The Association of Health Literacy with Intention to Vaccinate and Vaccination Status: A Systematic Review

Despite health literacy (HL) being recognized as a driver of health-promoting behavior, its influence on the vaccination decision-making process remains unclear. This study summarized current evidence on the association between HL and both intention to vaccinate and vaccination status. We searched PubMed, Scopus, and Web of Science, retrieving observational studies published until January 2022 that used HL-validated tools to investigate the above associations for any vaccine. Quality was assessed using the Newcastle–Ottawa scale. Twenty-one articles were included; of these, six investigated the intention to vaccinate and the remainder vaccination status. Articles on intention looked at SARS-CoV-2 vaccination using heterogeneous HL tools and were of high/fair quality. Vaccination status, mainly for influenza or pneumococcal vaccines, was explored using various HL tools; the quality was generally high. We found inconsistent results across and within vaccine types, with no clear conclusion for either vaccination intention or status. A weak but positive association was reported between a high HL level and influenza vaccination uptake for individuals aged more than 65 years. HL did not seem to significantly influence behavior towards vaccination. Differences in the methods used might explain these results. Further research is needed to investigate the role of HL in the vaccination decision-making process.


Introduction
Vaccines are one of the most effective and cost-effective tools for the primary prevention of infectious diseases [1]. They provide immunity against various illnesses, preventing death and disability in vaccinated individuals, but also protecting those who cannot be immunized by the attainment of herd immunity [2]. Despite their unquestionable success [3], immunization coverage for several diseases has plateaued or even dropped over the last decade [4]. In Europe, for example, only a few countries have reached the immunization coverage target of 95% against measles [5], and none has achieved an uptake of 90% for the final dose of the human papillomavirus (HPV) vaccine [6]. Key challenges to the attainment of such immunization goals have been unequal access to vaccine services, which particularly affects vulnerable populations, and vaccine hesitancy, a phenomenon that has attracted worldwide interest [7][8][9]. In addition, the COVID-19 pandemic has negatively impacted routine immunization services in most countries, interrupting vaccination demand and supply [10], with consequences yet to be fully quantified [11].
In this context, identifying the factors that might influence vaccination uptake has been the subject of intense research [9,[12][13][14][15]. One possible factor is health literacy (HL), which is a driver of population empowerment that may contribute to a reduction in health inequality [16][17][18]. Health literacy is also inextricably linked to the social and cultural context, which includes education, and it depends on the healthcare system organization, with all these aspects varying across countries [17]. A recent update to the definition of HL has emphasized its role in improving the health and well-being of people, underlining the importance of individual engagement in reaching this goal and acknowledging the fact that organizations need to address it equitably [19,20]. Besides being considered an independent determinant of an individual's health, HL also has a role in mediating the association between socioeconomic status and specific health outcomes, health-related behavior, and access to and use of health services [19,21]. Indeed, it has been demonstrated that people with low HL levels more frequently have poor health outcomes, such as increased hospitalization, higher rates of medication nonadherence, and a lower uptake of preventive interventions [21].
Nevertheless, while HL is a predictor of participation in cancer screening programs [22], a clear relationship between HL and vaccination behavior has not yet emerged [23]. In fact, despite the growing number of studies that have investigated the influence of HL on an individual's behavior towards vaccination [23,24], there are few reports of HL being a determinant [23]. In addition, factors including heterogeneity in the outcomes investigated and the instruments used for assessment, as well as differences in the vaccine type and the target populations, have made it difficult to generalize the results [23,25,26]. Therefore, the aim of this systematic review was to update and summarize findings on the association between HL and vaccination behavior. We considered both intention to vaccinate and vaccination status, aiming to provide a comprehensive picture of the vaccination decisionmaking process and to support the implementation of public health strategies that promote vaccination uptake.

Materials and Methods
This study was performed according to the Cochrane Handbook for Systematic Reviews and the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) statement [27,28]. The review protocol was registered at PROSPERO (identifier CRD42022302724). Because this study did not involve primary data collection, the protocol was not submitted for institutional review board approval and did not require informed consent.

Search Strategy and Study Selection
Three reviewers searched the bibliographic databases PubMed, Web of Science and Scopus using the following search string: ("health" AND "literacy") OR ("health literacy" OR "health literacy [MeSH Terms]") AND ("vaccin*" OR "immuniz*" OR "immunis*" OR "vaccination [MeSH Terms]" OR "vaccines [MeSH Terms]"). The string was adapted to fit the search criteria of each database (Supplementary Table S1). The search was conducted among reports published from database inception to 11 January 2022. No language or date restriction was applied. Duplicate articles were removed, and the title and abstract of all retrieved records were screened. Studies that did not meet the inclusion criteria were excluded. Full texts of potentially relevant articles were examined by three researchers. Disagreements were resolved through discussion and reasons for exclusion were recorded.

Inclusion and Exclusion Criteria
We included studies with the following characteristics: (i) reported in English or Italian, based on co-author language abilities; (ii) cross-sectional, case-control or cohort studies; (iii) investigated HL using a validated tool; (iv) provided raw data, unadjusted or adjusted estimates of the association between HL and vaccination intention and/or status in any population(s). Any statistical analysis was considered eligible. According to Sorensen et al., we adopted the following HL definition: "[people's ability] to make judgements and take decisions in everyday life concerning healthcare, disease prevention and health promotion to maintain or improve their quality of life" [29]. Articles that analyzed HL with non-validated tools, investigated only specific HL (e.g., oral HL), focused on specific subdomains only, or in which data or estimates of the associations of interest were not retrievable were excluded.

Data Collection and Quality Assessment
For each record included, three reviewers independently extracted the following information using a standardized data abstraction form: first author, year of publication, country, study design, main characteristics of the target population (age, ethnicity, recruitment process and number of participants), type of vaccine (e.g., against SARS-CoV-2, measles, etc.), tool used to assess HL, outcome definition and measurement, statistical analysis, main findings, and adjustment factors. Two main outcomes were distinguished: intention to vaccinate and vaccination status. Articles were then grouped according to the type of vaccine and a narrative synthesis was performed for each outcome. Three independent authors performed a quality assessment of the articles included using the Newcastle-Ottawa scale for cohort studies or its adapted version for cross-sectional studies [30] (Supplementary Table S2). Discrepancies were resolved by consensus. Articles were considered of high quality when the total score was ≥7, fair quality if the score was ≥5 and <7, and poor quality if the score was lower than 5.

Results
Overall, 3965 records were identified by database searching (Figure 1). After duplicate removal and screening by title and abstract, 95 articles were selected as eligible for fulltext analysis, from which 74 were excluded with reasons, giving a total of 21 articles ultimately included in the systematic review. Of these, six articles investigated intention to vaccinate [31][32][33][34][35][36], 14 records explored vaccination status [37][38][39][40][41][42][43][44][45][46][47][48][49][50], and one study [51] combined the two outcomes in a single analysis. In this last example [51], the composite outcome resulted from three questions, two of which referred to the vaccinations received in the previous years. For the purposes of this review, therefore, we considered this study to be an investigation of vaccination status.

Intention to Vaccinate
All studies investigating intention to vaccinate were published in 2021 and had a cross-sectional design (Table 1). Two were conducted in the United States [32,36], two in France [34,35], one in Japan [33] and one in Turkey [31]. In all studies but one [35], the authors specified the main characteristics of the target population: in one study, patients with chronic diseases were recruited from health clinics [32], while in three studies individuals were enrolled from educational settings (i.e., students or educators) [31,33,36], and in one study individuals attending homeless shelters were investigated [34]. About half the studies enrolled a large number of participants (i.e., more than 1000) [31,33,35]. All studies explored the intention to vaccinate against SARS-CoV-2. Quality was high in all cases except for one article [35], in which a lack of justification for the sample size and comparability between responders and non-responders were the main deficits (Supplementary Table S2).

Intention to Vaccinate
All studies investigating intention to vaccinate were published in 2021 and had a cross-sectional design (Table 1). Two were conducted in the United States [32,36], two in France [34,35], one in Japan [33] and one in Turkey [31]. In all studies but one [35], the authors specified the main characteristics of the target population: in one study, patients with chronic diseases were recruited from health clinics [32], while in three studies individuals were enrolled from educational settings (i.e., students or educators) [31,33,36], and in one study individuals attending homeless shelters were investigated [34]. About half the studies enrolled a large number of participants (i.e., more than 1000) [31,33,35]. All studies explored the intention to vaccinate against SARS-CoV-2. Quality was high in all cases except for one article [35], in which a lack of justification for the sample size and comparability between responders and non-responders were the main deficits (Supplementary Table S2).

Intention to Vaccinate
HL was assessed using self-reported comprehension items in all but one study [31,[33][34][35][36], which used a tool with reading comprehension and numeracy items (i.e., Newest Vital Sign) [32]. The HL level was then used in the analysis as a mean score in two cases [31,33], was categorized into two classes in two studies [35,36] or into three classes in the remaining study [32] ( Table 3). The intention to be vaccinated was generally explored with one question on attitude and willingness to receive the COVID-19 vaccine [33][34][35][36], and was expressed as a scale in half the studies [31][32][33] or as a categorization in the remaining three articles [34][35][36]. Accordingly, three studies performed multivariable linear regressions or ANOVA [31][32][33], while the other three used logistic regression or its extension [34][35][36]. Results were inconsistent: HL seemed not to influence the intention to be vaccinated in three cases [31,32,36], whereas a significant association was found in two articles [33,34], with low HL levels predicting vaccine hesitancy in one case [34] and higher HL associated with vaccination intention in the other [33]. Lastly, one study [35] recorded a significant association between poor HL and vaccination intention, but only when comparing vaccinehesitant and pro-vaccination individuals. All studies but one [32] conducted multivariable analyses; adjustment factors included were mainly socio-demographic characteristics.

Vaccination Status
Among studies investigating influenza vaccination, the instruments used to measure HL were heterogeneous, but mostly used reading or numeracy comprehension it-ems [37][38][39][42][43][44][48][49][50][51] (Table 4). In just one case, the authors assessed HL using three different tools [49]. HL was categorized into two or three levels in approximately half the studies [37][38][39]41,47,50]. Influenza vaccine uptake was explored using one or more self-reported questions in all studies but one, in which the immunization status was extracted from a registry [49]. The vaccination uptake was evaluated variously in the previous year [39,41,43,44,50], in one or more specific periods [40,49,51], or across the whole life of the individual [37,38], whereas Lorini et al. used a combination of questions on vaccination status and intention to vaccinate [51]. Vaccination uptake was expressed as a binary variable in almost all studies [37][38][39][40][41]43,44,47,49,50]. Results were contrasting: after adjusting mostly for socio-demographic, health status and health habit factors, inadequate but not marginal HL was strongly associated with vaccination refusal in the samples analyzed by Scott et al. [37] and Howard et al. [38]; low HL levels seemed to positively influence vaccination uptake in people aged less than 40 years and negatively influence it among people older than 65 years in one case [43]; high HL levels were significantly associated with vaccine uptake in two studies [44,50], whereas no relationship between HL and immunization status was obtained in five analyses [39][40][41]47,51]. Lastly, the study that used different tools to investigate the outcome found a significant association between high HL levels and vaccination uptake in one case out of three [49], but the analyses were unadjusted.
As for pneumococcal vaccine, all but one study [43] used the short version of the Test of Functional Health Literacy in Adults (S-TOFHLA) as the HL assessment tool, which groups HL into two [39] or three categories [37,38]. The other article used the National Assessment of Adult Literacy questionnaire. All studies investigated vaccination status with a selfreported question, at least once in the entire life [37][38][39] or during the previous year [43]. The outcome was always dichotomized into yes or no, and all articles provided adjusted estimates, using either logistic [37][38][39] or probit regression models [43]. No significant relationship was reported between HL and vaccination status except in one case [37] in which inadequate HL was associated with no vaccination uptake. Adjustment factors comprised mainly socio-demographic variables and health conditions. HPV vaccination status among girls was assessed in one study [48] in which the authors used the Rapid Estimate of Adult Literacy in Medicine to categorize parents' or caregivers' HL levels into three classes. The outcome was calculated as time to completion of three out of four vaccine doses and was divided into four categories in relation to the delay in completion: not delayed (≤12 months), delayed 12-24 months, delayed 24-36 months and delayed >36 months. A multinomial logistic regression analysis found no association between any HL level and the delay in completion of HPV doses for any of the interval times considered. The analysis was adjusted mainly for socio-demographic characteristics of caregivers and the target population.
The current vaccination status of children for hexavalent vaccination was investigated by Pati et al. [45], who used S-TOFHLA to classify the HL of mothers as 'adequate' or 'inadequate or marginal', whereas data on vaccination status were extracted from an immunization registry. After adjustment mainly for the mothers' socio-demographic variables, HL did not seem to predict the decision of mothers to vaccinate their children at three or seven months, according to multivariable logistic regression models.
Likewise, when exploring a similar population using the same HL tool several years later, but investigating the combination of hexavalent vaccine, MMR and pneumococcal vaccine, Pati et al. [42] did not find any relationship between maternal HL levels and the up-to-date immunization status of their children at 24 months, according to a univariable analysis.
The combined DTPa, MMR and HBV vaccination status was considered by Amit Aharon [46] using the Vaccine Health Literacy Scale to assess parents' HL, calculating its level as a continuous variable. With immunization data extracted from a registry, the authors performed a path analysis and found a direct effect between parents' communicative HL and the completion of the childhood vaccination protocol by the age of two, as well as an indirect effect between functional and critical HL and the same vaccination protocol.     Lastly, none of the three HL tools used by Castro-Sanchez et al. [49] detected any difference in the mean HL value of new mothers and pertussis vaccination received during pregnancy, according to a univariable analysis.

Discussion
The COVID-19 pandemic has rekindled interest in the importance of the population's adherence to immunization programs and, consequently, the need to identify factors associated with vaccination uptake [52]. Among such factors, we investigated HL, which is broadly considered a social determinant of health [21,53,54] and a driver of healthy behavior [22,55], but we did not find conclusive evidence of its influence on the vaccination decision-making process. Because vaccination intention does not always reflect real behavior, and predictors might differ between the two aspects [56,57], we distinguished intention to vaccinate from vaccination status. However, in line with the mixed evidence provided by a previous review [23], our findings were largely inconsistent for both outcomes, probably due to the high degree of heterogeneity in the methods used. The issue of the multitude of tools commonly used to quantify HL [58], and accordingly the different domains explored [59], is widely discussed in the literature. In our review, we also found that different tools were used by researchers, and only some measured the individual's capacity to read and understand actual material, minimizing the risk of an inaccurate self-assessment [17]. In this context, the development of a comprehensive instrument for HL evaluation is surely a challenge, but it would definitely enable a more precise estimation of the magnitude of the problem and a better comparison of evidence, even though HL remains strongly connected with cultural and social aspects that make it difficult to isolate this concept [17]. There are similar concerns about the measurement of outcomes, as recently demonstrated in a review that found different rates of vaccination acceptance according to the scale used to quantify COVID-19 vaccination intention [1]. The high degree of variability in the definition of vaccination status, which includes being up-to-date with vaccinations in the last few years, undergoing vaccination at least once over the individual's entire life, or delaying the completion of a vaccination cycle, was also a concern. In addition, the cross-sectional design adopted in most studies complicates the causal interpretation of the findings [23]. For these reasons, to help clarify the role of HL in the vaccination decision-making process, future research on the topic should devise a longitudinal approach with a standardized methodology for the definition and measurement of both exposure and outcomes [23].
Apart from these general methodological considerations, the studies we found on intention to vaccinate focused exclusively on COVID-19 vaccines, probably because researchers wanted to investigate perceptions and intentions regarding newly developed vaccines administered during an out-of-the-ordinary campaign [60]. Our results suggest that factors other than HL are likely to explain people's beliefs and intentions towards COVID-19 vaccines [31,32], such as trust in the government and institutions [32]. For this reason, communication strategies aimed at increasing public confidence in health authorities and helping people understand why recommended measures are useful to them and their community may be the most effective in promoting COVID-19 vaccine acceptance [61]. However, given the emergency context in which these surveys were conducted, more studies are needed to clarify the role of HL in the intention to vaccinate against SARS-CoV-2, but also against other diseases, particularly as, in the latter case, we were unable to find any relevant studies. In addition, to avoid polarizing the discussion around vaccines [60], future studies should differentiate between those who are hesitant and those who are openly against vaccination, as the determinants of the intention to vaccinate may be different in these two subgroups [62][63][64].
Besides the individual determinants that play a role in vaccination intention [65], a few factors may be critical in the actual administration of the injection [66], such as the availability and proximity of vaccination centers [66], the ease with which an appointment can be made [12], or the funding/reimbursement scheme [67]. In our review, the articles that investigated vaccination status mostly explored the determinants of influenza vaccination, probably because it is broadly recommended for the general population, there is an annual immunization program that struggles to reaches the desired coverage threshold, and there are huge variations in the uptake rates according to age and ethnicity [68]. Notably, the fact that a weak but positive association between HL and influenza vaccination uptake was mostly found in individuals aged more than 65 years [37,38,43,44,50] may be explained by the increased vulnerability of this age group to severe influenza outcomes, a factor often mentioned in the routine promotion campaigns that may encourage literate individuals to adhere to the recommended vaccination program [69]. In addition, the annual publicity for influenza vaccination delivery programs may promote vaccination uptake, which is different to what happens for pneumococcal vaccination, where there are less-widespread campaigns and the population's perceived risk is particularly low [67]. As for the few studies focusing on pediatric and adolescent vaccinations, since we did not find consistent association between vaccination status and parents' HL [42,45,46,48], more studies should be undertaken to investigate what influences vaccination uptake, especially considering that vaccine hesitancy in parents has contributed to the recent increase in vaccine-preventable disease outbreaks registered worldwide over recent years [9,70,71]. Furthermore, given that poor communication with parents was likely responsible for the association between vaccination rejection and a high level of education [49], increased attention should be given to communication strategies targeted to this particular group [72]. Specifically, tailored instruments and informative content that takes into account opinions, feelings and gaps in knowledge of the different vaccinations should be devised [73].
This study has some strengths and limitations. Firstly, we included observational studies that provided a general assessment of HL or that investigated all aspects of HL, excluding articles that analyzed only specific sub-domains. Nevertheless, given that we included articles that measured HL through validated and widely implemented tools, the resulting HL estimates can be considered reliable in relation to the multifaceted nature of the concept. Secondly, since our focus was HL generally, we excluded articles that investigated specific HL (e.g., cancer literacy, oral literacy). The other limitations are mostly related to the primary studies included in this review. Heterogeneity in the coding and measurement of HL and outcomes was found, largely limiting the opportunity to provide a quantitative synthesis. In addition, since our results are mostly based on self-reported outcomes, social desirability bias could affect the accuracy of our conclusions. Furthermore, since most studies were from the United States, and several of them analyzed specific subgroups, further research should be conducted both at the regional and national level to improve the generalizability of the findings. Nevertheless, to the best of our knowledge, this is the first review to perform an up-to-date systematic collection of evidence on the topic, expanding the findings provided in a previous review [23]. As a result, we were able to include information on COVID-19 vaccination. In addition, we were able to synthesize evidence on two different aspects of the vaccination decision-making process, namely, intention to vaccinate and vaccination status.

Conclusions
This review summarizes the current evidence on HL and intention to vaccinate and vaccination status. Despite some weak but positive results for influenza vaccination uptake in individuals aged more than 65 years, the relationship between HL and vaccination behavior remains scarcely supported by evidence. Differences in the methods used may explain the inconsistencies we found. Further research using a standardized approach is needed to clarify the role of HL in the vaccination decision-making process.
Supplementary Materials: The following supporting information can be downloaded at: https: //www.mdpi.com/article/10.3390/vaccines10111832/s1, Table S1: Search strategies used in the systematic review; Table S2: Quality assessment of the articles included in the systematic review according to the Newcastle-Ottawa scale for cohort and case-control studies and its adapted version for cross-sectional studies.

Conflicts of Interest:
The authors declare no conflict of interest.