Generic health literacy measurement instruments for children and adolescents: a systematic review of the literature

Three researchers (OO, MM, PP) developed the search strategy/method and algorithm. Two researchers (OO, EL) independently searched the databases using identical search algorithms for the respective databases and analysed the data. Three researchers (TMB, MM, JB) checked the initial results, and five researchers (OO, TMB, MM, JB, PP) critically discussed the search outcomes.

The search was performed in five databases: Medline via PubMed, the Cumulative Index to Nursing and Allied Health Literature (CINAHL) via EBSCOhost, the American Psychological Association (APA) PsycNET search platform, the Educational Resources Information Center (ERIC), and the German Fachinformationssystem Bildung (FIS).

The five abovementioned databases were initially searched from May – July 2015 using a composite search term that comprised a combination of three terms for papers published between January 1990 and July 2015. To combine search terms Boolean operators (AND/OR) were used. Wildcard characters were used in order to cover all spelling variations of the search terms. The first search term was “health literacy”, the second addressed the target population (“child*”, “adolescen*”, or “youth”), and the third term focused on either one of ten common terms/topics in the context of measurement tools (“measur*”, “test*”, “tool*”, “instrument*”, “questionnaire*”, “assessment*”, “screen*”, “survey*”, “psychometric*”, and “review*”). In the FIS, the search terms “Gesundheitsbildung” and “Gesundheitskompetenz” were used. Additionally, reference lists of included full-text articles were searched manually (OO) for potentially relevant publications. The following algorithm has been used in PubMed, while the search strategy used in PubMed was adapted to the specifications of the other four databases (see Additional files 1 and 2).

((health literacy[Title/Abstract]) AND ((child*[Title/Abstract]) OR (adolescen*[Title/Abstract]) OR (youth[Title/Abstract])) AND ((measur*[Title/Abstract]) OR (test*[Title/Abstract]) OR (tool*[Title/Abstract]) OR (instrument*[Title/Abstract]) OR (questionnaire*[Title/Abstract]) OR (assessment*[Title/Abstract]) OR (screen*[Title/Abstract]) OR (survey*[Title/Abstract]) OR (psychometric*[Title/Abstract]) OR (review*[Title/Abstract])))

For the purposes of this search, the following inclusion criteria were used (see Table 1): (a) time of publication between January 1990 and July 2015 (as the first health literacy tool was introduced in the early 1990s); (b) English or German language; (c) original publication describing the first use of a health literacy instrument; (d) a study population of children and/or adolescents or at least including these; (e) any setting, i.e., healthcare, schools or other; and (f) any country.

Five researchers (OO, EL, TMB, MM, JB) participated in the screening process. Screening took place in two steps: (1) title/abstract screening was performed by two researchers (OO, EL) independently, and (2) full-text analysis of the eligible publications was performed by three researchers (OO, EL, MM). Two researchers (TMB, JB) checked the results independently. To find consensus and validate the findings, expert consultations were performed by approaching the other authors (DB, EB, GSC, KS, LSN, DLZ, DS, UHB, JP, MT, UB, PP). Literature was imported to the reference management software CITAVI 5. The characteristics of the health literacy measurement instruments identified in the relevant publications were analysed and extracted by two researchers independently (OO, EL) based on the objectives of this review. The decision on what data to extract from the articles was based partly on (a) former systematic reviews of health literacy measurement tools [4, 38] and (b) an extensive discussion and consensus among the authors of this review and with further health literacy experts from the HLCA research consortium. If information on the validity and reliability of instruments was documented within the articles, these data were extracted and analysed as well.

Among the identified instruments (see Table 2), ten instruments are novel instruments that were developed especially either for children and/or adolescents [40‐47] or for adult age groups including adolescents 15 years and older [48, 49]. In relation to the teen version of the Rapid Estimate of Adolescent Literacy in Medicine (REALM-teen) [50], the adolescent version of the Test of Functional Health Literacy in Adults (TOFHLAd) [51], and the Newest Vital Sign (NVS) [52], our review identified child and adolescent adaptations of the most widely used fast-screening instruments in clinical and medical-related adult health literacy research. Another adaptation/validation study conducted in Austria [53] applied the population-based health literacy tool HLS-EU-Q47, which was developed for the HLS-EU [10]. This tool, which was originally developed and validated for adults (15 years +), was specifically validated as a long form of the tool and subsequently adapted as a short form for adolescents. Finally, one of the identified instruments is a shorter form of an already existing health literacy measurement tool for adolescent mothers, but analysis of the longer form of the tool is yet not published [54]. REALM-teen was the first health literacy instrument ever to be applied in a child population during a validation study [50]. The most recent study to use this instrument was conducted with the Maternal Health Literacy (MaHeLi) scale in Uganda in 2015. Questionnaire items are fully provided in seven of the studies [40, 43, 47, 49, 52‐54] but are not provided at all in seven other studies [41, 42, 44, 45, 48, 50, 51]. One study provides only some of the items [46].

Six of the health literacy measurement instruments were developed or applied in the United States [40, 41, 47, 50‐52], followed by three from Germany [43, 46, 48] and one each from Greece [42], the United Kingdom (England) [49], Canada [44], China [45], Uganda [54] and Austria [53]. Two articles were published in German [46, 53], and the rest were published in English. While one instrument was used in the general population and not in a specific setting [53], nine studies took place in schools or other educational settings [40‐46, 48, 52], four in healthcare settings [47, 49, 51, 54], and one in a mixed setting (educational and healthcare) [50].

The majority of studies (eleven) provided the exact age of the target group, but this was not specified in the other four studies [41, 44‐46]. In these studies, the type of school was specified; three of these studies were performed in secondary schools [41, 44, 46], and one study was performed in both elementary and middle schools [45]. While all instruments have been used in age groups older than 11 years [40‐54]; five of these instruments have been used with children 10 years old and younger as well [40, 43, 45, 50, 52], among which only three were specifically designed for children of this age: 9 to 13 years [40, 43], primary school-aged children [45].

The analysis showed that seven health literacy tools were based on subjective measurement using self-reporting questionnaires [40, 42, 45, 47, 49, 53, 54]. Another seven articles reported the use of objective measures assessing the actual performance in given tasks [41, 44, 46, 48, 50‐52]. One study adopted a mixed-method approach that combined both test methods [43]. All instruments were self-administered measures, except for two that adopted the use of structured interviews to collect the data [50, 53]. Four questionnaires were based on the pen-and-paper mode [46, 51, 52, 54] and another study used a computer-based questionnaire [40]. However, for eight questionnaires the specific administration mode (pen-and-paper or digital) was not provided by the authors [41‐45, 47‐49]; of these studies, one used two booklets given to the participants indicating that pen-and-paper may have been the mode of administration [44]. Two instruments used a mixture of open-ended and close-ended questions [44, 45], seven used close-ended questions only [40, 43, 47, 49, 52‐54], and five did not provide sufficient information on the specific question types [41, 42, 48, 50, 51].

Only two articles reported the involvement of participants in the development of the questionnaire by conducting qualitative research in order to derive meaning about health literacy from the target population [47, 49]. Massey et al. [47] conducted 12 focus groups with adolescents 13–17 years old and interviewed eight primary care providers who worked with adolescent populations. Chinn et al. [49] conducted expert consultation and focus groups with health and non-health professionals as well as with patients. However, these authors did not report the specific age of the participants. Cognitive pre-testing was conducted by six studies to obtain qualitative feedback by using methods, such as think-aloud techniques or interviews related to question comprehension or feasibility [40, 41, 44, 47‐49].

Nearly all the studies reported the psychometric properties of the instruments (see Table 2). The reported reliability differed across measures, with internal consistency ranging from α = 0.42–0.94. The highest reported rates were α = 0.94 [50], α = 0.92 [44], α = 0.90 [53] and α = 0.83 [47]. The weakest internal consistency was reported for the attitude scale α = 0.57 in the GeKoKids questionnaire [43] and for the critical health literacy scale α = 0.42 in the AAHLS questionnaire [49].

Reporting on the convergent validity of most instruments was insufficient. For those provided, these data were positive and moderate (r = 0.09–0.93). However, concurrent validities were often not tested or not reported. REALM-teen [50] correlated significantly with WRAT (r = 0.83) and SORT-R (r = 0.93). TOFHLAd [51] was strongly correlated with WRAT3 (r = 0.59) and REALM (r = 0.60). HLS-EU-Q-47 moderately correlated with NVS (r = 0.14), as did the short version, HLS-EU-Q16 (r = 0.09) [53]. In terms of concurrent validity, the short version was significantly correlated with the long version (r = 0.82). Test-retest reliability was reported for one instrument and was demonstrated to be strong (r = 0.98) for REALM-Teen [50]. Although the authors of the NVS study claim that the instrument is valid, they further explain that children younger than nine years had difficulties answering the questions [52].

Additionally, Hubbard and Rainey [41] selected items from the HEAP database, which they report to have proven content validity and reliability, as do Massey et al. [47], regarding some of their questionnaire items that were taken from other sources. Schmidt et al. reported the inclusion of some questionnaire items from the German KiGGS survey [55], which have been proven to be valid and reliable [43].

This review identified five instruments that measured health literacy related contextual factors [42, 43, 47, 49, 53]. These asked for stated satisfaction with healthcare professionals, media sources, or in the context of health learning [42, 47], parent-child communication [43], perceived availability of social resources in form of help and support [49], or perceived difficulties in the social or media interaction [53]. The other instruments focused solely on individual skills or did not provide any information related to contextual factors. However, most studies focused on three main context arenas: healthcare, prevention, and health promotion, including health education (see Table 2).

Each instrument used a different, study-specific understanding of health literacy, and two instruments were underpinned with a health literacy definition that was specifically developed for the instrument [40, 53]. In the other 13 studies [41‐52, 54], the researchers referred to different existing definitions but did not make clear whether or not these definitions were underlying their instruments, and of these, six studies [42, 43, 46‐49] referred to the definition provided by Nutbeam [8, 56]. A further five articles [42, 47, 48, 51, 52] quoted the Healthy People 2010 definition [57], the definition of Ratzan and Parker [58] or that of Parker and Ratzan [9], and the definition by Zarcadoolas and Pleasant [59]. One study [54] did not refer to a specific definition. Instead, the authors emphasised two different models as important for their tool, the HLS-EU model [1] and the Health Belief Model (HBM) [60]. Finally, one study [46] referenced the definition developed by Mancuso [61].

The analysis revealed three different conceptual approaches that underpin the instruments: health literacy was based on (i) a functional literacy model in three studies [50‐52], (ii) school health education standards from different countries in five studies [40‐42, 45, 46], and (iii) broader multidimensional models in seven studies [43, 44, 47‐49, 53, 54]. In one of the studies the researchers claimed that their instrument measured functional health literacy by assessing knowledge, whilst the knowledge questions were based on school health education curriculum [46]. Although knowledge is highlighted as a main component of functional health literacy [8], this instrument was nevertheless instead classified under the second category rather the ‘functional literacy’ category.

To ascertain the scope of the measurement approach, a content analysis of the measured components (sometimes referred to as “dimensions”) was performed that could identify a wide array of skills, competencies, abilities, or certain actions covered by the instruments. To closely relate to the original published study and thereby avoid any false interpretation, we decided to label a specific component as if the authors of the original article had named it explicitly. For example, in Brown et al. [40], the authors explicitly state that their instrument measures understanding, accessing, and applying health information as well as beliefs, attitudes, and interests towards health. Based on this, six components could be identified: understand, access, apply, belief, attitude, and interest. In Chisolm and Buchanan [51], the authors explain that their instrument intends to measure reading ability and numeracy skills. Therefore, we extracted two components, reading and numeracy. We applied this coding scheme to all identified articles. In total, the 15 instruments comprised 31 different components (see Table 3 and Fig. 2). The ability to understand health information was identified in six studies [40‐42, 44, 48, 53] and was the most prominent approach to measure health literacy, followed by four studies that asked about the ability to access health information [40, 41, 49, 53]. The health literacy components appraise, attitude, communicate, knowledge, reading, and numeracy were found in each of the three instruments. Many different components, such as problem-solving, self-management, coping, and self-efficacy, were each assessed in only one instrument.

While seven health literacy instruments (nearly 50%) have been used in North America [40, 41, 44, 47, 50‐52], all but one [44] in the USA, six instruments have been developed and used in Europe [42, 43, 46, 48, 49, 53]. However, health literacy measurement tools for children and adolescents remain a marginalised area of research in Asia [45] and Africa [54], with only one instrument found for each region. The authors of this review are aware of instruments used in Asia that had yet not been published in English while preparing this article [62‐64]. However, our search has not identified any generic health literacy instruments for South America or Australia. Given that Australia is one of the pioneering areas of health literacy research and action in childhood and adolescence, primarily in the context education [28, 65], this result is especially surprising. In relation to Africa, while the short form of the MaHeLi scale was adapted to the settings in Uganda, it was originally a European-based instrument [54]. Health literacy research in Africa, a continent with approximately 60 countries with more than one billion inhabitants and serious health threats to children and adolescents [66], has only recently begun and thereby is calling for an ‘African health literacy scheme’ and proportionate solutions [67]. Given the social, cultural, economic, and political differences between regions, it is especially important that research in these regions engages in health literacy with children and adolescents to facilitate better health promotion despite the challenges that these populations face. Development of more culturally sensitive concepts and measurement tools in these regions would also support this approach. In turn, and due to migration to Western countries from within these regions, health promotion and prevention in migrant and refugee child/adolescent populations in Europe and North America in particular could benefit by such approaches when adapting these tools and concepts.

Regarding the settings in which health literacy was measured, it seems that schools and other educational settings are main loci of interest for measuring the health literacy of children and adolescents. This review found that many health literacy tools are based on school health topics or curricula [40‐46, 48]. When researchers aim to derive an understanding of health literacy for children and adolescents, an existing health literacy-related school curriculum could provide the necessary information and would also ensure comparable instruments and models. Shaping of school health topics to match health literacy content, therefore, would foster further development and the use of comparable instruments, at least within countries, which would allow for consistent monitoring and evaluation of health literacy in these settings.

The findings of this review suggest that the age groups from 11 to 18 years are well reflected by existing health literacy tools (see Table 2). This review was able to identify a total of 15 instruments that were used with children within those age groups [40‐54]. However, only five of these tools had been used to measure health literacy of children younger than 11 years old [40, 43, 45, 50, 52]. One instrument was used with children aged 7 to 12 years old within a parent-child study [52]. In this study, the authors reported difficulties for children younger than 9 years old when administering the questionnaire [52]. This finding is supported by another study using this instrument in that age group [68]. One instrument was used in a study specifically conducted in the primary school setting in China without providing the exact ages of the children who participated [45]. However, we assume that the children were most likely between 6 and 10 years old, which is the common age of primary school-aged children in China. The same study also used their instrument with secondary school children, but they did not provide any information on differences in using this tool between age groups. Two instruments were used with secondary school children aged 9 to 13 years [40, 43]. Although it is most common that children 9 and 10 years old usually attend primary schools rather than secondary schools, in those two studies, the children were nevertheless reported to be students at secondary schools. Together with the Chinese instrument [45], these two [40, 43] were also the only instruments that were specifically developed for children younger than 11 years old. The other two instruments were adaptations of the REALM [50] and NVS [52] studies. With only one instrument specifically developed for the use in primary school-aged children [45], our findings emphasise that there might be only one health literacy instrument available for primary school children. Unfortunately, the authors do not provide adequate information that might facilitate a better understanding of their instrument. In conclusion, there is a lack of adequate information on health literacy instruments for primary school children and no instruments available to assess the health literacy of children younger than seven except the Chinese tool [45].

Health literacy is important for young children, but the conceptualisation and measurement of health literacy are different from the approach taken with adults. Existing health literacy concepts specific to adults address how people understand, evaluate, and use health information, which may not be appropriate for children of primary school age who are not making all the health-related decisions that affect them. Not only do such approaches demand more complex skills to process health information, but an individual also needs more experience and autonomy to be able to judge the quality of information or to act based on given, sometimes complex, information. Primary school-aged children are still developing their formative skill sets that are required to process health information, and they will have acquired far less health-related knowledge and experience and have less autonomy than adults or older children [14, 17, 18, 20, 22, 26‐28]. Therefore, health literacy concepts and their assessment in this age group have to be less complex and should focus on more basic aspects of health promotion and education, such as healthy eating, physical activity, or knowledge about their bodies and the environment. Arguably in healthcare contexts, the health literacy of the parents of young children is more likely to be important and have greater impact than is that of the young children themselves, as the parents are those who communicate with professionals and make the influential health-related decisions. These assumptions are also supported by two studies that reported on cognitive difficulties experienced by younger children during the administration of health literacy instruments [52, 68].

Half of the identified tools were subjective measurement tools and asked the participants for an assessment of their self-perceived health skills and half were objective measurement tools and assessed the actual capabilities by performance testing. Only one tool combined both assessment methods [43]. Although self-reported health literacy measurement has been used in many studies and instruments [4, 32], there is much criticism concerning self-reporting as a valid measure of health literacy [33, 44, 49, 69‐72]. It is argued that self-reporting measures are oversimplifying the given complexity of health literacy [49], or, in some cases, assess self-efficacy rather than health literacy [33]. Furthermore, a discrepancy between self-reported health literacy and the actual performance related to health literacy has been highlighted indicating concerns with self-reporting measures in terms of accuracy [33, 44, 49]. Other research suggests that self-reporting methods are just as valid as objective methods [73]. Ethical concerns related to objective measures, however, have highlighted the potential discomfort that participants with low health literacy skills may feel if they are ashamed or embarrassed by their abilities [73, 74]. For future health literacy research with children and adolescents, adopting a combination of both methods might prove a valuable way in order to generate more detailed and profound data. It would also allow for the comparison of differences in health literacy as measured with these two different methods. Applying different research methods may also generate richer results and, therefore, would support the development of better and more problem-centred interventions to address specific health literacy weaknesses.

Including participants in the development of a health literacy questionnaire has proven to be a sound method to improve the quality of the measurement tool [75]. Regarding the development of new and innovative health literacy tools, it is recommended always to include the target population in the development process as early as possible [76]. However, analysis of this review showed that children are mostly excluded from the development of the instrument and/or the conceptual construction that underpins the questionnaire. With only two reporting qualitative research with adolescents and/or adults during the development process of the instrument [47, 49], child and adolescent involvement was poor. In none of the studies were children younger than 13 years old actively involved, except for completing the questionnaires. Accordingly, this review found adult experts setting out to define a concept for children and young people without consulting them to determine their understanding of health literacy and assuming the health-related skills and knowledge and health behaviours and practices that may be important for them in their everyday lives. This conflicts with theories and findings from childhood studies that consider children and adolescents to be active citizens, social agents, and co-constructors of their social worlds [77, 78]. Researchers from this field also highlight that children should be understood as a social minority living in a “childhood” with unequal power relations, that they suffer from an uneven distribution of rights, and that one major weakness in research and practice is that childhood is mainly constructed by adults [28, 77, 78]. Although two studies valued contributions from adolescents as useful sources of qualitative information, there is considerable room for improving child/adolescent participation in the health literacy research process, especially as there are yet no studies including young children in the development of instruments. Therefore, little is known about what health literacy means to children, what knowledge and skills are important to them in order to promote their health, or how, where, when, with whom, and why (or why not) they are interested in developing skills in order to promote their own health. To learn more from children and to learn collaboratively with them, their active involvement and participation in the social and cultural construction of health literacy and its measurement is a specific challenge that would be highly beneficial to overcome for research with children. In not doing so, an important voice, the children’s perception of health and health literacy, is not recognised and given the consideration that it deserves in the field of health literacy research. Although, few studies have performed qualitative health literacy research with children [29, 62], in recent years, especially in health-related disciplines (for example, mental and dental health), qualitative research with children has increased [79‐82]. Researchers argue that the benefits of children’s involvement by using qualitative research methods are obvious and include enhancing child empowerment, producing better knowledge and understanding of children’s views and priorities, and developing better tools and practice measures for more effective action [79, 81]. Additionally, as such, it could be argued that applying qualitative methods in health literacy research with children would have similar effects and thus produce more precise findings and unravel children’s own perspectives and knowledge related to health. Such findings could be used to derive a child-centred understanding of health literacy and elucidate the underlying tenets of children’s health literacy. Furthermore, children’s participation could also support the development of more child-adjusted measurement tools and facilitate the development of tailored health literacy interventions that better match the needs and demands of this population.

The results concerning the validation processes are diverse. Only a few studies provide validation data on the instrument [44, 48‐51, 53], and most instruments have weak or only moderate validity. The focus of reporting in these studies is on statistical population data, results, and sample characteristics, while reporting on methodological data and psychometric properties is scarce and not well described. These results confirm findings from other systematic reviews with different foci [31, 32, 37]. Only one article made test-retest-reliability data available [50]. To support better clarity on the characteristics of tools, the authors should provide more detailed information on the development and validation process, psychometric properties, and assessment characteristics.

There has been much debate in the conceptual and methodological discussion regarding adult populations as to whether health literacy is only associated with individual abilities, or if beyond individual abilities, context also might have an essential impact on the health literacy of individuals [1, 9, 10, 16, 21, 33, 58, 59, 70, 71, 83‐85]. Similarly, the influence of context is also discussed within the scientific literature related to children’s health literacy [14, 20, 22‐28]. It seems that the scientific community generally agrees that context is an important dimension that should be considered appropriately and that health literacy itself is a relational concept that is influenced by more than only individual abilities. In general, when the relatedness of health literacy is being discussed, the context dimension is mainly meant to capture the complexities and demands of the health system and the health literacy of health professionals, particularly in relation to the health literacy of individuals that they interact with in certain health literacy related situations [9, 16, 58, 85]. This also includes information materials being tailored and user-friendly and communication being based on understandable and simple language [85]. In this context, the identified measurement tools, however, did not provide much information because most of the tools focused on measuring individual abilities rather than contextual factors with impact on health literacy. However, when the context dimension is extended to also address contextual factors beyond health system complexities, there are some findings that can be reported. Overall, the analysis could identify five instruments that included questionnaire items addressing contextual factors or situational determinants that might influence health literacy [42, 43, 47, 49, 53]. In relation to health literacy, one instrument assessed the stated satisfaction of adolescents in the context of health learning in schools [42]. Another questionnaire asked adolescents about the confidence that they feel when they try to access health information from personal and media sources [47]. Two instruments included broader context questions in relation to self-perceived difficulties with respect to health literacy and thereby asked about the trustworthiness and meaning of health information and sources, the possible impact of the social and political environment on health, and the interaction with health professionals [49, 53]. Finally, one instrument included questions related to children’s communication habits with their social environment (friends/parents) [43]. Apart from collecting data on individual abilities, asking such questions related to emotions, attitudes, opinions, and interests within the context of specific health literacy situations seems to be a promising approach that could support the derivation of a broader understanding of the interplay between individuals and their health-related environment. In addition, it allows for encountering valuable insights from individuals and how they perceive the influence of context on their health literacy and health practices.

Context can also be understood as the setting in which individuals use their health literacy or interact with others, such as in healthcare and medicine, disease prevention, health promotion, everyday life, and educational settings. In this context, our review has found several health literacy instruments that are related to different contexts and their inherent characteristics. While the health literacy instruments used within the educational setting had a dominant focus on health promotion resources and prevention [40‐46], studies conducted in hospitals or medical centres addressed various aspects of the healthcare environment by asking questions, i.e., on treatment, self-management, and interaction with health professionals [47, 49‐52, 54]. Two instruments were found that addressed mixed contexts, including healthcare, prevention, and health promotion and education [48, 53].

However, to provide meaningful data on the impact of context on individual health literacy, future health literacy measurement tools for children and adolescents should include more questions related to contextual factors. In addition, tailored questionnaires could also be administered simultaneously in dependent populations within the same study, for example, involving teachers and students, patients and health professionals, or parents and children, including questions on the various inherent intricacies of complex contexts as discussed within the conceptual health literacy literature. This would ensure the capture of both sides of the health literacy concept, the individual and its related context.

Health literacy measurement tools for both adults and children/adolescents have been criticised for different reasons [33, 38, 86]. First and foremost, the lack of a consistent general understanding of health literacy is highlighted as a major weakness within the field [1, 33, 76, 85] that affects the development of comparable methods to accurately measure the concept [33, 85]. This review supports this finding, as each of the identified instruments introduced an individual and/or study-related understanding of health literacy. This makes it difficult to compare the instruments or the results generated by these instruments. Even instruments that apply similar conceptual approaches vary considerably in how they transfer the underlying concepts into questionnaire items at the operational level. Three instruments, for example, measure functional literacy [50‐52], which can be described as a unidimensional approach to health literacy and sometimes is referred to as medical health literacy [1, 83, 84, 86]. Even though REALM-teen [50], TOFHLAd [51], and NVS [52] are tools that measure functional literacy, they are however testing different components of the literacy construct, such as reading, comprehension, and numeracy. Another tool addressing functional ‘health literacy’ is the “health quiz” measurement tool [46]. However, it is based on the functional concept of health literacy as proposed by Nutbeam [8], measuring knowledge rather than literacy, as in the previously mentioned tools [50‐52].

Compared to the narrow ‘functional literacy’ approach to health literacy, all other instruments (n = 12) are informed by a multidimensional, multifactorial, or multifaceted and thus broad understanding of health literacy. Still, each understanding is more or less unique, and therefore, comparisons across these health literacy assessment instruments and associated study results are difficult to undertake. Additionally, as the instruments do not cover all their proposed dimensions and components at the operational level, there seem to be some uncertainties regarding how to transform all aspects of the theoretical model into a testable concept within a measurement tool. It is also not clear why a certain model or definition is provided at the conceptual level when at the operational level, something else is indeed being measured. However, when school health education (or other educational contents and contexts) is the foundation of the health literacy framework, it seems that there is more precise clarification as to why, how, and for what purpose a model was designed and operationalised [40, 42‐45]. If an instrument has been based on a clear conceptual model, it is much more feasible to clearly understand what is assessed by the construct [43, 47‐49, 53, 54]. However, in some studies, the operational definition covers only select aspects of the conceptual definition [40, 41, 43, 53, 54].

When researchers measure health literacy, there is a wide array of components at the operational level that can be assessed by instruments [4, 38]. This review was able to identify a total of 31 different components that were addressed by 15 different instruments (see Table 3). As these studies each provide a different understanding of health literacy, including the chosen components to reflect upon, neither the conceptual/operational definitions nor the instruments were found to be comparable to one another, although some instruments apply and assess similar components. Moreover, not all authors provide their questionnaire items or make them accessible elsewhere, which makes understanding the tools difficult and prevents an accurate evaluation of exactly how each component was measured or indeed if it was measured at all. Thus, it is not possible to assess how far the predefined concept is represented in the questionnaire items or scales. Accordingly, we argue here that authors should make it quite clear which aspects of the model they intend to measure and which they actually do measure, and explain why they measure in that way. Of course, providing the questionnaire items could additionally support a better understanding of the instrument and the operationalisation of the underlying concept, if applicable. Despite the importance given to the social and cultural dimensions of health literacy [1, 8, 14, 16, 21, 26, 27, 56, 59], most instruments fail to address measurable components related to cultural competencies or social skills making it difficult to evaluate their actual impact on individual’s health literacy. Further components that are considered important to health literacy, such as empowerment, attitude, and self-efficacy [8], can be found in few instruments [40, 43, 45, 47].

This review has several limitations. Not all identified studies provided the questionnaire items which makes it difficult to undertake an accurate evaluation of the instrument without knowing how the construct was operationalised. This becomes even more critical if the components are only loosely described. In some cases, in which studies did not provide the questionnaire, the authors could be asked to provide the instruments to undertake further analysis with more detail. Unfortunately, several promising studies on health literacy measurement tools for children and adolescents, were not available in German or English [62‐64]. In the systematic review of Ormshaw et al. [38], further domain-specific and generic health literacy measurement tools for children and adolescents are reported. Having also analysed these studies during our review process, the authors of the original studies cited by Ormshaw et al. [38] did not provide necessary information to match the inclusion criteria of this review. Either some of the described instruments were used in studies we had already included, such as adaptations of TOFHLA, REALM, and NVS, or the articles did not describe the instruments in the necessary detail. Thus, these instruments were not included in this present review. However, these instruments may provide useful information for the development of new health literacy tools for children and adolescents. Furthermore, as health literacy is seen by some as a repackaging of the already existing concepts of health promotion and education, such as empowerment, self-efficacy, life skills, and the community approaches [87, 88], it may be useful to consider the methodology by which these concepts are being measured. For example, life skills measures could provide useful information for health literacy as well. This impression is supported by Sørensen and Brand [89], who reflected on the different translations and contexts of health literacy in which life skills or health-related competencies are used sometimes as synonyms for health literacy.

Regarding general limitations in the context of the health literacy measurement tools that were investigated during this study, the measurement of health literacy in children and adolescents is particularly difficult for two reasons. Firstly, there is no commonly accepted health literacy model or definition. In fact, there are approximately 250 different definitions or models available for adults [90], while there is a significantly smaller number available for children and adolescents [91]. Various constructions co-exist but are indeed fundamentally different, and they are apparently associated with uncertainties with respect to their measurement. The understanding that informs an assessment tool defines what kind of health literacy interpretation is measured. It is no surprise that across studies and tools dedicated to the assessment of health literacy, health literacy itself can be a different construct in each case. To progress in the field of measurement tools, consensus is essential in the conceptual field. Secondly, and in addition to conceptual concerns, another important issue is related to the cognitive and social development of children, which should be seriously considered when addressing the measurement of health literacy. Developing adequate health literacy levels requires certain prerequisites, such as skills; knowledge; level of experience with regard to health; autonomy and independence; and acquired social skills. Considering current conceptualisations of health literacy and how they are being measured, it is questionable whether young children will have acquired the competencies required to undertake the complex processing of health information as demanded by most of the existing health literacy models. As the achieved skill sets of children vary enormously across age groups and diverse backgrounds, there might be a need for age-appropriate and developmental-stage-adjusted concepts and their operationalisation considering the actual (social and cognitive) capabilities of children at different ages during childhood and adolescence.