondly, at the institutional sub-ecosystem, heavy workloads leave VSPs little time to communicate with the public about vaccines. With more vaccines introduced, public demand for vaccination has surged, leading to an inadequate number of VSPs available to meet the demand [34]. Our findings among VSPs who participated in the study in Shenzhen city, for example, were less likely to recommend non-EPI vaccines due to the pressure on their time than those in Shaanxi and Anhui provinces. There is a much higher proportion of young migrant workers and a more developed economy in Shenzhen than the other two provinces [35], which translated to a greater demand for local vaccination services. Moreover, a lack of electronic information systems also contributed to the overload of the VSPs. Dan Gong et al. found that insufficient infrastructure was one of the main barriers of delivering additional vaccines through the national EPI schedule [36]. While most provinces have an immunization information system capable of managing vaccine stocks and keeping official vaccination records, it cannot support vaccination services [37]. For example, vaccination appointment procedures were primarily traditional, using reservation books and oral notification. In addition, VSPs have to take on additional responsibilities, such as chronic disease management [38,39], and this extended work scope has exacerbated the shortage of the VSPs. Therefore, to ensure the reasonable workload of VSPs and leave time for health communications, the government should promote the construction of the whole-process health information system and strengthen the public health workforce through both retaining and recruiting staff, using financial and nonfinancial incentives [36].
Thirdly, at the public policy sub-ecosystem, there were no performance assessments or financial incentives for VSPs to recommend non-EPI vaccination. In China, there is strict performance assessment for EPI vaccination coverage for each vaccination clinic and CDC, and their performance is related to the staff merit pay, meanwhile there is no performance assessment for VSPs regarding non-EPI vaccination as their main responsibility is around EPI vaccination. Therefore, the non-EPI vaccination should also be covered as a part of performance assessment [36]. Moreover, to address the phenomenon of over-prescriptions, China issued the zero mark-up drug policy (including non-EPI vaccines) by disengaging prescribing from profits in 2009 [40]. It was reported that the policy promoted rational use of medicines [41,42]. Unlike drugs such as antibiotics, which are overprescribed and could lead to adverse health consequences, vaccines are preventative and need to be promoted by VSPs through incentives. Previous studies highlighted that HCWs' recommendation behaviors were notably influenced by financial incentives [43], and the financial incentives were effective in improving the uptake and delivery of health services [44][45][46]. However, implementing financial incentives could bring additional concerns, including neglect of non-incentivized tasks and distorted motivation among HCWs [47]. Thus, to avoid excessive and unnecessary non-EPI provision for economic benefit, it is important to establish an appropriate income distribution system [43], which could balance basic salary and performance-based incentives (e.g., avoiding overly high incentives and overly low basic salary) [48]. In addition to the above measures, governmental engagement can also contribute to the promotion of non-EPI vaccines. Taking rabies vaccine as an example, in order to meet the goal of eliminating dog-mediated rabies by 2030 [49], the Chinese government promoted rabies prevention education programs, particularly in high-risk provinces; meanwhile, the Chinese national reference laboratory for animal rabies provided training to more than 500 laboratory staff from provincial and municipal animal disease control centers [50]. These measures greatly improve the awareness of HCWs and the access to post-exposure prophylaxis, including the rabies vaccine. Since peaking in 2007 with more than 3000 reported human rabies deaths, substantial progress has been made in reducing these deaths [51].
Our study provides important insights into recommendation practices and the different communication styles among VSPs for non-EPI vaccines in China from a socioecological perspective. While previous studies have investigated HCWs' recommendation of influenza vaccines and their intrapersonal determinants (including knowledge and attitudes towards influenza and influenza vaccines) in China [15,16], there has been less attention paid to the influence of macro-level factors, such as demand, system capacity and public policy on VSPs' recommendation practice. Given that many childhood vaccines are optional and paid out of pocket in low-and middle-income countries [52], the implications of our study could be valuable for China and other countries with similar contexts.
Our study has several limitations. First, the recommendation behavior of VSPs was self-reported and potentially influenced by recall bias. Second, our study only covered three provinces, and our findings may not be generalized to all parts of China. Third, we only interviewed the VSPs who deliver vaccination services, and did not interview general HCWs who are not responsible for vaccination services, but may give health education on vaccination during clinical services. Finally, the study focused on recommendation for non-EPI vaccines in general. Recommendation behaviors may vary across different non-EPI vaccines, and further studies need to consider recommendation for specific non-EPI vaccines.

Conclusions
Our study reveals a low frequency of VSPs recommending non-EPI vaccines. Fears of potential conflicts with patients over recommending paid medical services, heavy workload, and the lack of performance assessment and financial incentive are the major barriers to VSPs' recommending practice. The multi-level ecosystem around non-EPI vaccination should be improved to incentivize and support VSPs and the public, which include a better system of legal redress to resolve potential disputes between the VSPs and the public, more effective workload management through the whole-process health information system and strengthening the public health workforce, the introduction of performance assessment and appropriate income distribution system for non-EPI vaccination, and more governmental engagement in infectious disease prevention programs. Informed Consent Statement: Informed consent was obtained from all subjects involved in the study. Written informed consent has been obtained from the patient(s) to publish this paper.

Data Availability Statement:
The corresponding author had full access to all the data in the study and had final responsibility for the decision to submit for publication.

Acknowledgments:
We thank the data collection teams from Fudan University, China CDC, the provincial and county CDCs, vaccination clinics who facilitated the research fieldwork, and all the survey participants.

Conflicts of Interest:
The Vaccine Confidence Project, which HL leads, receives collaborative grants with Astra Zeneca, GlaxoSmithKline, J&J, and Merck in addition to public sector grants. None of those research grants are related to this paper. Note: * the prices of non-EPI vaccine come from Shanghai public resources trading platform in 2022.

Appendix A
Appendix B Table A2. The questionnaire used in the quantitative survey. How often did you recommend Category II vaccines to the public?