Health Reports
Determinants of non-vaccination against seasonal influenza

View the most recent version.

Archived Content

Information identified as archived is provided for reference, research or recordkeeping purposes. It is not subject to the Government of Canada Web Standards and has not been altered or updated since it was archived. Please "contact us" to request a format other than those available.

Archived

This page has been archived on the Web.

by Maxime Roy, Lindsey Sherrard, Ève Dubé and Nicolas L. Gilbert

Seasonal influenza is a contagious disease that affects millions of Canadians each year.^{Note 1} It is an important cause of morbidity and mortality.^{Note 2} It is estimated that influenza is responsible for an average of 12,200 hospitalizations and 3,500 deaths in Canada every year.^{Note 3Note 4}

Despite its sometimes limited effectiveness, vaccination remains the best way to prevent influenza.^{Note 5} The National Advisory Committee on Immunization (NACI) recommends that all individuals aged six months and older get the seasonal influenza vaccine every year, and especially people at high risk for influenza-related complications or hospitalization, such as people with a chronic medical condition (CMC) and people aged 65 years and older.^{Note 6}

All Canadian provinces and territories run annual influenza vaccination campaigns and various activities to promote vaccination. Moreover, influenza vaccines are free for people at risk of complications or hospitalization, and those who could transmit the virus to someone at risk. In some provinces and territories, influenza vaccines are also free for adults aged 18 to 64 years with no CMC.^{Note 7} Yet in spite of these measures, vaccine coverage for these specific groups remains well below the national target of 80% for adults aged 18 to 64 years with a CMC, adults aged 65 years and older and health care professionals.^{Note 6Note 8} This target was adopted by the federal government and approved by the provinces and territories in 2017.^{Note 9} A number of studies have dealt with the relationship between a particular factor (e.g., ethnic origin or having a chronic condition) and influenza vaccination while controlling for several confounding factors.^{Note 10Note 11Note 12}

This study aims to identify the health and sociodemographic factors associated with non-vaccination among those most cited in the literature as being potentially associated with vaccination or being of public health interest.^{Note 8Note 10Note 11Note 12Note 13Note 14} Studying non-vaccination supports identifying groups where coverage is low, to better inform health promotion activities.^{Note 12} Understanding what contributes to the decision to remain non-vaccinated provides insight on attitudes and concerns, and allows for the exploration of possible solutions to guide vaccination promotion initiatives.^{Note 15} The purpose of the study is therefore to identify the factors associated with non-vaccination and the reasons why Canadian adults did not get the influenza vaccine in 2013/2014.

Data and methods

Population of study

Data from the Canadian Community Health Survey (CCHS) were used. The CCHS is a cross-sectional study conducted nationwide by Statistics Canada every year and consists of computer-assisted interviews on a number of health-related topics.^{Note 16} The survey uses a complex sample design to obtain a representative sample of the Canadian population. Béland (2002) described the CCHS methodology in detail.^{Note 17} The following groups are excluded from the survey: persons living on First Nation Reserves and on Crown lands, full-time members of the Canadian Armed Forces, institutional residents (such as long-term care facilities and prisons), and residents of two Quebec health regions (Nunavik and Terres-Cries-de-la-Baie-James). These groups combined represent less than 3% of the target population.^{Note 16}

The analysis uses the most recent data available, which are from the 2013/2014 cycle of the CCHS. The response rate was 66.2%, and 94.7% of respondents gave Statistics Canada their consent to share the data collected about them with partner organizations, including the Public Health Agency of Canada.^{Note 16} Therefore, the analysis file comprises 111,790 respondents aged 18 years and older.

Variables

The dependent variable was self-reported influenza vaccination status within the past 12 months. This was determined by the responses to the following questions: “Have you ever had a seasonal flu shot?” and “When did you have your last seasonal flu shot?” Respondents who answered “Less than 1 year ago” were considered vaccinated.

The independent variables were age, sex, province of residence, level of education, total household income, country of birth, Aboriginal identity, mother tongue, place of residence, having or not having a family doctor, having or not having a CMC, and self-perceived health. These variables are most often cited in the literature as being associated with vaccination or being of public health interest.^{Note 8Note 10Note 11Note 12Note 13Note 14}

The analyses were done on subpopulations considered at risk of complications, including adults aged 18 to 64 years with a CMC and adults aged 65 years and older. Adults aged 18 to 64 years with no CMC were also included, even though they are not considered at high risk of complications. Respondents who reported a diagnosis of one of the following conditions were categorized as having a CMC: asthma, chronic obstructive pulmonary disease (emphysema or chronic bronchitis), diabetes, heart disease, effects of a stroke and cancer.^{Note 6}

Respondents were able to select multiple reasons for non-vaccination from a pre-established list. These reasons were divided into three major categories: reasons related to the decision to refuse vaccination, reasons related to barriers to access, and other reasons. “Cost,” “transportation problem” and “language problem” made up the subcategory “other barriers to access”. All responses not reflected in the survey questionnaire’s pre-established list made up the “other” category.

Statistical analyses

The associations between the various factors and non-vaccination were measured with unadjusted and adjusted odds ratios (ORs) estimated using simple and multiple logistic regressions. Factors with a p value of less than 0.1 in the simple logistic regression were included in the multiple models and were retained as long as their p remained below 0.1. Additionally, the province or territory with the lowest proportion of non-vaccinated persons was considered the reference category in the logistic regression. All proportions and ORs calculated using the final sample (therefore excluding the response rate) were weighted. Weighting took into account age, sex and health region of residence. The absolute numbers presented are unweighted. Variances, 95% confidence intervals (CIs), and coefficients of variation were estimated using the bootstrap method to account for the complex sample design.^{Note 16} SAS Enterprise Guide 5.1 was used to perform the statistical analyses.

Results

A total of 111,790 adults aged 18 years and older responded to the 2013/2014 CCHS, among which 2.8% (n=3090) were excluded for non-response to at least one of the questions used to create the dependent variable. As such, 108,700 respondents were included in the analyses (Table 1). Of the three groups studied, the proportion of non-vaccinated persons was lowest among adults aged 65 years and older (n = 33,664), or 36.2% (95% CI: 35.2 to 37.1). In this group, the lowest proportion was observed among adults aged 65 years and older with a CMC, or 28.0% (95% CI: 26.8 to 29.2) (Table 2). Among adults aged 18 to 64 years with a CMC (n = 14,036) for whom vaccination was also recommended, the proportion of non-vaccinated persons was 62.2% (95% CI: 60.8 to 63.7) (Table 3). Lastly, the proportion of non-vaccinated persons was 77.8% (95% CI: 77.2 to 78.3) among adults aged 18 to 64 years with no CMC (n = 61,002) (Table 4).

Determinants of non-vaccination

In all the groups studied, being young, having a lower level of education (high school or less and postsecondary below the bachelor’s level), and not having a family doctor were factors significantly and independently associated with non-vaccination. Also, the proportion of non-vaccinated persons varied significantly among the provinces and territories for all groups (Tables 2 to 4).

Among adults aged 65 years and older, factors independently associated with non-vaccination also included having a low household income ($0 to $29,999 compared to $90,000 or more), being born outside Canada, having a mother tongue other than English (compared to English), living in a rural area, not having a CMC, and perceiving one’s health as excellent (compared to poor). The factor most strongly associated with non-vaccination in this group was not having a family doctor (OR 3.57; 95% CI: 3.01 to 4.24) (Table 2).

Among adults aged 18 to 64 years with a CMC, the factors independently associated with non-vaccination also included being a man, having a mother tongue other than English or French (compared to English), and perceiving one’s health as excellent, very good or good (compared to poor) (Table 3).

Among adults aged 18 to 64 years with no CMC, being a man, having a household income of less than $90,000 (compared to $90,000 or more), and living in a rural area were also independently associated with non-vaccination (Table 4).

Reasons for non-vaccination

The reasons most commonly cited by respondents in each group studied were the following: “respondent didn’t think it was necessary,” “have not gotten around to it,” and “bad reaction to previous shot.” Each of these reasons was considered a personal decision to refuse vaccination. Of the reasons related to access, unavailability of the vaccine when required and not knowing where to get it were the two most common. Less than 5% of respondents mentioned reasons related to access (Table 5).

Discussion

Vaccine coverage among high-risk individuals was much lower than the national target of 80%. In other words, the proportion of non-vaccinated persons was greater than 20%.^{Note 9} The proportions of non-vaccinated adults aged 18 to 64 years with a CMC (62.2%) and with no CMC (77.8%) were higher than the proportions posted in the United States for the same year (53.7% and 66.1%, respectively).^{Note 18} In the United States, organizations that provide services to people with certain chronic illnesses, such as the Diabetes Quality Improvement Project, have in recent years been working proactively to increase the number of people who receive the influenza vaccine.^{Note 19} Among adults aged 65 years and older, the proportion observed in the United States (35.0%) was quite similar to the one in Canada (36.2%).^{Note 18} However, these proportions were higher than in the United Kingdom, where the proportion of non-vaccinated persons among adults aged 65 years and older for the same season varied from 23.1 % to 31.7%, depending on the country.^{Note 20}.

Determinants of non-vaccination

The results show that non-vaccination against influenza is associated with certain sociodemographic factors. Firstly, non-vaccination was more common among the youngest people across the three groups studied, similar to the results of other studies.^{Note 8Note 13Note 21Note 22} Advanced age is one of the only factors consistently associated with vaccination in general, and vaccination against influenza in particular.^{Note 14Note 21Note 22Note 23} The fact that the proportion of non-vaccinated persons is lowest among adults aged 65 years and older suggests that vaccination programs are reaching this at-risk group.^{Note 23} Perceived vulnerability could also be playing a role in the decision to get the influenza vaccine. Younger people may not feel that they are as much at risk as older people, and therefore fewer get vaccinated.^{Note 24}

A person’s sex was also associated with non-vaccination. Among adults aged 18 to 64 years with or without a CMC, fewer men than women were vaccinated; however, this difference was not observed in adults aged 65 years and older. These results are consistent with what is generally reported in the literature, by other studies that also compared men and women in the same age group.^{Note 19Note 21Note 22Note 23Note 24} However, other studies conducted in France showed that more men got vaccinated, or that there was no difference between the sexes.^{Note 25Note 26} The difference between men and women could be explained by the fact that women use preventive health services more than men, and as a result, they are more frequently in contact with health care professionals who may recommend that they get vaccinated.^{Note 27Note 28}

Differences between provinces and territories were observed for the three groups studied. Not all provinces and territories offer free vaccination for adults aged 18 to 64 years without CMC, but these differences among the programs do not seem to completely account for the differences in vaccine coverage.^{Note 21} Moreover, less than 1% of respondents cited the cost of the vaccine as a reason for not getting an influenza vaccine. It is also possible that vaccines not provided free of charge by the government are viewed as less important by a segment of the population, including some health care professionals.^{Note 29}

There may be a correlation between level of education and household income. The results suggest that a level of education below a university degree is an explanatory variable for non-vaccination, similar to results observed in other studies.^{Note 10Note 13} For example, an American study showed that the proportion of vaccinated persons was greater among those who had at least gone to college compared with those who had not. However, this result is not significant for all age groups.^{Note 19} Another study conducted in the United States did not find any differences based on level of education for adults aged 18 to 64 years.^{Note 23} Education and income are two indicators of socioeconomic status (SES). It is possible that those with a lower SES may be generally less inclined to adopt preventive health practices such as vaccination, particularly because of a lower literacy level.^{Note 30Note 31}

Living in a rural area was significantly associated with non-vaccination among adults aged 18 to 64 years with no CMC, and adults aged 65 years and older. The few studies conducted to examine this relationship, in France and in Canada, also showed that living in a rural area was associated with non-vaccination.^{Note 10Note 25} This result therefore suggests that those living in a rural area may have difficulty accessing the vaccine. However, when we analyze the reasons for non-vaccination, we see that less than 5% of respondents in all groups cited problems accessing the influenza vaccine. Future studies should examine this factor to better determine its relationship with non-vaccination.

Being born outside Canada was a factor associated with non-vaccination only among adults aged 65 years and older. This was not seen in the other two groups. An American study showed that non-vaccination was more common among immigrants than non-immigrants if they had been living in their host country for less than 10 years. After 10 years, there were no longer any differences between the vaccination status of immigrants and non-immigrants.^{Note 19} A Canadian study based on CCHS data that looked exclusively at ethnic differences revealed that in all ethnic groups, except those who identify as Black, more people received the influenza vaccine than white Canadians.^{Note 12} This relationship is not very clear however, and, although some studies show that mother tongue and beliefs may have played a greater role in explaining this relationship, the results are not consistent.^{Note 12} Mother tongue was associated with non-vaccination in adults aged 65 years and older, and to some extent among adults aged 18 to 64 years with a CMC. Aboriginal identity was not associated with non-vaccination in any group.

Lastly, having a CMC was associated with a greater probability of vaccination in all groups. Not having a family doctor increased non-vaccination in the three groups, whereas perceiving one’s health as excellent was associated with non-vaccination among adults aged 18 to 64 years with a CMC and adults aged 65 years and older. These results were similar to those of other Canadian studies that examined these three factors.^{Note 10Note 14} An American study revealed that a greater proportion of adults aged 18 to 64 years with a CMC who perceived their health as poor or fair were vaccinated than those who perceived their health as good, very good or excellent.^{Note 13} Self-perceived health and actual health are related, and perceiving one’s health to be poor could lead to more medical visits. Another American study showed that vaccine coverage increased with the number of medical visits.^{Note 19} However, in this study, self-perceived health and the presence of a CMC were factors independently associated with non-vaccination. Moreover, having a health care professional recommend the vaccine (a factor not measured in this study) increases the prevalence of vaccination.^{Note 13Note 19Note 27Note 32}

Reasons for non-vaccination

The Health Belief Model describes four dimensions that can guide the adoption of healthy behaviour: 1) perceived susceptibility of developing the disease, 2) perceived severity of the disease, 3) perceived benefits, and 4) perceived barriers to adopting the behaviour.^{Note 33} In this study, the reasons for non-vaccination were related to different dimensions of this model. Among the three groups studied, the most common reason for not getting the influenza vaccine was that the respondent thought it was not necessary, which could correspond to low perceived susceptibility to, and severity of, influenza and to low perceived benefits of vaccination.^{Note 24} In a European study that also used this model, the vast majority of the reasons cited for non-vaccination were due to low perceived susceptibility.^{Note 34} In a Quebec study, the main reasons cited were associated with low perceived susceptibility to, and severity of, influenza.^{Note 22} The second category of most commonly cited reasons relates to high perceived barriers that Santos et al. (2017) described as emotional or cognitive, namely fear and reactions to previous vaccines.^{Note 34} According to them, emotional and cognitive barriers are the dimension that best predicts the adoption of healthy behaviour.^{Note 34} In this study, “fear” (4% to 5%) and “reaction to previous shot” (6% to 12%) were among the four most commonly cited reasons, along with “vaccine not necessary” (66% to 74%) and “have not gotten around to it” (11% to 17%).

Strengths and limitations

The main strength of this study is the quality of the CCHS. First, using data from the 2013/2014 cycle makes it possible to have a large sample size. Second, probability sampling yields results that are representative of the Canadian population, except the few excluded groups. Specifically, 36% of CCHS respondents had a household income over $90,000 and 25% had at least a university degree. These proportions are similar to those observed during the 2011 National Household Survey, where 37% of the Canadian population reported an income over $80,000 and 26% had at least a university degree.^{Note 36Note 37} Finally, the survey had a high participation rate (66.2%), which reduces non-response bias.

Nevertheless, this study contains certain shortcomings. Vaccine status is self-reported and may therefore be subject to recall bias. However, the validity of this information has been proven in the past.^{Note 38Note 39} Moreover, the survey does not include institutionalized populations, such as seniors in long-term care facilities, which could result in overestimation of the proportion of non-vaccinated persons among adults aged 65 years and older. Other vulnerable populations, such as persons living on Aboriginal reserves, are also excluded from the survey. In addition, the CCHS did not include questions on certain chronic diseases that, according to NACI, may increase the risk of influenza-related complications or hospitalization. This may have influenced the proportion of non-vaccinated persons among adults aged 18 to 64 years with or without a CMC. Lastly, since the vaccine status of respondents is known for only a single season, it is impossible to distinguish respondents who sometimes get vaccinated from those who never get vaccinated.

Conclusion

Vaccine coverage for influenza in Canada remains below the 80% target for groups at risk of complications. This study identified several factors associated with non-vaccination against influenza, including a lower level of education and not having a family doctor, two factors that increase social inequalities in health. A better understanding of the determinants of non-vaccination can guide promotion activities to better reach the most vulnerable populations. Future efforts should particularly target the population with a CMC and seek to educate it about the severity of influenza and the increased risk of complications on account of their condition, given that most non-vaccinated persons in this group perceived the vaccine as not necessary.

Acknowledgements

The authors would like to thank Heather Gilmour, Lyne Cantin and Jennifer Pennock for reviewing a preliminary version of this article.