The authors have declared that no competing interests exist.
Conceived and designed the experiments: YCF JHC. Performed the experiments: YCF. Analyzed the data: TCC DWW. Contributed reagents/materials/analysis tools: YCF JHC. Wrote the paper: TCC YCF DWW JHC.
The paper examines the factors associated with both receiving pandemic (H1N1) 2009 vaccines and individuals’ intentions to get the next seasonal influenza vaccine in Taiwan.
We conducted a representative nationwide survey with in-person household interviews during April–July 2010. Multivariate logistic regression incorporated socio-demographic background, household characteristics, health status, behaviors, and perceptions of influenza and vaccination.
We completed interviews with 1,954 respondents. Among those, 548 (28.0%) received the pandemic (H1N1) 2009 vaccination, and 469 (24.0%) intended to get the next seasonal influenza vaccine. Receipt of the H1N1 vaccine was more prevalent among schoolchildren, the elderly, those who had contact with more people in their daily lives, and those who had received influenza vaccinations in previous years. In comparison, the intention to receive the next seasonal influenza vaccine tended to be stronger among children, the elderly, and those who reported less healthy status or lived with children, who received a seasonal influenza vaccination before, and who worried more about a possible new pandemic.
Children, the elderly, and those who had gotten seasonal flu shots before in Taiwan were more likely to both receive a pandemic H1N1 vaccination and intend to receive a seasonal influenza vaccine.
Epidemic and pandemic influenza infections cause tremendous social impacts in addition to generating serious threats to the health and lives of the global population
Due to the limited budget for purchasing influenza vaccines and for other reasons, vaccination coverage rates vary widely across countries. According to surveys in 11 European countries during the 2007/08 influenza season
In Taiwan, government-funded influenza vaccinations started in 1998, when the coverage rate among the elderly was only 9.9%
High-risk subpopulations have a chance to receive vaccinations free of charge beginning in October of each year. Because such policies create significant incentives to receive vaccines, we summarize the seasonal and pandemic vaccination plans during the 2009/2010 influenza season (
Pandemic (H1N1) 2009 vaccine | Seasonal influenza vaccine | ||
Priority groups | Starting Date | Priority groups | Starting Date |
Residents affected by Typhoon Morakot | Nov. 1 | Adults ages ≥65 years of age | Oct. 1 |
Healthcare and public health workers | Nov. 2 | Patients and workers at long-term care centers | Oct. 1 |
Infants ages 6–11 months | Nov. 9 | Grade 1–4 schoolchildren with school-based vaccination | Oct. 1 |
Pregnant women | Nov. 16 | Children ages 6 months - 6 years | Oct. 1 |
Children ages 1–6 years | Nov. 16 | People with severe illnesses/injuries | Oct. 1 |
People with severe illnesses/injuries | Nov. 16 | Healthcare and public health workers | Oct. 1 |
Elementary school students with school-based vaccination | Nov. 16 | Poultry and livestock industry workers | Oct. 1 |
Middle school students with school-based vaccination | Nov. 23 | People engaged in animal disease control | Oct. 1 |
High school students with school-based vaccination | Nov. 30 | All others | Dec. 1 |
Adults ages 19–24 years | Dec. 1 | ||
People with specific chronic diseases | Dec. 7 | ||
All others | Dec. 12 |
The seasonal influenza vaccination was launched earlier, on October 1, 2009, and the pandemic (H1N1) 2009 vaccination started on November 1, 2009, followed by a few waves of vaccinations for different priority groups. During the 2009/10 influenza season, 11.9% of the population received at least one dose of the seasonal influenza vaccine, including 28.5% of children ages 6 months to 6 years, 79.6% of schoolchildren in grades 1 to 4, and 37.6% of the elderly, according to the Influenza Vaccine Information System (IVIS)
Despite the advancement of public health and increased budget allocation, the current level of vaccination coverage in Taiwan, as well as in Europe and the United States, was unable to reach the effectiveness of herd immunity
One such reason lies in people’s motivation or intention. Most studies have shown that past experience with seasonal influenza vaccination is an important factor for understanding the intention to be vaccinated during the next influenza season. Like past experiences, the intention to be vaccinated is also likely to lead to actually getting vaccinated against influenza. Such behaviors and intentions, in turn, have been attributed to various socioeconomic and psychosocial factors, including gender, education status, risk perception of influenza, and trust of influenza safety
In line with the well-recognized linkage between past and future vaccination uptakes, we contribute to the study of influenza vaccination in two ways. First, we identify and compare what kinds of socioeconomic, household, and psychosocial factors are associated with both the behavior (H1N1 2009 pandemic) and intention of vaccination uptakes. Second, we incorporate both the experience and perception of a pandemic (H1N1 2009) influenza into modeling the intention to receive the vaccination during the next seasonal influenza.
With a systematic analysis of data collected from a nationwide representative survey in 2010, we provide insights into the cognitive differences between the actual vaccination behavior (such as past seasonal influenza vaccination history and pandemic (H1N1) 2009 influenza vaccination) and the intention to get vaccinated. Such insights should yield policy implications for influenza vaccination strategy and risk management during the next pandemic.
This study was approved by the institutional review board (IRB) of the Academia Sinica for research ethics (IRB#: AS-IRB-HS 02-13020). To minimize potential ethical issues that might arise during data collection, we took several steps to protect our research subjects. First, all field interviewers and research staff signed an agreement on maintaining the confidentiality of respondents’ personal information. Second, when respondents were younger than age 8, we asked a parent or a guardian to answer the questions on the child’s behalf. Third, we required written informed consent prior to the interviews from respondents, including parents or guardians if the respondents were under 18 years old. Fourth, we maintained all individual survey results under anonymity and retained none of the identifiable personal information (including the names listed in the contact diaries). Finally, after data collection, we destroyed the sample list; and after data cleaning, we destroyed the completed paper questionnaires
Following our extensive experiences with nationwide general social surveys in Taiwan, we designed a representative survey that included questions about residents’ knowledge of and experiences with both the pandemic (H1N1) 2009 and seasonal influenza, as well as 24-hour contact diaries among the whole population
We ensured that our successful survey sample would be as representative as possible by adopting two critical survey designs, as detailed in a previous report
We asked our respondents both “Have you received a pandemic (H1N1) 2009 vaccination?” and “Will you receive a seasonal influenza vaccination this coming flu season?” The answer to the pandemic (H1N1) 2009 vaccination was simply either Yes or No. For the seasonal influenza vaccination, we pooled the original answers into a dichotomy, indicating the intention to receive the vaccination (including “definitely will” and “probably will”) or not to receive the vaccination (including “probably will not,” “definitely will not,” and “don’t know”).
We used such socio-demographic variables as gender, age groups, education status, and work status as the adjusted confounders in the models. The classifications of age groups differed in the two analyses because the vaccination policy assigned different priority groups for the pandemic (H1N1) 2009 and regular seasonal influenza in Taiwan. The survey did not track household members’ vaccination history. Nor did it ask these members about their intentions toward vaccination. As an alternative, we examined whether one’s cohabitants included someone who belonged to a high-risk subpopulation or was a medical professional. In the analysis, we examine whether respondents’ behaviors and intentions regarding influenza vaccination vary by the number of contacts, the percentage of bodily contacts, the number of household members, the numbers of children and the elderly, and whether someone in the household worked in the medical industry.
To compare high and low frequencies of the number of household members and contact number, and partly to simplify the analyses, we selected the median values as the thresholds. Variables used for the household included whether the number of household members was equal to or larger than 5 (median = 4), whether there was any household member with a medical background, and whether there was any household member under age 12 or over age 65. Furthermore, we constructed two measures from the contact diaries to indicate contact intensity: whether the number of contacts with people within the past 24 hours was equal to or larger than 10 (median = 9), and the percentage of bodily contacts among all contacts within 24 hours.
We also took into account several other factors that might affect the behaviors of receiving and intentions to receive influenza vaccinations. The self-reported health condition reflected how respondents perceived their health condition in general. We recoded the answering categories, from 1 to 5, into: poor, fair, good, very good, and excellent, so that a higher score always indicates a better health condition. How often one went to public places (from 1 “almost never” to 5 “almost every day”) serves as a proxy indicator of exposure to the influenza virus from human gatherings and environmental contamination. Watching intense political talk shows about the adverse effects of influenza vaccination may affect how one perceives the safety of vaccinations, which in turn may influence individuals’ intentions and subsequent behaviors. Both the perception of the severity of the (H1N1) 2009 pandemic in Taiwan and the level of worry about a new pandemic serve as measurements of risk perception. Respondents’ past experience of influenza vaccinations could also be an important factor: If respondents had received influenza vaccination at least once in the past five years, we coded their experience as 1, and those with no vaccination were coded as 0.
We first used Pearson’s chi-square test to compare the categorical variables’ frequency distributions, and the Wilcoxon rank-sum test to compare the continuous variables’ medians in the two separate analyses, one about the pandemic (H1N1) 2009 vaccination, and the other about the intention to receive a seasonal influenza vaccination. Due to different age limits set for government-funded vaccination plans, we used age groups under age 18 in two ways (ages 0–10 and ages 0–18) when analyzing who received the pandemic H1N1 vaccination and who intended to get the seasonal influenza vaccine. For the pandemic (H1N1) 2009 vaccine, all children under age 18 belonged to the priority group for free vaccination (
We selected the variables that had a p-value less than 0.2 for the stepwise multivariate logistic regression to determine the factors associated with receiving pandemic (H1N1) 2009 vaccines and the intention to get influenza shots in the next season and to estimate their adjusted odds ratios (AOR) with 95% confidence intervals (CI) as well. The selection criterion of p-value used for both the chi-square test and the Wilcoxon rank sum test served to filter out the variables unrelated to the dependent variables
The survey data of this study were collected during April–July 2010, immediately after the (H1N1) 2009 pandemic in Taiwan. We finished the survey with 1,954 cases of individual questionnaires and 1,943 24-hour contact diaries, at a response rate of 51%. There were 548 (28.0%) respondents having received pandemic (H1N1) 2009 vaccines and 469 (24.0%) persons classified as intending to get seasonal influenza vaccines in the next season, respectively.
The socio-demographic variables, such as age group, education, and working status, all turned out to be significant factors in distinguishing between those who received a pandemic (H1N1) 2009 vaccination and those who did not (all p<0.001,
Variables | Did not receive Vaccination | Received Vaccination | |||
Gender | 0.691 | ||||
Male, no.(%) | 712 | (50.6) | 283 | (51.6) | |
Female, no.(%) | 694 | (49.4) | 265 | (48.4) | |
Age groups, no. (%) | <0.001 | ||||
0–18 | 166 | (11.8) | 318 | (58.0) | |
19–35 | 411 | (29.2) | 42 | (7.7) | |
36–50 | 359 | (25.5) | 54 | (9.9) | |
51–64 | 309 | (22.0) | 45 | (8.2) | |
65 & older | 161 | (11.5) | 89 | (16.2) | |
Education, no. (%) | <0.001 | ||||
Elementary | 367 | (26.2) | 257 | (46.9) | |
High school | 548 | (39.1) | 210 | (38.3) | |
College or higher | 488 | (34.8) | 81 | (14.8) | |
Working status, no. (%) | <0.001 | ||||
Pre-school | 90 | (6.4) | 49 | (8.9) | |
Students | 143 | (10.2) | 269 | (49.1) | |
Work | 793 | (56.4) | 123 | (22.5) | |
Unemployed | 379 | (27.0) | 107 | (19.5) | |
Self-reported health status, median (IQR) | 3 | (2–4) | 3 | (2–4) | <0.001 |
Frequency of visiting public place, median (IQR) | 4 | (3–5) | 5 | (4–5) | <0.001 |
Habit of watching political talk shows, no. (%) | 497 | (35.4) | 112 | (20.4) | <0.001 |
Perception of severity of pandemic in 2009, median (IQR) | 3 | (2–3) | 3 | (2–4) | <0.001 |
Level of worry about new pandemic, median (IQR) | 3 | (2–3) | 3 | (2–3) | 0.041 |
Previous vaccination against seasonal influenza, no. (%) | 231 | (16.4) | 320 | (58.4) | <0.001 |
Household, no. (%) | |||||
# members> = 5 | 554 | (39.4) | 257 | (46.9) | 0.003 |
with med. background | 135 | (9.6) | 52 | (9.5) | 0.951 |
someone under age 12 | 449 | (31.9) | 191 | (34.9) | 0.217 |
someone over age 65 | 388 | (27.6) | 169 | (30.8) | 0.154 |
Contact diary | |||||
# people> = 10, no. (%) | 669 | (47.9) | 322 | (59.1) | <0.000 |
% bodily contact, median (IQR) | 0.24 | (0.1–0.2) | 0.38 | (0.2–0.7) | <0.000 |
IQR interquartile range (25th percentile–75th percentile).
In the final model of the multivariate logistic regression (
Variables | Multivariate Analysis | ||
Odds Ratio | 95% CI | ||
Age groups | |||
0–18 | 23.5 | 11.8–46.8 | <0.001 |
19–35 | 1 (Reference) | ||
36–50 | 1.5 | 0.9–2.3 | 0.128 |
51–64 | 1.3 | 0.8–2.2 | 0.331 |
65 & older | 2.2 | 1.2–4.0 | 0.007 |
Working status | |||
Pre-school | 0.2 | 0.1–0.4 | <0.001 |
Students | 1.3 | 0.6–2.8 | 0.457 |
Work | 1.0 | 0.6–1.5 | 0.973 |
Unemployed | 1 (Reference) | ||
Previous vaccination against seasonal influenza | 6.4 | 4.7–8.9 | <0.001 |
Contact diary | |||
# people> = 10 | 1.4 | 1.0–1.8 | 0.043 |
Socio-demographic variables, such as age group, education, and working status, were all significant factors (p<0.001) affecting respondents’ decisions to receive seasonal influenza vaccines (
Variables | Do not intend to get vaccine | Intend to get vaccine | |||
Gender, no.(%) | 0.407 | ||||
Male | 764 | (51.5) | 231 | (49.3) | |
Female | 721 | (48.6) | 238 | (50.8) | |
Age groups, no. (%) | <0.001 | ||||
0–10 | 134 | (9.0) | 108 | (23.0) | |
11–18 | 180 | (12.1) | 62 | (13.2) | |
19–35 | 404 | (27.2) | 49 | (10.5) | |
36–50 | 353 | (23.8) | 60 | (12.8) | |
51–64 | 300 | (20.2) | 54 | (11.5) | |
65 & older | 114 | (7.7) | 136 | (29.0) | |
Education, no. (%) | <0.001 | ||||
Elementary | 366 | (24.7) | 258 | (55.0) | |
High school | 630 | (42.5) | 128 | (27.3) | |
College or higher | 486 | (32.8) | 83 | (17.7) | |
Working status, no. (%) | <0.001 | ||||
Pre-school | 78 | (5.3) | 61 | (13.0) | |
Students | 298 | (20.1) | 114 | (24.3) | |
Work | 766 | (51.6) | 150 | (32.0) | |
Unemployed | 342 | (23.1) | 144 | (30.7) | |
Self-reported health status, median (IQR) | 3 | (2–4) | 2 | (2–4) | 0.055 |
Frequency of visiting public place, median (IQR) | 4 | (3–5) | 4 | (3–5) | 0.964 |
Habit of watching political talk shows, no. (%) | 485 | (32.7) | 124 | (26.4) | 0.011 |
Perception of severity of pandemic in 2009, median (IQR) | 3 | (2–3) | 3 | (2–3) | 0.436 |
Level of worry about new pandemic, median (IQR) | 3 | (2–3) | 3 | (2–3) | 0.005 |
Previous vaccination against seasonal influenza, no. (%) | 194 | (13.1) | 357 | (76.1) | <0.001 |
Household, no. (%) | |||||
# members> = 5 | 593 | (39.9) | 218 | (46.5) | 0.012 |
with med. background | 130 | (8.8) | 57 | (12.2) | 0.029 |
someone under age 12 | 460 | (31.0) | 180 | (38.4) | 0.003 |
someone over age 65 | 395 | (26.6) | 162 | (34.5) | 0.001 |
Contact diary | |||||
# people> = 10, no. (%) | 742 | (50.3) | 249 | (53.3) | 0.251 |
% bodily contact, median (IQR) | 0.25 | (0.1–0.8) | 0.33 | (0.1–0.7) | <0.001 |
IQR interquartile range (25th percentile–75th percentile).
In the final model of the multivariate logistic regression (
Variables | Multivariate Analysis | ||
Odds Ratio | 95% CI | ||
Age groups | |||
0–10 | 2.4 | 1.5–4.0 | 0.001 |
11–18 | 1.4 | 0.8–2.3 | 0.208 |
19–35 | 1 (Reference) | ||
36–50 | 1.3 | 0.8–2.0 | 0.323 |
51–64 | 1.3 | 0.8–2.0 | 0.403 |
65 & older | 3.0 | 1.8–4.9 | <0.001 |
Self-reported health status | 0.9 | 0.8–1.0 | 0.050 |
Level of worry about new pandemic | 1.3 | 1.1–1.5 | 0.003 |
Previous vaccination against seasonal influenza | 16.1 | 11.9–21.5 | <0.001 |
Household | |||
someone under age 12 | 1.4 | 1.0–1.9 | 0.030 |
In this study, we have explored the factors that influenced whether the general public received pandemic (H1N1) 2009 vaccinations and, at the same time, their intentions to get seasonal influenza vaccines in the coming influenza season. Most current surveys have separated the actual vaccination factors and intentions into different surveys
One study on pandemic (H1N1) 2009 vaccination intention in Taiwan
The actual pandemic (H1N1) vaccination rate in the 19–35 age group, however, was only 9.27% (42 out of 453) in our study, which was the lowest vaccination rate among all age groups. The fact that few young adults got pandemic (H1N1) shots might be due to the passing of the pandemic peak (Nov. 22–Nov. 28, 2009) before they were eligible to get immunized
The previous study in Taiwan
Compared with schoolchildren, pre-school children (ages ≦6 years) were found to have a lower pandemic (H1N1) vaccination rate. The possible reasons might be accessibility and concerns about vaccine safety. During the pandemic period, the vaccination campaign was implemented in all school settings (including elementary, middle, and high schools), which made the campaign more efficient and extensive. It was hard to reach pre-school children without parents’ help, however. The other potential concern was the safety of the pandemic influenza vaccines for young children. One review study in Europe found that the vaccine effectiveness was moderate to good, and that the safety of non-adjuvanted trivalent inactivated influenza vaccines was excellent among children
There were some concerns related to the validity of the responses from the children under age 8. In our survey, we instructed the interviewers that “if the targeted respondents were under age 8, all interviews and records must be answered and taken by a parent or a guardian.” So the information obtained under such circumstances refer to that received from the parent. In addition, because Taiwan’s civil law stipulates that children under age 8 lack behavioral competence, the acceptance of vaccination must be decided by a parent or guardian.
We included all age groups in the analysis, because the priority groups for free influenza vaccinations were set mostly by age limits, and young children are often a top priority. To verify that the results were not altered by the inclusion of young children, we conducted a separate analysis on the respondents under age 8. There were 168 respondents in this subsample (i.e., ages 0–7, about 8.6% of the full sample). The results from the analysis of this subsample in terms of both receiving the pandemic (H1N1) 2009 vaccination and intention to get the seasonal influenza vaccine showed that only receipt of the previous seasonal influenza vaccination was a significant explanatory factor (for receiving the pandemic 2009 vaccination, odds ratio = 4.01, 95% CI: 1.8–9.2, p<.001; for intention to take the next seasonal influenza vaccine, odds ratio = 8.75, 95% CI: 3.6–21.5, p<.001). Because such results are identical to those from the full sample, the inclusion of young children under age 8 in the models did not distort the analysis.
From the viewpoint of vaccination policy, it was desirable to examine why some respondents received either the pandemic (H1N1) 2009 vaccine or past seasonal influenza vaccines but did not intend to receive the next seasonal influenza vaccine. To explore the possible reasons, we summarized the differences in the respondents’ characteristics in
People with more contacts were more likely to get pandemic (H1N1) 2009 vaccines. A direct explanation for this phenomenon was not found in previous literature. Indirect evidence from Europe, however, indicated that vaccination target populations, such as the elderly, healthcare workers, and people suffering with chronic illness, would like to get seasonal influenza vaccines because they do not want to infect their friends and family
One interesting finding from this study was the negative effect of the habit of watching political talk shows on either receipt of the pandemic (H1N1) 2009 vaccination or intention to receive a seasonal influenza vaccine. During the pandemic (H1N1) 2009 vaccination in November 2009, some negative discussions on the safety of the vaccine were featured on TV political talk shows every day. That might partially explain why those respondents were less likely to get a pandemic (H1N1) 2009 vaccine or had a lower intention to receive seasonal influenza vaccine. One study in Canada also found that having negative beliefs about the pandemic (H1N1) 2009 vaccine and deciding not to be vaccinated were highly correlated
There were some limitations in this study. The study was initiated right after the 2009/10 flu season. The acceptance of pandemic (H1N1) 2009 vaccinations might not be the same as the acceptance of seasonal influenza vaccine, due to differing perceptions of disease risk. The pandemic (H1N1) 2009 vaccination rate approximated the official coverage rate among the whole population. Therefore, the representativeness of the data could be assured. In Taiwan, most government-funded influenza vaccines were targeted on the priority populations. Therefore, some socio-demographic factors, such as age or working status, were also highly related to the specific population. In future studies, understanding the factors affecting healthy adults’ decisions to get influenza vaccinations at their own expense would be beneficial for health education.
To our knowledge, this was the first study to compare the factors of pandemic (H1N1) 2009 vaccination and intentions to receive seasonal influenza vaccine at the same time. Since children, the elderly, and those with previous vaccination experiences are more likely both to have received pandemic vaccines and to intend to receive seasonal influenza vaccines, the school-based vaccination program and government-funded vaccines for the priority groups play crucial roles for promoting influenza vaccination in Taiwan. Successful vaccination campaigns during annual influenza seasons will be strong support for promoting the acceptance and delivery of novel influenza vaccines during pandemic periods. Perceptions of pandemic or worries about friends or family being infected are related to people’s intention to receive influenza vaccinations. Thus, prompt and clear risk communication about an influenza epidemic or pandemic through mass media can help generate the correct perception of the disease among the public and enhance acceptance of the vaccine.
(DOCX)
(DOCX)
(DOCX)
We thank Szu-Ying Lee for helping with data collection and analyses.