Factors associated with adults' perceived need to vaccinate against hepatitis B in rural China

ABSTRACT

Hepatitis B virus (HBV) infection is a serious public-health issue in China. While the hepatitis B vaccine is effective in preventing HBV infection, the HBV vaccination coverage rates among Chinese adults remain low. From a survey of rural adults from 7 provinces in China, we identified a unique HBV at-risk group: rural adults who had no history of HBV vaccination and had no plan to HBV vaccinate in the future. We divided this ‘no history-no plan’ group into those who identified No-need to vaccinate and those that perceived a Need to vaccinate (even if they had no plan to do so). We found age, marital status, health status, perceived HBV infection environment, perceived HBV infection risk and perceived HBV infection severity explained differences between the ‘No-need’ and ‘Need” to HBV vaccinate groups. Education, occupation and knowledge of hepatitis B and HBV transmission were not associated with HBV vaccination need. Our results showed that free HBV vaccinations and reimbursement for vaccinating could significantly increase the HBV vaccinate take-up rate for both Need and No-need rural adults. A tailored public health HBV campaign, especially targeting the No-need subgroup, would increase vaccination rates by better informing rural adults about HBV transmission routes, the dangers of HBV infection, the effectiveness of HBV vaccinations and the safety of HBV vaccinations.

KEYWORDS:

• HBV
• HBV vaccinations
• need for HBV vaccinations
• perceived HBV risks
• rural adults

Introduction

Hepatitis B virus (HBV) infection is a serious public-health issue in China.¹ HBV kills up to 300,000 people in China every year,¹ mostly by cirrhosis and hepatocellular carcinoma which are common long-term sequelae of chronic HBV infection.² The positive rate of hepatitis B surface antigen (HBsAg) in the Chinese population (aged 1–59) was 7.2% in 2006, and an estimated 93 million people are HBV carriers in China.³ In a study of the economic burden of chronic hepatitis B in Beijing and Guangzhou,⁴ Hu and Chen found the disease imposed a ‘substantial economic burden on patients, families and the society in China urban areas’. In Shandong province, Lu et al. estimated the in-patient costs of hepatitis B-related diseases ranged from $US2954 to $US10635 per patient,⁵ and Zhang et al. found that in Jiangsu province the direct economic costs of HBV-related diseases ranged from $US107 for out-patients to $US3193 for in-patients.⁶ All these studies advocated the need for public health officials to intervene to reduce HBV-related diseases.

Vaccines are regarded as an effective public health intervention when combating infectious diseases.⁷ The hepatitis B vaccine is an excellent intervention for preventing HBV infection,^8,9 both substantially reducing the economic burden ofhepatitis B-related diseases,¹⁰ and attenuating discrimination against hepatitis B patients or HBV carriers.¹¹ For all newborns in China, the Expanded Programme on Immunization (EPI) included free hepatitis B vaccines from 2002 and free HBV vaccines and vaccinations from 2005. Catch-up vaccination was performed from 2009 to 2011, for children born from 1994 to 2001. It has been estimated the Chinese HBV vaccination program have prevented more than 4 million HBV-related deaths.¹² Vaccination rates are generally lower in rural areas than in urban areas,¹³ and lower among adults than children.^14-16 Currently, HBV vaccination of adults is not publicly funded under China's immunization policy. Compared to children, adults who are infected with HBV are less prone to develop chronic HBV infection, but HBV vaccinations for adults are still beneficial for several reasons: it reduces acute infections; it indirectly protects children, especially by reducing the number of non-immune women who contract HBV from their partners before or during pregnancy; and it raises public HBV awareness, reducing widespread HBV stigma in China. A recent economic evaluation has found that vaccination of adults may be cost-effective.¹⁷

Previous studies of adults' HBV vaccinations have mainly focused on demographic and socioeconomic factors related to HBV vaccination,^15-20 or on high-risk groups such as drug-users, surgeons, oral healthcare personnel and other healthcare workers.^21-25 As part of a larger project, we conducted a comprehensive HBV survey in 42 villages in 7 provinces.¹⁵ Our focus was on the adults in the general rural population who have not received a HBV vaccination (no history) and who do not have current plans for getting vaccinated (no plan), which comprised about one third of all adult participants. We divided this no history-no plan sample into 2 groups based on their perceived need for HBV vaccinations: ‘Need’ and ‘No-need’. The No-need group poses a unique challenge to achieving high vaccination rates. Using descriptive statistics and logistic regressions, we analyzed the (No) Need to vaccinate and the potential influencing factors of the perceived need to take the HBV vaccine.

Results

Our sample consisted of 3050 unvaccinated adults without plans for obtaining the HBV vaccination. When asked why they did not want to get vaccinated, the pre-defined responses for not getting HBV vaccinated comprised: do not need the vaccine (69.67%); the price is too high (13.18%); worrying about the safety and side-effect of the vaccine (1.77%); and not confident about the vaccine effectiveness (1.48%). The open-ended “other reasons” (13.90%) responses included: respondents were uninformed about when and where to get HBV vaccination; respondents did not know about the HBV vaccine; and respondents were too old. Our analysis compared 2 subgroups: those who stated that they did not need the HBV vaccine (No-need, N = 2125) and those that stated they needed the vaccine (Need, N = 925). In the Need group, 58.73%, and in the No-need group, 46.46%, stated they would take the vaccine if it was offered for free (p-value = 0.000).

Table 1 presents definitions of the variables and their assigned values. Females were overrepresented in the sample (59.51%), but gender was not significantly associated with Need/No-need (p-value 0.971). The age level, marital status, educational level, occupation, and relative health status were significantly associated with Need/No-need (0.000 < p-value < 0.043). The variables measuring respondents' HBV-related perceptions (perceived HBV infection around, perceived risk of HBV infection, perceived severity of HBV infection) were all associated with Need/No-need (0.002 < p-value < 0.038), while hepatitis B knowledge was not (p-value 0.738).

Table 1. Descriptive statistics for independent variables (N=3050).

Variable (value)	Variable definition^*	%	n	Need	No-need	P^**
Male	male	40.49	1235	375	860	0.971
Female	female	59.51	1815	550	1265
Agelevel (1)	aged 18–27	4.52	138	56	82	0.017
Agelevel (2)	aged 28–37	11.70	357	123	234
Agelevel (3)	aged 38–47	23.67	722	212	510
Agelevel (4)	aged 48–57	23.64	721	202	519
Agelevel (5)	aged 58-	36.46	1112	332	780
Not married	never married or divorced or widowed	11.38	347	127	220	0.007
Married	married	86.62	2703	798	1905
Education level (1)	years of schooling ≤ 6 years	57.97	1768	568	1200	0.036
Education level (2)	6 < years of schooling ≤ 9 years	34.49	1052	290	762
Education level (3)	years of schooling > 9years	7.54	230	67	163
Occupation (1)	jobless	6.82	208	59	149	0.043
Occupation (2)	farmer	74.89	2284	717	1567
Occupation (3)	migrant worker	9.41	287	86	201
Occupation (4)	professional or technical personnel etc.	7.80	238	55	183
Occupation (5)	staff in government/doctors/teachers	1.08	33	8	25
Relative health status (1)	health status is worse or much worse	14.26	435	173	262	0.000
Relative health status (2)	health status is about the same	16.98	518	153	365
Relative health status (3)	health status is better or much better	65.75	2097	599	1498
Hepatitis B knowledge (1)	Hepatitis B knowledge score is 0–8	34.92	1065	318	747	0.738
Hepatitis B knowledge (2)	Hepatitis B knowledge score is 9–16	47.28	1442	447	995
Hepatitis B knowledge (3)	Hepatitis B knowledge score is 17–26	17.80	543	160	383
Perceived HBV infection around (1)	infection is decreasing in the village	25.31	772	215	557	0.002
Perceived HBV infection around (2)	infection rate has no change	68.98	2104	638	1466
Perceived HBV infection around (3)	infection is increasing in the village.	5.70	174	72	102
Perceived risk of HBV infection (1)	unlikely or very unlikely	66.07	2015	584	1431	0.004
Perceived risk of HBV infection (2)	unknown	25.54	779	241	538
Perceived risk of HBV infection (3)	likely or very likely	8.39	256	100	156
Perceived severity of HBV infection (1)	HBV is not serious for adults > 18years.	8.30	253	63	190	0.038
Perceived severity of HBV infection (2)	unknown	38.26	1167	340	827
Perceived severity of HBV infection (3)	HBV is serious or very serious for adults > 18 years	53.44	1630	522	1108

Notes.

* All variables in the table are dichotomous and take the value 1 if the condition is satisfied and 0 otherwise.

** Chi-squared tests of association between Need/No-need and sets of dichotomous variables.

Table 2 presents results from a logistic regression analysis of Need/No-need on the variables listed in Table 1. Gender and hepatitis B knowledge were not statistically significant in Table 1 and omitted from the regressions. In the regressions, the variables measuring “education level” and “occupation” were not statistically significant. “Perceived HBV infection around,” “Perceived risk of being HBV infected” and “Perceived severity of HBV infection” were positively associated with HBV vaccination need, while age level, marital status and health status were negatively associated with HBV vaccination need.

Table 2. Logistic regression. Need for HBV vaccination against explanatory factors (N = 3050).

	Coef.	SE	OR	P
Agelevel	−0.127	0.035	0.881	0.000
Married	−0.302	0.122	0.739	0.013
Education level	−0.110	0.068	0.896	0.105
Occupation	−0.093	0.059	0.911	0.116
Relative health status	−0.227	0.055	0.797	0.000
Perceived HBV infection around	0.190	0.078	1.209	0.015
Perceived risk of HBV infection	0.165	0.064	1.179	0.010
Perceived severity of HBV infection	0.221	0.067	1.248	0.001

Note. Reference group: No need.

The lack of significant association in Table 1 between hepatitis B knowledge and Need/No-need was unexpected and investigated further. Table 3 presents detailed scores (0–26) for hepatitis B knowledge for the 2 groups. More than 25% of the participants had a zero score and the overall median was 11. There was no association between the scores and Need/No-need (p-value=0.745), nor were the distribution of scores significantly different between the 2 groups (Wilcoxon rank sum test, p-value = 0.7263).

Table 3. Hepatitis B knowledge score, by perceived need for vaccination (N = 3050).

	Need			No-need
Score	n	%	Cum.	n	%	Cum.
0	238	25.73	25.73	534	25.13	25.13
5	2	0.22	25.95	3	0.14	25.27
6	5	0.54	26.49	18	0.85	26.12
7	43	4.65	31.14	127	5.98	32.09
8	30	3.24	34.38	65	3.06	35.15
9	47	5.08	39.46	109	5.13	40.28
10	39	4.22	43.68	111	5.22	45.51
11	57	6.16	49.84	113	5.32	50.82
12	58	6.27	56.11	141	6.64	57.46
13	62	6.70	62.81	131	6.16	63.62
14	62	6.70	69.51	145	6.82	70.45
15	62	6.70	76.22	120	5.65	76.09
16	60	6.49	82.70	125	5.88	81.98
17	29	3.14	85.84	85	4.00	85.98
18	27	2.92	88.76	70	3.29	89.27
19	40	4.32	93.08	88	4.14	93.41
20	28	3.03	96.11	66	3.11	96.52
21	13	1.41	97.51	38	1.79	98.31
22	16	1.73	99.24	21	0.99	99.29
23	2	0.22	99.46	8	0.38	99.67
24	2	0.22	99.68	6	0.28	99.95
25	2	0.22	99.89	1	0.05	100.00
26	1	0.11	100.00	0	0	100.00

Note. Chi-square test for association, p = 0.745.

Table 4 presents the joint distribution of the perceived severity of HBV infection and the perceived risk of infection. For both variables a notable proportion gave an uncertain response, 38.3% (1167/3050) for HBV severity and 25.5% (779/3050) for perceived risk of infection. Among those who were not uncertain (N = 1878), 88.6% responded that HBV was serious or very serious and 88.7% responded that it was unlikely or very unlikely for them to be infected. The association between the 2 variables was significant when all observations were considered (p < 0.000), but not when the uncertain respondents were excluded (p = 0.171).

Table 4. Perceived severity of HBV infection by Perceived risk of HBV infection. Frequencies. (N = 3050).

How serious do you think hepatitis B is for adults 18 y	Perceived risk of being infected in next 3 years
or older?	Unlikely or very unlikely	Unknown	Likely or very likely
Not serious	217	1	35
Unknown	350	774	43
Serious or very serious	1448	4	178

Note. Chi-square test for association, p = 0.000 (all observations) and p = 0.171 (omitted “unknown”).

Table 5 presents the joint distribution of the knowledge of HBV transmission score and the perceived risk of being infected. Among those with a zero score, 98.8% (771/780) were uncertain about their risk of being infected in the next 3 y. One-third (66.1%) believed they were unlikely or very unlikely to being infected and 8.4% answered it was likely or very likely that they would be infected. The association between the 2 variables was significant when all observations were considered (p < 0.000) but not when the uncertain respondents were excluded (p = 0. 856).

Table 5. Knowledge on HBV transmission by Perceived risk of HBV infection. Frequencies. (N = 3050).

	Perceived risk of being infected in next 3 years
Knowledge on HBV transmission (score)	Unlikely or very unlikely	Unknown	Likely or very likely
0	9	771	0
1	131	0	17
2	564	5	67
3	378	2	42
4	404	1	58
5	363	0	48
6	140	0	20
7	26	0	4

Note. Chi-square test for association, p = 0.000 (all observations) and p = 0.856 (omitted “unknown”).

Table 6 presents the joint distribution of the knowledge of HBV transmission score and the perceived severity of HBV infection. Among those with a zero score, 99.1% (773/780) were uncertain of whether HBV was serious or not. The mean score was higher (3.60) among those who believed HBV was serious than for those who did not (3.14). The association between the 2 variables was significant, both when all observations were considered (p < 0.000) and when the uncertain respondents were excluded (p = 0.001). Knowledge of HBV transmission was not associated with Need/No-need (p = 0.510, corresponding to Table 1) and was not a significant when included among the other independent variables in the logistic regression.

Table 6. Knowledge on HBV transmission by Perceived severity of HBV infection. Frequencies. (N = 3050).

	How serious do you think hepatitis B is for adults 18 y or older?
Knowledge on HBV transmission (score)	Not serious	Unknown	serious
0	1	773	6
1	23	31	94
2	80	198	358
3	49	53	320
4	48	59	356
5	41	46	324
6	9	7	144
7	2	0	28

Note. Chi-square test for association, p = 0.000 (all observations) and p = 0.001 (omitted “unknown”).

Discussion

This study is part of a larger project that found that vaccination fees and indirect costs (travel cost, time spent) acted as barriers for adult vaccination, and that discrimination against HBV infected individuals is associated with vaccination status and HBV related knowledge.^11,15 The current sample consists of unvaccinated adults with no plans to get vaccinated against HBV. About 70% of the sample stated they did not need the vaccine. For a substantial part of the respondents this might actually be correct, because China is a high prevalence country where many adults are chronic carriers or have acquired natural immunity from past infections.² Due to ethical considerations, we did not obtain data on the participants' infection status and, therefore, the need for HBV vaccination should be regarded as a subjective, and not an objective, measure.

The finding that 46.46% of the No-need group stated they would take the vaccine if it was offered for free may indicate that the respondents' distinction between need and affordability, or willingness to pay, is somewhat unclear. But, in combination with the shares worrying about side-effects or doubting vaccine efficacy, it may also be seen as an expression of trust in public vaccination programs. That is, if a vaccine is offered for free, respondents may interpret that as signal from the health authorities that vaccination is important.¹⁵

The individual characteristics found significant in Table 2 (Agelevel, Married, Relative health status) all had negative coefficients, and were found significantly associated with Need/No-need in Table 1. This indicates that unvaccinated adults that are older, married, or with relatively good health on average thought they did not need the HBV vaccine. Education level and occupation were associated with Need/No-need in Table 1, but their negative coefficients in Table 2 were not significant. Gender was insignificant in Tables 1 and 2 and appears not to be associated with Need/No-need.

The three perception variables were significant in Tables 1 and 2 and had positive coefficients. This indicate that respondents who perceived that HBV prevalence was increasing, that they were likely to be infected in the next 3 years, or that HBV was serious for adults, more often tended to think they needed HBV vaccination.

It was unexpected that hepatitis B knowledge and HBV transmission knowledge were not associated with Need/No-need. However, HBV transmission knowledge was associated with perceived risk of HBV transmission and perceived severity of HBV infections, and both these variable were associated with Need/No-Need. It is possible that the questionnaire items for the perception variables were more understandable for the respondents while those used to construct the knowledge scores were too detailed. In any case, the low scores on HBV transmission routes should be a cause of concern.

We are not aware of studies directly comparable to ours, but our results agree with general conclusions in studies from other contexts. Urban adults in Beijing were found to reject HBV self-paid vaccination due to perceived lack of need, while being more willing if the vaccine was offered for free.²⁶ Outside China, HBV vaccination has also been associated with perceived low risk of contracting HBV,²⁷ and low levels of perceived susceptibility have been reported.²⁸ For other types of vaccines, a Chinese study reported that the majority of surveyed individuals were unwilling to pay the market price forself-paidvaccines, but that willingness to pay was associated with perceived risk and severity of disease.²⁹ The associations between vaccine uptake and perceived risk of infection or severity of disease have been found in other countries.^30-32

Roughly one-third of the no history-no plan respondents reported a need to HBV vaccinate. Age, marital and health status and prevalence, risk and severity of HBV were significant in determining the need to vaccinate. Although knowledge of HBV was not significant, knowledge of HBV transmission and HBV severity and HBV infection were significant. About half of the respondents, whether Need or No-need, would get HBV vaccinated if it was offered for free. One policy implication is that free HBV vaccinations and reimbursements for vaccinating could significantly increase the HBV vaccinate take-up rate for both the Need and No-need rural adults. Another policy suggestion is to promote a tailored public health HBV campaign, especially targeting the No-need subgroup, which would increase vaccination rates by better informing rural adults about the HBV transmission routes, the dangers of HBV infection, the effectiveness of HBV vaccinations and the safety of HBV vaccinations.

Participants and methods

Sampling procedure and study participants

To measure HBV vaccination rates, we selected 7 provinces with notable regional, economic, and epidemiological diversity: Beijing, Hebei, Shandong, Heilongjiang, Hainan, Ningxia and Jiangsu. We stratified the counties by level of economic development (low, medium, high), and then stratified the villages in each county by the distance (short, medium, long) to hepatitis B vaccination sites. In larger villages, we randomly chose households by using household size as sampling weights (probability proportionate to size, PPS), while we included all households in small villages. The participation rate was about 82%, with the main reason for nonparticipation due to household members having moved to industrialized regions.

We surveyed 32,311 participants coming from 42 villages in 13 counties from 7 provinces.¹⁵ We identified a sub-sample of 7879 adult participants who had not received the HBV vaccination and had no plan to HBV vaccinate in the future. We retained 3050 adults (≥18 years) who answered the survey themselves, excluding respondents who were absent at the survey time and for whom other household members gave partial answers on vaccination history.

We employed a self-designed questionnaire, which included questions about individual and household characteristics, individual vaccination history, willingness to get vaccinated in the future, and knowledge about hepatitis B. Well-trained staff conducted the survey after a pilot survey.

Definition and measurement of variables

The subgroup of unvaccinated respondents without plans for future vaccination was asked about the main reason for why they would not get vaccinated. The five options comprised: (1) I don't need the vaccination; (2) I think I need the vaccination, but (2.1) I worry about the safety and side-effect of the vaccine; (2.2) The price is too high; (2.3) I am not confident about the effectiveness of vaccine; (2.4) other reasons. We defined the dependent variable as a categorical variable with 2 outcomes: ‘No-need’ for HBV vaccination and ‘Need’ for HBV vaccination. The No-need group were unwilling to vaccinate because they did not think they needed the vaccination. The Need group believed they needed the vaccination, but they also had no past vaccination history and no plan for future vaccination.

The knowledge of HBV transmission infection risk and infection severity was defined as 2 separate scores (Tables 5 and 6). Respondents were asked to identify transmission routes from a pre-defined set, including 5 true transmission routes (during birth, if the mother is infected; unclean medical or dental equipment; unprotected sex; unhygienic practices of tattoo or piercing; sharing shaving equipment used by an infected person) and 2 false transmission routes (eating with HB patients or HBV carriers; mosquitoes or insects bites). The score was defined as the number of correct answers, so that a higher score indicates higher knowledge.

The hepatitis B knowledge score was defined in a similar manner. In addition to the HBV transmission score it included 3 other aspects: (1) symptoms of hepatitis B (general ill-health; loss of appetite; nausea; vomiting; body aches; fever; dark urine; jaundice, skin becomes yellow; jaundice, whites of the eyes become yellow; (2) treatment of hepatitis B (no treatment is necessary; taking medicine; injections; cupping; religious healing; herbal medicines); (3) prognosis of hepatitis B without treatment (the sick person gradually recovers to full health; the sick person recovers but can still transmit the disease to others; the infection becomes chronic and the person is chronically ill; the person may develop serious liver disease, cirrhosis or liver cancer; the sick person may die after a short time). Again, the score was defined as the number of correct answers. As the range of the score was large (Table 3), we defined the variable hepatitis B knowledge by dividing the hepatitis B knowledge score into 3 groups (0–8, 9–16, 17–26).

Table 1 presents definitions for all independent variables involved in the logistic regression. The participants were asked to describe their health status compared to the health of other people at their age in their village. The options include “much better,” “better,” “about the same,” “worse,” and “much worse.” The participants were also asked: “Was HBV infected population increasing or decreasing in your village?” According to their answers, 3 groups were generated: people who thought HBV infection was decreasing in the village; people who thought the infection rate was unchanging in the village; and people who thought that HBV infection was increasing in the village. The perceived risk of HBV infection variable was assigned values depending on how likely the respondent found it to be infected the next 3 years: 1 if it was unlikely or very unlikely; 2 if unknown; 3 if likely or very likely. The perceived severity of HBV infection variable was assigned values depending on how serious the respondent believed HBV was for adults older than 18 years: 1 if HBV is not serious; 2 do not know; and 3 HBV is serious or very serious.

Statistical analyses

All data were double inputted using Microsoft Access and checked for consistency. Statistical analysis was performed using STATA 12.0. Survey data was analyzed using descriptive statistics. We used logistic regression analyses to assess the associations between each independent variable and perceived need for HBV vaccination. A two-tailed p-value of 0.05 was considered statistically significant.

Ethics

Hepatitis B is a sensitive issue, especially in China. Participation was voluntary and the participants were carefully informed that they could refuse to answer any question. We did not ask about infection status and no biological samples were collected. The project was approved by the Medical Ethics Committee at the Shandong University School of Medicine (Grant No. 201001052).

Disclosure of potential conflicts of interest

No potential conflicts of interest were disclosed.

Funding

This work was supported by the Norwegian Research Council (Project no. 196400/S50).