The trends of socioeconomic inequities in full vaccination coverage among children aged 12–23 months from 2000 to 2017: evidence for mitigating disparities in vaccination service in Zhejiang province

ABSTRACT

Objective

This study aimed to evaluate the trends and changes in inequities in the completeness of the primary vaccination (CPV) scheduled before the first year of age among children aged 12–23 months, from 2000 to 2017. Methods: Data were extracted from five rounds of the provincial vaccination coverage survey (PVCS) in 2000, 2004, 2008, 2014 and 2017, respectively. The household income per month was used as an index of socioeconomic status for the inequity analysis. The concentration index (CI) was used to quantify the degree of inequity, and the decomposition approach was applied to quantify the contributions from independent factors to inequity in the CPV. Results: The CPV was significantly improved from 2000 to 2017, with 67.0% for the 2000 PVCS and 86.0% for the 2017 PCVS. The CI value decreased from 0.29839 for the 2000 round to 0.03601 for the 2017 round. The decomposition analysis indicated that independent variables such as birth order, ethnic group, mother’s education, maternal employment status, residence, immigration status and the percentage of the total health spending allocated to public health could explain the inequity in the CPV in varying degrees. Conclusions: A sharp reduction in socioeconomic inequity in the CPV was observed from 2000 to 2017. Policy recommendations to reduce the inequality in the CPV should focus on children with the risk factors found in this study, for better outcome in full vaccination and long-lasting herd immunity.

KEYWORDS:

• Completeness of the primary vaccination
• inequity
• concentration index
• decomposition

Introduction

Morbidity and mortality resulting from vaccine-preventable diseases (VPDs) can be significantly reduced through vaccination, which have played an important role in achieving one of the aims of United Nations Millennium Development Goal (MDG) 4 to reduce the mortality of children under 5 years by 2/3 in the period of 1990–2015.¹ Usually, the success of routine vaccination program is estimated by the incidence of VPDs and the implementation of vaccination program is evaluated by the vaccination coverage, especially by the coverage for vaccinations scheduled before first year of age among children aged 12–23 months.²

In China, the expanded program on immunization (EPI) has been initiated against six target diseases (four vaccines) since 1978, while it has been expanded to 12 target diseases (11 vaccines) since 2008. As we know, the thresholds of vaccination coverage to stop the transmission of VPDs are different for different VPDs. However, the Chinese Advisory Committee on Immunization Practice (CACIP) sets the same goal of 90% for the vaccination coverage for all antigens. There may be two reasons: first, it is required to reach herd immunity and stop the transmission of the most contagious VPDs. Furthermore, the coverage is set at 95% with 2 doses of measles-containing vaccine (MCV) for interrupting measles transmission.³ Second, a unified target is easy to implement in practice in China.

In this study, we defined the primary vaccination as the vaccinations scheduled before first year of age, which included one dose of Bacillus Calmette–Guérin vaccine (BCG) scheduled at birth; three doses of hepatitis B vaccine (HBV) scheduled at birth, 1 month and 6 months of age; three doses of diphtheria-tetanus-pertussis combined vaccine (DTaP) scheduled at 3, 4 and 5 months of age; three doses of polio vaccine (PV) scheduled at 2, 3 and 4 months of age; one dose of MCV scheduled at 8 months of age; one dose of Japanese encephalitis vaccine (JEV) scheduled at 8 months of age; and two doses of meningococcal polysaccharide vaccine type a (MenV-a) at 6 and 9 months of age. It is recommended that all children receive the primary required 14 vaccinations before 12 months of age and all vaccinations included in EPI are free of charge.

With the great efforts of EPI, China has achieved the excellent success and maintained existing high vaccination coverage. For example, China has eradicated poliomyelitis since 2000,⁴ and the infection rate of hepatitis B among children under 5 years has reduced from 4.05% in 1992 to 0.32% in 2014.⁵ However, the provincial vaccination coverage in some sub-population has decreased as well as significantly fluctuated.

Since the average vaccination coverage is inadequate in reflecting the performance of EPI and may mask the gap between the different sub-populations, the inequity in this indicator has become more prominent than before. Traditionally, the concentration index (CI) and concentration graph approaches were used to evaluate the inequities in health outcomes as well as the vaccination coverage rate.⁶ Besides, some studies decomposed the general inequity to detect the pathways or determinants of the observed health inequities.^7,8

Data showing disparities in vaccination coverage at sub-provincial level were not previously monitored in China, despite the programmatic potential for identifying communities at risk for VPD outbreaks. Based on our previous analysis, a potential inequity in vaccination coverage within Zhejiang province existed, which could have a substantially negative influence on EPI.⁹ Achieving the high general coverage of childhood vaccination involves creating equal accessibility, as well as identifying the vulnerable children who are facing the risk of under-vaccination. To our knowledge, some determinants from child-specific, parental and household levels had been identified as the risk factors for suboptimal vaccination coverage in previous studies.^10–12 For policy-making purposes, it is critical to detect the trends in inequity in vaccination coverage over the years and to understand the determinants which contribute to the EPI implementation and its effectiveness in the sub-populations.

Therefore, the aim of this study was to evaluate the trends and changes in the socioeconomic inequalities in vaccination coverage of the primary vaccination among children aged 12–23 months from available data of the five rounds of the provincial vaccination coverage survey (PVCS) in 2000, 2004, 2008, 2014 and 2017, by using the completeness of the primary vaccination (CPV) as a main indicator.

Methods

Data source

The rounds of the PVCS were developed upon the household-based cluster survey method recommended by WHO to monitor the EPI performance.¹³ These PVCSs were usually conducted at provincial level with an interval of 3–5 years. In this study, we used the raw data from the 2000, 2004, 2008, 2014 and 2017 round of PVCS.

Zhejiang is a middle-income, developing province in east China, with an area of 1105.5 square kilometers. Administratively, it is divided into 11 cities, 90 counties and 1319 towns. The total population of Zhejiang province was 54.8 million in 2000, 58.7 million in 2004, 63.2 million in 2008, 66.1 million in 2014 and 70.4 million in 2017. The average annual birth rate ranged between 1.0% and 1.3% in these 5 years, with a peak of 1.3% in 2017. The increase in annual birth in recent years was due to the opening of the two-child policy in China since 2015.

Sampling design

The sample size of all five rounds of PVCS wascalculated based on the formula as follows:¹³ $N=deff\times \frac{z_{\left(1-\frac{\alpha }{2}\right)}^2\times p\times \left(1-p\right)}{d^2}$. For the 2000, 2004 and 2008 round of PVCS, the parameters were set as a two-tailed α error of 5%, a permissible error (d) of 0.1, a design effect (deff) of 2, and the expected CPV at 0.5. The final sample size required for each city was 192, corresponding to 2112 at provincial level. For 2014 round, the permissible error (d) was changed to 0.08 and the expected CPV was changed to 0.7. The final sample size required for each city was 252, corresponding to 2772 in total. For 2017 round, the permissible error (d) was changed to 0.08 and the expected CPV was changed to 0.9. The final sample size required for each city was 70, corresponding to 770 in total.

Field survey procedures

For the 2000, 2004, 2008 and 2014 round of PVCS, the sample size at city level was divided in six towns (cluster) of 32 or 42. For the 2017 round, the sample size at city level was divided in seven towns of 10. Although the sample size at cluster level was different in the different rounds of PVCS, the field survey procedures were similar: first, target towns for each city were selected from a list of all towns (with the population size), on the basis of the probability proportional to population size. Second, one community was randomly selected from each town through the simple ballot from the list of all communities. Third, the index household was selected randomly from the list of all households by using the table of random numbers. Fourth, the adjacent household right to the previous one was selected. Only one eligible child per household was chosen. Households were re-visit if there was somebody living but without any response and they were excluded after three attempts. The closest community was selected to survey the rest children if adequate sample could not be obtained in the selected community.

Information on the vaccination was collected from immunization cards held by caregivers. For the 2000 and 2004 round of PVCS, the vaccination information of children who missed the immunization cards could be found in the immunization records kept by vaccination providers. For the other rounds of PVCS, the vaccination information of children who missed the immunization cards could be found in Zhejiang provincial immunization information system, which covered all vaccination clinics since 2005. If there were no written records or the immunization cards were missing, we considered these children as under-immunized.

Dependent variable

The main study outcome variable (or dependent variable) was whether or not children aged 12–23 months completed the primary vaccination schedule in the five rounds of PVCS (completed = 1, not completed = 0). Since the vaccination doses for the primary vaccination were different in these five rounds, we gave the three different definitions of CPV for the five rounds of PVCS as follows: (1) the CPV of the 2000 round was defined that children received one dose of BCG, three doses of DTaP, three doses of PV and one dose of MCV; (2) the CPV of the 2004 round was defined that children received one dose of BCG, three doses of HBV, three doses of DTaP, three doses of PV and one dose of MCV; (3) the CPV of the 2008, 2014 and 2017 rounds was defined that children received one dose of BCG, three doses of HBV, three doses of DTaP, three doses of PV, one dose of MCV, one dose of JEV and two doses of MenV-a.

Independent variables

Two sets of independent variables were included. First set was individual/household variables including sex and birth order and ethnic group of children, education level and age of mother, maternal employment status, household socioeconomic status, household resident status and household immigration status. Individual/household variables were collected through a pre-tested, standard questionnaire developed by Zhejiang Provincial Center for Disease Control and Prevention. All variables and their levels were similar in the five rounds of PVCS. Second set was city-specific variables including the gross domestic product (GDP) per-capital and the percentage of public health spending of the total health spending. All city-specific variables extracted from the statistics of Zhejiang province of the years when the PVCSs were implemented.

The socioeconomic distribution was categorized into tertiles according to the household income per month, with the first tertile representing the poorest group and the last tertile representing the least poor group.

Analytical method

The analytical procedures included three aspects as follows: First, the distribution of dependent and independent variables was described. Second, the degree of socioeconomic inequity in CPV was assessed by the CI through the formula:¹⁴ $\mathrm{C}\mathrm{I}=\frac{2}{\mu }\mathrm{c}\mathrm{o}\mathrm{v}\left(h,r\right)$. Here, h represented the values for the CPV of each observation, and μ was the mean percentage of CPV, and r indicates the rank of the household socioeconomic status, with i = 1 for the most disadvantage and i = N for the least. Third, we applied the concentration curve (CC) to estimate the degree of socioeconomic inequity in CPV. The CC plotted the cumulative percentage of the CPV (y-axis) against the cumulative percentage of the children, which was ranked by household income per months, beginning with the poorest and ending with the least poor (x-axis). If there was a perfect equity in CPV, the CC would be a 45-degree line (the diagonal line) running from the bottom left-hand corner to the top right-hand corner. Fourth, logistic regression model ($\mathrm{L}\mathrm{o}\mathrm{g}\left({\hat{p}}_i\right)=\sum _k{\beta }_k{x}_{ki}+{\epsilon }_i$) was applied to detected the determinants of the CPV, including the significant variables in the univariate analysis (p < .1). The $\mathrm{L}\mathrm{o}\mathrm{g}\left({\hat{p}}_i\right)$ was the logistic of the predicted probability of being full vaccinated, and the ${\beta }_k$ was the coefficient for the risk factor $x_k$, and the ${\epsilon }_i$ was the residual estimate. The CI could be represented as:¹⁴ $\mathrm{C}\mathrm{I}=\sum _k\left(m_k{\bar{x}}_k/\mu \right)C{I}_k+G{C}_{\epsilon }/\mu$, where μ was the mean percentage of the CPV, ${\bar{x}}_k$ was the mean of the $x_k$, $C{I}_k$ was the CI of the $x_k$. The marginal effect was defined as $m_k=\lambda \left(\sum _k{\beta }_k{\bar{x}}_k\right){\beta }_k$. The $\lambda ()$ was the logistic density function. The $m_k$ meant the average change in the probability of the full vaccination when the risk factor $x_k$ changed one unit. The $G{C}_{\epsilon }$ meant the CI of the residual (${\epsilon }_i$) was defined as: $G{C}_{\epsilon }=\frac{2}{N}\sum _{i=1}^N{\epsilon }_i{r}_i$. $\left(m_k{\bar{x}}_k/\mu \right)C{I}_k$ was the contribution of the CI. The “bootstrap” procedure in a five-step manner to obtain the standard errors (S.E.) of the CI was also applied.

All statistical analyses were conducted by Stata^® 15 (StataCorp LLC, USA), using the weighting of variables for children in the dataset. The Coindex package for the Stata^® software was used to calculate the CI and to draw the chart. A two-tailed statistical significance of 0.05 was set in this analysis.

Results

The average percentage of the CPV of the total five rounds was 75.7% (7482/9878), with 67.0% for the 2000 PVCS, 70.0% for the 2004 PVCS, 77.0% for the 2004 PVCS, 83.0% for the 2014 PVCS and 86.0% for the 2017 PCVS, respectively. In the univariate analysis, high birth order of children, minority ethnic group, low maternal education, mother with fixed jobs, migrant household were found as the significant risk factors for CPV. Furthermore, younger mother and rural residence were associated with the CPV in the 2008, 2014 and 2017 round of PVCS, and the higher socioeconomic status of household was associated with the CPV in the 2000, 2004, 2008 and 2014 round of PVCS (Table 1).

Table 1. Demographic and socioeconomic characteristics and the completeness of the primary vaccination (CPV) during the five rounds of provincial vaccination coverage survey (PVCS) in 2000, 2004, 2008, 2014 and 2017

Variables	Round 2000			Round 2004			Round 2008			Round 2014			Round 2017
	CPV		p	CPV		p	CPV		p	CPV		p	CPV		p
	n/N	%		n/N	%		n/N	%		n/N	%		n/N	%
Gender of child			>0.05			>0.05			>0.05			>0.05			>0.05
Male	707/1055	67.0		740/1056	70.1		810/1054	76.9		1152/1388	83.0		327/382	85.6
Female	708/1057	67.0		738/1056	69.9		816/1058	77.1		1149/1384	83.0		335/388	86.3
Birth order			<0.01			<0.01			<0.01			<0.01			<0.01
1	1193/1689	70.6		1208/1680	71.9		1321/1667	79.2		1665/1889	88.1		443/496	89.3
2	211/353	59.8		240/350	68.6		278/378	73.5		533/655	81.4		198/234	84.6
≥3	11/70	15.7		30/82	36.6		27/67	40.3		103/228	45.2		21/40	52.5
Ethnic group			<0.01			<0.01			<0.01			<0.01			<0.01
Han	1339/1816	73.7		1352/1788	75.6		1398/1710	81.8		2145/2506	85.6		624/708	88.1
Others	76/296	25.7		126/324	38.9		228/402	56.7		156/266	58.6		38/62	61.3
Age of mother (years)			>0.05			>0.05			<0.05			<0.01			<0.01
<30	1172/1762	66.5		1246/1788	69.7		1375/1766	77.9		1588/1816	87.4		594/682	87.1
≥30	243/350	69.4		232/324	71.6		251/346	72.5		713/956	74.6		68/88	77.3
Maternal education			<0.01			<0.01			<0.01			<0.01			<0.01
No/primary	65/241	27.0		63/206	30.6		63/176	35.8		56/132	42.4		17/29	58.6
Second level	1018/1486	68.5		1071/1514	70.7		1128/1446	78.0		1465/1775	82.5		384/451	85.1
College level	332/385	86.2		344/392	87.8		435/490	88.8		780/865	90.2		261/290	90.0
Maternal employment			<0.01			<0.01			<0.01			<0.01			<0.01
Home fulltime	188/204	92.2		198/218	90.8		214/235	91.1		360/388	92.8		169/183	92.3
Employed	1227/1908	64.3		1280/1894	67.6		1412/1877	75.2		1941/2384	81.4		493/587	84.0
Residence			>0.05			>0.05			<0.05			<0.01			<0.01
Urban	709/1058	67.0		730/1055	69.2		772/1054	73.2		1068/1368	78.1		318/386	82.4
Rural	706/1054	67.0		748/1057	70.8		854/1058	80.7		1233/1404	87.8		338/384	88.0
Immigration status			<0.01			<0.01			<0.01			<0.01			<0.01
Resident	1374/1821	75.5		1336/1690	79.1		1167/1377	84.7		1611/1802	89.4		400/448	89.3
Migrant	41/291	14.1		142/422	33.6		459/735	62.4		690/970	71.1		262/322	81.4
Socioeconomic status			<0.01			<0.01			<0.01			<0.01			>0.05
Q1 (poorest)	416/704	59.1		432/704	61.4		501/704	71.2		730/924	79.0		215/257	83.7
Q2	472/704	67.0		510/704	72.4		545/704	77.4		770/924	83.3		221/257	86.0
Q3 (least poor)	527/704	74.9		536/704	76.1		580/704	82.4		801/924	86.7		226/256	88.3

χ² test was applied to detect the significance of the difference in CPV between levels of each variable.

The disparities in CPV in the five rounds of PVCS showed that the CPV was consistently lower in the poorest group than the least poor group in common, but the inequity of CPV decreased from the 2000 round to the 2017 round, with the highest difference of 6.5% observed in the 2000 round and lowest difference of 2.1% observed in the 2017 round (Figure 1).

Figure 1. The distribution of the completeness of the primary vaccination by socioeconomic status in the five rounds of provincial vaccination coverage survey from 2000 to 2017

The CI value was 0.08413 (95% CI: 0.04628–0.12197) for the five rounds of PVCS from 2000 to 2017, with the CI of 0.29839 (95% CI: 0.19614–0.40064) for the 2000 round of PVCS, 0.26176 (95% CI: 0.17340–0.35012) for the 2004 round, 0.24703 (95% CI: 0.15975–0.33431) for the 2008 round, 0.09594 (95% CI: 0.02393–0.16795) for the 2014 round and 0.03601 (95% CI: −0.02724–0.08966) for the 2017 round. Although the inequity of the CPV was increasing narrow, Figure 2 showed that most of the CCs were below the line of equity, indicating that children in the highest socioeconomic stratum were most likely to complete the primary vaccination (except for the 2017 round).

Figure 2. The concentration curve of inequity in the completeness of the primary vaccination in the five rounds of provincial coverage survey from 2000 to 2017

The CI of each factors significant in the univariate analysis as well as its relevant contributions to the inequity in the CPV in the five rounds of PVCS were presented in Table 2. We found that the birth order one, the Han ethnic group, the high maternal education level, the home fulltime mothers, living in rural areas, the resident children and the high percentage of the public health spending of the total health spending were significantly associated with the concentration of the CPV in the least poor group.

Table 2. The concentration index and contributions of determinants to inequities in the completeness of the primary vaccination (CPV) during the five rounds of provincial vaccination coverage survey (PVCS) in 2000, 2004, 2008, 2014 and 2017

Variables	Level	Concentration index (95%CI)^a	$m_k$^b	${\bar{x}}_k$^c	% of contribution
CPV	–	0.08413 (0.046282–0.121978)
Birth order of child	≥3	Ref
	2	0.03011 (0.02203–0.05201)	0.04211	0.1922	0.4
	1	0.14213 (0.10223–0.18104)	0.18103	0.5842	14.2
Ethnic group of child	Others	Ref
	Han	0.15203 (0.11006–0.19203)	0.13042	0.5012	15.6
Maternal education level	No/primary	Ref
	Second level	0.01163 (0.00805–0.12108)	0.03411	0.1228	0.1
	College level	0.16829 (0.13178–0.20393)	0.14582	0.5988	23.1
Maternal employment status	Employed	Ref
	Home fulltime	0.05772 (0.03671–0.0824)	0.05017	0.2576	1.2
Residence	Urban	Ref
	Rural	0.07219 (0.04736–0.10561)	0.09767	0.3618	4.0
Immigration status	Migrant	Ref
	Resident	0.18034 (0.12625–0.22182)	0.16223	0.5722	26.3
Per-capital GDP (CNY)		0.00562 (−0.00277–0.00836)	0.00001	8533	0.8
% of public health spending of the total health spending		0.10255 (0.08265–0.15355)	0.02033	4.0274	13.2

^a95% of confidence interval.

^bThe marginal effect of $x_k$.

^cThe mean of $x_k$.

Discussion

Chinese EPI is considered as one of the most cost-effective public health interventions as it has remarkably reduced the morbidity and mortality of the VPDs. For example, small pox and poliomyelitis has been eradicated and the infection rate of hepatitis B and measles has fallen to the lowest level in the history. However, China, like many developing countries, still faces the challenges of inequity of vaccination coverage that may lead to sub-population or the specific geopolitical zones with a vaccination coverage under the threshold set by CACIP, which may induce the endemic of VPDs.

By analyzing the continuous rounds of PVCS, an increasing achievement in the CPV was observed, with a dramatic growth from 67.0% in 2000 round to 86.0% in 2017 round. This trend might be explained by the efforts of EPI, especially by the policy changes in vaccination service in Zhejiang province in the last decade. First, on the basis of the overall picture of Zhejiang provincial growing socioeconomic context in the last decade, the increase in public spending for EPI might have a positive impact on the trend of the CPV. At the meantime of the increasing investment from Zhejiang provincial government, the increased rate of support on vaccination higher than other public service was also observed in recent years. As the vaccine procurement was covered by Chinese central government, the provincial budget mainly covered these aspects as follow: the first aspect was the financial incentives to providers (5 CNY per dose), which could encourage the providers to actively search drop-out or under-immunized children; the second was to support two rounds of EPI training covered all vaccinators in Zhejiang province in 2008 and 2014; the third was to invest 20 million to replace and update the cold chain equipment in all vaccination clinics. Second, more than 95% of the vaccination clinics extended opening time or increased the frequency of sessions to improve the accessibility of service since 2014. Third, CACIP had set the milestone of EPI for measles elimination by 2012,¹⁵ maintaining polio eradication and further reducing the infection rate of hepatitis B among children under 5. It also required all vaccination providers to work hard on achieving or maintaining a high CPV among children before their first birthday.

Our study not only shed light on the decreasing trend in socioeconomic inequality in the CPV during the five rounds of PVCS but also provided important insights for the public health system regarding the needs of specific interventions target on addressing the remaining socioeconomic inequity in the CPV in Zhejiang province. In this study, the socioeconomic inequity in CPV sharply decreased as the CI decreased from 0.298 in 2000 to 0.036 in 2017. However, the inequity in CPV between the poor and the wealth still remained. Previous evidence had demonstrated that some factors of children and their families could impede childhood vaccination.¹⁶ Furthermore, household income was considered as a key factor of receiving the vaccination service during the first year of life in low-income or developing countries.^17,18 In this study, although the inequity was reducing from the 2000 round to the 2017 round, we still found that children from wealthier households were more likely to complete the primary vaccination compared with those from the poorer households. It was consistent with the previous findings in Africa,¹⁹ which reported that the inequality existed among children from poor households. Our finding indicated that the poverty was an important determinant for the utilization of vaccination service during the surveyed time period. The reasons for the observed lower CPV among children from the poorer households might include: (1) a negative attitude on the acceptability of childhood vaccination or a misunderstanding of the EPI vaccines that needed out of pocket expense; (2) caregivers might face a significant indirect cost (travel fee or working time lost) when they bring their children to get vaccinations; (3) poor people preferred to spend time on salary-generating activities to improve the living standard or (4) limited freedom of decision making due to the poor knowledge on vaccination. However, we did not observe the inequity in CPV between the poor and the wealth in the 2017 round of PVCS, which might be considered as a persistent effort of Zhejiang provincial EPI in perfecting the service network from grassroots level to provincial level, making sure all EPI vaccines be accessible and free to both the wealth and the poor children.

Child of a high birth order probably had an incomplete primary vaccination, which was in line with the previous studies.^20,21 The possible explanation was the caregiver’s fatigue on vaccination. It meant that the demand of receiving vaccination waned if the number of children increased in the same family. Besides, our finding might reflect that the resource competition among siblings would become fiercer in a household with more children, leading to the neglect of vaccination. On the other hand, we suggest that community mobilization and advocation should be implemented to raise the importance of vaccination and the awareness of bring all children to get vaccination.

China is known as a multiple-ethnic nation, with 56 ethnic groups in total, and each ethnic group has its own culture characteristics. We found ethnicity was a key factor existing as the possibility of the CPV favored the Han group. This finding was consistent with the reports from other developing countries,^22,23 where ethnic minorities were marginalized and lack education, and they probably lived in a lower socioeconomic status leading to a lack of the awareness of primary healthcare. Another reason could be the negative cultural beliefs behind the acceptance of vaccination in the ethnic minorities. However, an attenuated effect from ethnicity was observed in recent rounds of PVCS, which could be explained by the social mobilization of policy of EPI and vaccination knowledge toward population from minority ethnic groups was strengthened and improved in recent years. We recommend that a special focus should be continuously necessary to increase vaccination awareness among the ethnic minorities in future.

After the decomposition analysis of the determinants of the inequity in CPV, maternal education level stood out as one of the largest contributors. This result was consistent with the previous studies from other developing countries and in other setting of health service.^24,25 It indicated that education advantaged mothers exerted a positive effect on the CPV, though improving the accessibility of knowledge or information on vaccination, and facilitating the communication with health providers. We recommend that it should be necessary to improve the health education platform to inform mothers on the importance of vaccination, leading to a better understanding of vaccination schedules and practice. Additionally, we suggest that awareness on vaccination should be created in mothers or other decision-makers in the family, irrespective of literacy status.

According to our results, employed mother was a negative factor of fully receiving the primary vaccination. We assumed that mother with fixed jobs might have logistical difficulties in bringing their infants on time for vaccination, due to where they work and the hours they work.

Previous reports from some developing countries had indicated the residence area as a determinant of CPV, with a higher CPV observed in urban area.^26,27 These studies assumed the possible reason lied in the increased availability and quality of vaccination service in urban areas. However, our finding was different from the previous reports, with a significant rural advantage in the CPV. Similarly, one domestic study also found that there was a rural–urban disparity in the coverage of seasonal influenza vaccine in elderly people under the free vaccination policy of influenza.²⁸ It might be explained by the differing health provision system. After the implementation of the policy of reform and opening up by Chinese central government, the process of urbanization was pushed forward rapidly due to the high-speed economic development in Zhejiang province. The connection between providers and caregivers was lacking or loose in urban areas, and the providers were less likely to disseminate the vaccination knowledge or to conduct the health education to parents and residents. Conversely, vaccination providers had a closer relationship with parents in rural areas. It was convenient to inform and encourage parents to bring their children to get vaccinations. Thus, we recommend that health education should be enhanced through the diverse methods to improve the parent’s awareness on vaccination and the equity of CPV.

Immigration status was observed as the largest contributor to the inequity of CPV in this analysis, with the lower CPV favored in the migrant children. The significant association between migrant children and low vaccination coverage had been identified in our other studies.^29,30 Similarly, the possible reason was the challenges faced by migrant people when they flown into a new living sociocultural environment, such as employment, children’s education, housing, medical security, social insurance, etc. Any maladjustment from these aspects would influence the accessibility to or awareness of the utilization of vaccination service. On the contrary, local people were better able to avail themselves of vaccination services, as they were well adapted to the current living areas.

The last determinant of CPV was the proportion of public health spending of the total health spending. Our result was in line with a study from Taiwan,³¹ which found the difference in appendicitis cases narrowed as the national health insurance program was made universal. We suggest that all local governments increase the investment in public health service to improve the service utilization and its quality.

There were several limitations worth noting. First, reporting bias might occur in collecting some sensitive information, such as the monthly household income. We assumed that this bias would probably underestimate the CI value due to the socioeconomic advantage people might underreport the monthly household income. Second, this analysis did not assess the cultural aspects related to childhood vaccination due to the study design and data limitation, such as the acceptability and attitudes toward the uptake of vaccination among parents. Cultural beliefs and preferences of parents might affect their vaccination service-seeking behavior and function as a confounding factor or an interacting variable in the evaluation of the association between the independent variables and the CPV. Third, the data from the cross-sectional study design could not be assessed sequentially by month for the outcome variable in all five rounds of PVCS. Additionally, the outcome variable was obtained from multiple vaccine doses scheduled before the first year of life, so the assessment focusing on the month was difficulty to be performed. Forth, we could not evaluate the disparities in the quality and convenience of vaccination services between poor and wealthier communities as those data were not collected in this study. The limitation would be addressed in future surveys by collecting the accessibility, quantity and quality of health workers, service satisfaction from caregivers, communications, etc.

Conclusions

This study indicated a sharp reduction in socioeconomic inequity in the CPV among children aged 12–23 months in Zhejiang province over a decade, from 2000 to 2017. The socioeconomic advantaged children tended to complete the primary vaccination more actively than the disadvantaged counterpart. Based on our analysis, we recommend the interventions as follow to maintain or promoting an equal and high full vaccination coverage: first, increasing the governmental investments on vaccination service at all levels by increasing the budget on procurement of vaccines and incentives to health workers. Second, adjusting the vaccination sessions to reflect the current needs of the community to improve the accessibility. Third, enhancing the vaccination service system by implementing the training target on health workers and conducting the performance assessment of vaccination clinics. Forth, implementing the health education on the awareness and importance of vaccination target people with low education background or from poor household, ethnic minority groups and migrant population.

Disclosure of potential conflicts of interest

No potential conflicts of interest were disclosed.

Author contributions

Y.H. and HK.L. conceived and designed the experiments; H.L. and Y.C. performed the experiments; Y.H. and Y.W. analyzed the data; Y.W. contributed reagents/materials/analysis tools; Y.H. wrote the paper.

Ethics approval and consent to participate

This study was approved by the ethical review board of Zhejiang provincial CDC. Written informed consent was obtained from a parent or a legal caregiver of each eligible child enrolled in this study.

Acknowledgments

The authors would like to thank the immunization staff from 11 CDCs at city level for their investigation and data collection.

Additional information

Funding

This study was funded by medical and health science and technology project of Zhejiang province (Grant number: 2020KY522).