Loss of confidence in vaccines following media reports of infant deaths after hepatitis B vaccination in China

Abstract

Background: China reduced hepatitis B virus (HBV) infection by 90% among children under 5 years old with safe and effective hepatitis B vaccines (HepB). In December 2013, this success was threatened by widespread media reports of infant deaths following HepB administration. Seventeen deaths and one case of anaphylactic shock following HBV vaccination had been reported.

Methods: We conducted a telephone survey to measure parental confidence in HepB in eleven provinces at four points in time; reviewed maternal HBV status and use of HepB for newborns in birth hospitals in eight provinces before and after the event; and monitored coverage with hepatitis B vaccine and other programme vaccines in ten provinces.

Results: HepB from the implicated company was suspended during the investigation, which showed that the deaths were not caused by HepB vaccination. Before the event, 85% respondents regarded domestic vaccines as safe, decreasing to 26.7% during the event. During the height of the crisis, 30% of parents reported being hesitant to vaccinate and 18.4% reported they would refuse HepB. Use of HepB in the monitored provinces decreased by 18.6%, from 53 653 doses the week before the event to 43 688 doses during the week that Biokangtai HepB was suspended. Use of HepB within the first day of life decreased by 10% among infants born to HBsAg-negative mothers, and by 6% among infants born to HBsAg-positive mothers. Vaccine refusal and HepB birth dose rates returned to baseline within 2 months; confidence increased, but remained below baseline.

Conclusions: The HBV vaccine event resulted in the suspension of a safe vaccine, which was associated with a decline of parental confidence, and refusal of vaccination. Suspension of a vaccine can lead to loss of confidence that is difficult to recover. Timely and credible investigation, accompanied by proactive outreach to stakeholders and the media, may help mitigate negative impact of future coincidental adverse events following immunization.

Hepatitis B vaccination, coincidental death, vaccine confidence, communication

Introduction

In China, over 750 000 babies are born annually to hepatitis B virus (HBV) carriers, becoming HBV-exposed. ¹ HBV is transmitted efficiently during childbirth, and approximately 90% of infants infected when born develop chronic HBV infection that can perpetuate a chain of transmission and cause liver cirrhosis and cancer later in life. ² Administration of hepatitis B vaccine (HepB) to the infant within 24 h of birth followed by two additional doses during infancy is a post-exposure prophylaxis strategy that can prevent 95% of these HBV vertical transmissions and provide lifelong protection from HBV. ³ China successfully implemented this strategy, beginning in 1992 by making a recombinant HepB available, but as a vaccine for which parents had to pay out of pocket, and accelerating in 2002 with inclusion of HepB into the Expanded Program on Immunization (EPI) paid for by government. Between 1992 and 2006, the strategy of timely birth dose reduced by 90% the prevalence of chronic HBV infection among children less than 5 years of age—from 9.76% to 0.96%—a major public health accomplishment that was verified by the World Health Organization (WHO) in 2012. ^3,4

China’s accomplishment was threatened by events that took place in December 2013 and January 2014, in which Chinese and international media reported that 17 infants died following administration of HepB made by Biokangtai in Shenzhen, China. ⁵ This was followed by a loss of confidence in and decreased use of vaccine. The Chinese Center for Disease Control and Prevention (China CDC) monitored parental and provider confidence in vaccines and the health impact of this event. We report the event, the responses by Chinese health authorities and temporally-associated changes in confidence and use of HepB.

Methods

Settings

Social media and mainstream media in China reach national audiences; hepatitis B vaccine is used in all parts of China. To monitor temporally-associated changes in confidence and use of vaccines, we selected several provinces based on three factors: (i) willingness of the provincial CDC to participate in the study; (ii) adequate penetration of the Immunization Information Management System throughout the province so that vaccination data could be collected; and (iii) representation of east, middle and western parts of China. Thus, province selection was, in part, a convenience sample because willingness to participate was a prerequisite. Completeness of the immunization information management system implied that 100% of the province’s townships/districts were included, and that all children, locally registered and migrant, were included. Not all provinces were able to participate in all parts of the study. Within provinces, for determination of use of the timely birth dose of hepatitis B vaccine, hospitals were selected at random from among those providing birthing services. Sampled counties and townships/districts were selected randomly from the provinces. Parents were selected out of listings from the Immunization Information Management Systems; providers were selected from EPI clinician lists maintained by CDCs.

Adverse event and vaccine administration reporting systems

The Chinese government established a national adverse events following immunization (AEFI) surveillance system in 2005. Reported adverse events are initially classified by county-level CDCs into one of five categories: vaccine quality, vaccine reactions, programme errors, coincidental illnesses and psychogenic reactions. AEFI that are fatal or that result in disability and clusters of AEFI are required to be reported within 2 h of their occurrence. Each county, prefecture and provincial CDC must organize an expert panel to investigate AEFI and assess causality. ^6–9

Hospital and birthing centre personnel administer the birth dose of HepB and EPI clinics complete the three-dose series, vaccinating at 1 and 6 months of age. All hepatitis B vaccine used in China is monovalent; hepatitis B-containing combination vaccines are not used. Doses administered are recorded in an Immunization Information Management System (IIMS). China CDC obtained IIMS data from Beijing, Guangdong, Shandong, Hubei, Anhui, Guangxi, Zhejiang, Fujian, Henan and Jiangxi provinces from November 2013 to April 2014, to determine vaccine utilization patterns.

Surveys of parent and provider confidence in HepB vaccination

We conducted telephone surveys to assess parental confidence in HepB and other EPI vaccines in Hunan, Guangdong, Hebei, Ningxia, Gansu, Jiangsu, Xinjiang, Henan, Shandong, Anhui and Heilongjiang provinces in late December 2013 (early period), mid January 2014 (interim period), May 2014 (middle period) and August 2014 (late period). One county/district was chosen from each province, and two villages/towns were chosen from each county/district. Telephone contact numbers of guardians of children less than 2 years of age were acquired, and 50 subjects were randomly chosen to participate in the survey. In each selected village and town, we conducted in-person interviews with EPI clinicians about their knowledge of hepatitis B immunization strategy and their confidence in the safety of HepB and other EPI vaccines.

Survey questions are shown in Table 1 . Some of the questions used in the HepB confidence survey were the same as questions used in ongoing evaluations of parental confidence in vaccines in China. In November and early December 2013, 1 month before the HepB event, a total of 2402 guardians of children in six provinces (Beijing, Shanghai, Anhui, Henan, Yunnan and Xinjiang) had been surveyed about their confidence in vaccines. Forty villages had been randomly selected in each province and 10 households with children aged under 5 years were interviewed in each village. Urban and rural communities were included. Results from this pre-event survey were used as a baseline.

Table 1.

Questions and possible responses for the survey of parents about their confidence in vaccines

Question	Possible responses
What relation is the respondent to the child?	Mother; father; grandparent; other
What is your highest educational level?	No schooling; primary school; middle school; high school; college or above
Have you heard about the infant deaths following hepatitis B vaccination in China?	Yes; no
Among those hearing of the event: from what source did you hear of the event?	One selection only: internet; relatives or friends; television; radio; newspaper; doctor; other
Among those hearing of the event: do you believe that the infant deaths were caused by the vaccine?	Yes, the vaccine certainly caused the deaths; I suspect that there is a relation to the vaccine; no, the vaccine did not cause the deaths; I don’t know
At the moment, do you think domestically produced hepatitis B vaccine is safe?	Very safe; safe; neutral; probably not safe; very unsafe; don’t know
Is this event likely to influence your decision to have your child receive his or her next vaccination?	Yes; no; perhaps—hesitant but undecided; don’t know
Will you have your child receive his or her next vaccination on time?	Yes, will go to vaccination on time; no, will not go temporarily; no, will not have my children vaccinated; don’t know
Have you heard of any other events about vaccine safety in China?	Yes; no
For healthcare workers only: what impact do you think the hepatitis B event may have on vaccination?	Huge impact, a persistent decrease in vaccination rates; big impact, temporary decrease in vaccination rates; minor impact, only leading to insufficient supply of hepatitis B vaccine; no impact; don’t know

Hospital record review

Use of HepB in hospital birth units was determined through a review of medical records in Hunan, Guangdong, Hebei, Shandong, Henan, Anhui, Gansu and Sichuan provinces. In each province, one provincial hospital, two prefectural hospitals and six county hospitals that provided childbirth services were selected at random. Medical records of children born live between November 2013 and March 2014 were reviewed to determine whether the baby received HepB within 24 h of birth and also the HBsAg status of their mother.

Results

Description of event

In mid December 2013, Hunan Provincial Economic Television News reported that two infants died after receiving HepB manufactured by Biokangtai. Other news media picked up the story. The China Food and Drug Administration (CFDA) and the National Health and Family Planning Commission (NHFPC, formerly the Ministry of Health) sent experts to investigate. On 19 December, there were reports of two more infant deaths following HepB vaccination. Two days later, CFDA and NHFPC suspended use of Biokangtai HepB and required local authorities to recall the vaccine, replacing it with another manufacturer’s HepB. On 24 December 24, CFDA and NHFPC held the first media briefing of the event, stating that an investigation was ongoing.

CFDA and NHFPC conducted clinical/epidemiological investigation of the cases to determine causality, re-tested vaccine from the implicated lots, reviewed the manufacturer’s production history and conducted a regulatory inspection site review of the manufacturer. The clinical causality assessment was conducted by the legally-required province-level panel of experts and consisted of three parts: (1) review of interviews of the infants’ attending physicians, (ii) review of the infants’ medical records, and (iii) autopsies and reviews of autopsy results.

WHO experts were invited to participate in the manufacturer site visit and to review all evidence from the investigations. During the investigation, additional infant deaths following Biokangtai HepB vaccination were reported in the media, although most of these had happened before mid December. By the end of December, 18 AEFI had been reported: 17 deaths and one case of allergic shock that recovered. China CDC announced publicly that use of HepB decreased by 30% in December 2013, compared with December 2012.

On 3 January 3 2014, CFDA and NHFPEC provided media with preliminary results of the investigation, indicating that the deaths appeared to be coincidental and not caused by the vaccine. WHO experts supported the interim findings in media interviews. On 17 January , CFDA and NHFPEC announced the final results: the vaccine was within quality specifications; the manufacturing practices were consistent with licensure standards; none of the infant deaths was caused by Biokangtai HepB; and the infant with allergic shock may have had a reaction to HepB or vitamin K. The expert panel had been unable to determine whether the allergic shock had been caused by HepB or vitamin K. The causes of death are shown in Table 2 . CFDA and NHFPC issued a statement to resume the use of Biokangtai HepB.

Table 2.

Final classification of the 18 cases reported during the hepatitis B vaccine event in China

Case	Gender	Age	No. doses of HepB	Interval vaccination-death	Diagnosis	Causal categorization	Reported before or after the suspension
1	M	1 month	2	11 h	Severe pneumonia	Coincidental event	Before
2	M	1 month	2		Anaphylaxis	Vaccine product-related reaction	Before
3	M	1 month	2	24 h	Haemorrhagic pneumonia	Coincidental event	Before
4	M	1 month	2	48 h	Severe, acute bronchopneumonia	Coincidental event	Before
5	M	8 months	3	19 h	Infant muggy syndrome	Coincidental event	Before
6	M	1 month	1	16 h	Asphyxia	Coincidental event	After
7	M	0 day	1	1 hour	Neonatal haemorrhagic pneumonia / asphyxia	Coincidental event	After
8	M	1 day	1	52 h	Amniotic fluid aspiration pneumonia	Coincidental event	After
9	M	2 months	2	25 days	Asphyxia	Coincidental event	After
10	M	5 days	1	7 days	Renal failure	Coincidental event	After
11	M	6 months	3	5 days	Severe infant diarrhoea and dehydration	Coincidental event	After
12	F	1 days	1	73 h	Neonatal necrotizing enterocolitis with digestive tract perforation	Coincidental event	After
13	M	10 days	1	8 days	Meconium aspiration pneumonia	Coincidental event	After
14	M	1 day	1	20 h	Severe pneumonia with respiratory failure	Coincidental event	After
15	M	1 month	2	6 days	Infant muggy syndrome	Coincidental event	After
16	F	15 days	1	15 days	Congenital heart disease	Coincidental event	After
17	M	3 days	1	2 days	Neonatal asphyxia	Coincidental event	After
18	M	1 month	2	2 days	Sudden infant death syndrome	Coincidental event	After

M, male; F, female.

Parental surveys

Sample sizes (response rates) were 1138 of 1201 (94.75%) in early-period (20–24 December 2013), 1704 of 1773 (96.11%) in interim-period (10–15 January 2014), 1926 of 2025 (95.16%) in May 2014 and 1501 of 1604 (93.58%) in August 2014. In the early period, 63% of parents indicated that they were aware of the event; in the interim period, 74% of parents indicated awareness; and in the late period, 70.2% indicated awareness. Parents’ main sources of information were the internet (43.1%), television (40.4%) and relatives or friends (10.8%).

Before the event, around 85% of respondents regarded domestic vaccine as safe; during and after the event, respondents’ beliefs in the safety of domestic vaccines varied from 26.7% in the early period to 67.9% in August ( Figure 1 ) among those hearing of the event. Before the event, 5% of guardians were hesitant to allow their children to be vaccinated. During the event, 30.1% parents were hesitant in the early stage, 26.3% during the interim stage, 10.4% were hesitant 5 months after the event and 8.1% were hesitant 8 months after the event. Before the event, 3% of the guardians refused vaccination for their children; and during the event, reported refusal rates varies from 18.4% to 3.6% ( Figure 2 ).

Figure 1

Perception of safety of domestic vaccines among guardians before and after the hepatitis B vaccine event (Percentage of respondents answering ‘very safe’ or ‘safe’ to the question of ‘At the moment, do you think domestically produced hepatitis B vaccine is safe?').

Figure 2

Percentage of vaccine hesitancy and refusal among guardians before and after the hepatitis B vaccine event (Hesitant: guardians answering ‘hesitant’ to the question of ‘Is this event likely to influence your decision to have your child receive his or her next vaccination?’; Refuse: guardians answering ‘no’ to the question of ‘Will you have your child vaccinated at the next vaccination visit?’).

Vaccination

In the monitored provinces, use of HepB decreased by 18.6%, from 53 653 doses the week before the event to 43 688 doses during the week that Biokangtai HepB was suspended. Two months later, use of HepB averaged 52 000 doses per week ( Figure 3 ). Coverage for the third dose of hepatitis B vaccine declined by 10 percentage points, from 95% to 85%.

Figure 3

Trend curve of the weekly vaccination doses for the three doses of hepatitis B vaccination in 10 provinces/municipalities.

A total of 52 850 hospital records were reviewed from 8 provincial, 22 municipal, and 70 county hospitals for births between mid November 2013 and mid March 2014. Before the HepB event (mid November to mid December), HepB vaccination rates of newborns with HBsAg-positive and HBsAg-negative mothers were above 95%. However, during late December to January, use of the birth dose decreased by 6 percentage points for babies born to HBsAg+ women and 10 percentage points to HBsAg- women, increasing back to baseline within 2 months ( Figure 4 ).

Figure 4

Hepatitis B vaccination rates of newborn infants within 24 h of birth in 100 hospitals of eight provinces before and after the event by maternal HBsAg status (Birth dose vaccination and HBsAg positivity are based on hospital medical record review).

Clinician perceptions

We surveyed 349 township- and village-level immunizing clinicians and 233 obstetric vaccinators from the eight provinces during the initial month of the event and again in August 2014. Among respondents, 47%, 36%, 6% and 5% obtained their information about the event from the internet, television, government sources and colleagues or friends, respectively. Clinician perceptions of causality and safety varied by time since the event. In the early stage, 28.9% believed that the deaths were caused by the vaccine whereas 65% believed that domestic vaccines are generally safe. At the middle stage, the percentages changed to 16% believing that the vaccine caused the deaths and 85.6% believing that domestic vaccines are generally safe. In the early stage, belief that the vaccine caused the deaths varied from 37.5% to 6.3% by province. Among clinician respondents, 77.3% felt that the event would influence routine vaccination, with the percentage varying from 70% to 93.8% by province. There were no meaningful differences by clinician type (data not shown).

Discussion

We have described the December 2013 hepatitis B vaccine event in China and its temporal associations. Parental and provider confidence in HepB decreased substantially following the media reports of infant deaths and the recall of the implicated HBV vaccine. Over a quarter of providers initially believed the deaths to have been caused by the vaccine, even though most providers thought that domestic vaccines were generally safe. The decline in parental confidence was striking, especially among those who recalled hearing about the event, and the decrease in confidence was associated with a measurable decrease in use of HepB and other vaccines. Hospitals’ usage of the HepB birth dose decreased by 6 percentage points for babies born to HBsAg+ women and 10 percentage points to babies born to HBsAg- women.

Within 2 months of the health officials’ announcement that the investigation showed the vaccine to have been manufactured, handled and administered properly, and that the deaths were coincidental to administration of HepB and not caused by the vaccine, confidence increased although not back to baseline. Use of vaccine increased back to baseline, and use of the HepB birth dose returned to baseline. Given that over 700 000 babies are born annually to HBV carriers in China, 1 month of a 6%-decrease in use of the birth dose has potential to result in several hundred babies acquiring chronic HBV infection.

Chen and colleagues conducted an Internet-based analysis of media and public reactions toward this hepatitis B event in China. ⁵ Their analysis showed that the media attention arose in waves, one of which corresponded to the suspension of Biokangtai’s vaccine. They concluded that the event raised great concern among the public in China and that suspension of the vaccine can trigger even greater reactions and negative sentiments. Our results are consistent with Chen and colleagues’ results and extend their findings by showing associations between the event and: the health outcome of missing a post-exposure prophylactic vaccination; utilization of vaccine; missed opportunities to immunize; and survey-measured confidence by providers and parents. Although speculative on our part, reports of deaths before the event may have been stimulated by media coverage.

Li and colleagues conducted a mathematical modelling study of a HepB confidence crisis in Vietnam that happened in 2013, to estimate the number of HBV infections and deaths that were caused by the event. ¹⁰ Their methods involved using vaccination coverage measurements as input to a model linking coverage to HBV-related outcomes. They estimated that the Vietnam event resulted in over 90 000 chronic HBV infections that will lead to over 17 000 deaths. Our study used different methods to estimate impact, and so our findings provide a qualitatively different approach to event analysis in comparison. The great difference in magnitude of the Vietnam event compared with the China event is probably not due to differences in the methods, but rather to differences in the duration and decline in HepB birth dose utilization. As Li and colleagues point out, the hepatitis B vaccine events in Vietnam happened repeatedly over years, beginning in 2007 and continuing through 2013. Thus, the difference in magnitude of events in China and Vietnam also reflects a chronic versus acute event.

Why is hepatitis B vaccine likely to be associated with infant deaths? HepB is the only vaccine for which the routine immunization schedule recommends administration of two doses during the first month of life. In China, the official infant mortality rate is 1.2%, and most of these deaths happen in the first month of life. ¹¹ The United Nations estimated that in China in 2012, over 150 000 live-born babies die within one month of their birth. ¹² Because the neonatal period has the highest mortality rate and is a time when two HepB vaccinations need to be given, some infants who are vaccinated will die from unrelated causes following a vaccination. This coincidence in time can give the appearance that a death might be caused by the vaccine, even though that is almost never the case. ^13,14

Although infant immunization is generally accepted by the public, vaccination confidence crises have been observed in several countries. The French Ministry of Health suspended routine HepB vaccination of adolescents in schools in 1998 following concerns that the vaccine may have been associated with multiple sclerosis. ¹⁵ WHO expressed concern that suspension of HepB in France might reduce public confidence in the vaccine in other countries. ¹⁶ Following promotion of a false link of autism to measles, mumps,and rubella vaccine (MMR), coverage of MMR declined in the UK, resulting in outbreaks of measles and necessitating a large catch-up campaign. ^17–21 Since elimination of indigenous measles in the Usa 15 years ago, intentionally unvaccinated children have comprised a majority of measles cases. ²² In 2003, belief in false rumours catalyzed a resurgence of polio, which spread to 15 other countries. ^23,24 During a 2008 nationwide measles–rubella campaign in the Republic of Georgia, safety fears following television coverage of a cluster of adverse events following MR vaccine among schoolchildren led to a dramatic decline in vaccination. ²⁵

We believe our study provides information useful in improving response to vaccine crises. The hepatitis B event was publicized widely and spread rapidly through social media—within a month, over two-thirds of parents reported knowing about the event, primarily through social media or television. Confidence declined sharply immediately following the suspension of Biokangtai vaccine yet, despite the steep decline in confidence, a much smaller decline in actual vaccine use was seen. Among the most at-risk newborn babies—those born to HbsAg+ women—the decline in birth-dose use was modest and limited to less than 2 months’ duration. Thus the crisis, although very large in the media, was managed by the government of China in a way that was associated with limiting the missed post-exposure prophylaxis HepB birth doses to a relatively small number and short duration.

Actions that contributed to a successful response included: supplying an alternative HepB vaccine upon suspension of Biokangtai vaccine; immediately launching an investigation to determine clinical situations of death, quality of the suspended vaccine, and manufacturer regulatory history and practices; inviting WHO participation and review; holding media briefings during and after the investigation; and monitoring confidence and vaccine use to provide information to decision makers.

Limitations

Limitations to our study include that it was an observational study, most similar to a time series following an event. As such, inferences of causality by the event and its management on parental and provider confidence, vaccine use and health impact cannot be stated conclusively. An unplanned strength of the study is that a survey of parental confidence in vaccines had been conducted in the month preceding the HepB event, allowing a comparison of pre- and post-event confidence. The comparison is imperfect because most of the provinces in the pre-event survey were different from the surveys used in this study. However, we believe that the results are roughly comparable since the difference in confidence was very large and fitted into a logical time pattern. Although supplying an alternative hepatitis B vaccine was an important part of the response management, there may have been stock-outs of hepatitis B vaccine. We did not specifically monitor for stock-outs, and none were spontaneously reported. Therefore, although we believe that missed opportunities due to lack of substitute vaccine did not happen, we can offer no data to support our belief other than a lack of spontaneous reports of any stock-outs. Although hospitals were selected at random, they are too few to be considered representative.

Recommendations

Our study showed that confidence in vaccines is not fully restored in China, and even 9 months after the event, many still held the belief that HepB caused infant deaths. Thus, more needs to be done to reassure parents and providers that the vaccines and immunization strategies being used are safe and effective. Development of a comprehensive communications strategic plan to raise and sustain confidence and to manage crises is a logical step supported by this study. It can be difficult for people to understand that deaths following vaccination are seldom caused by vaccination; thus, education of professional journalists, and proactive media outreach about vaccine safety and AEFI, could be part of a strategic plan. An immediate and thorough investigation into the cause of serious AEFI is essential. Timely, official communication that is responsive to questions and concerns of the media and the public can be based on this evidence, as was done for the hepatitis B vaccine event. The timeliness of the investigation and its communication may have played an important role in regaining confidence and use of hepatitis B vaccine. Periodic monitoring of parental and provider confidence and understanding about vaccines and immunization can help identify concerns and misconceptions.

During the event, the most rapid decline in confidence happened immediately after the suspension of Biokangtai vaccine. A key lesson, then, is that a decision to suspend a vaccine will likely have consequences, possibly leading to a lasting drop in confidence in the vaccine and other vaccines. Therefore, when a vaccine suspension is being considered, unintended loss of confidence is a risk that should be included in the decision making.

Funding

This work was supported by the World Health Organization Representative Office of China, contract number CHN-14-EPI-003578. The World Health Organization Representative Office of China supported the protocol development, field investigation and data analysis and article preparation.

Conflict of interest: The authors declare no conflicts of interest.

Acknowledgements

We thank the staff of the provincial, prefectural and county CDCs and township hospitals and village clinics for conducting the adverse event surveillance, telephone surveys measuring parental confidence, hepatitis B vaccination survey for the newborn in hospitals and coverage assessment of hepatitis B vaccine. We thank Professor Ignatius Yu, the Chinese University of Hong Kong, for his comments and recommendations for editing the manuscript.