Establishing the HLS-Q12 short version of the European Health Literacy Survey Questionnaire: latent trait analyses applying Rasch modelling and confirmatory factor analysis

Background The European Health Literacy Survey Questionnaire (HLS-EU-Q47) is widely used in assessing health literacy (HL). There has been some controversy whether the comprehensive HLS-EU-Q47 data, reflecting a conceptual model of four cognitive domains across three health domains (i.e. 12 subscales), fit unidimensional Rasch models. Still, the HLS-EU-Q47 raw score is commonly interpreted as a sufficient statistic. Combining Rasch modelling and confirmatory factor analysis, we reduced the 47 item scale to a parsimonious 12 item scale that meets the assumptions and requirements of objective measurement while offering a clinically feasible HL screening tool. This paper aims at (1) evaluating the psychometric properties of the HLS-EU-Q47 and associated short versions in a large Norwegian sample, and (2) establishing a short version (HLS-Q12) with sufficient psychometric properties. Methods Using computer-assisted telephone interviews during November 2014, data were collected from 900 randomly sampled individuals aged 16 and over. The data were analysed using the partial credit parameterization of the unidimensional polytomous Rasch model (PRM) and the ‘between-item’ multidimensional PRM, and by using one-factorial and multi-factorial confirmatory factor analysis (CFA) with categorical variables. Results Using likelihood-ratio tests to compare data-model fit for nested models, we found that the observed HLS-EU-Q47 data were more likely under a 12-dimensional Rasch model than under a three- or a one-dimensional Rasch model. Several of the 12 theoretically defined subscales suffered from low reliability owing to few items. Excluding poorly discriminating items, items displaying differential item functioning and redundant items violating the assumption of local independency, a parsimonious 12-item HLS-Q12 scale is suggested. The HLS-Q12 displayed acceptable fit to the unidimensional Rasch model and achieved acceptable goodness-of-fit indexes using CFA. Conclusions Unlike the HLS-EU-Q47 data, the parsimonious 12-item version (HLS-Q12) meets the assumptions and the requirements of objective measurement while offering clinically feasible screening without applying advanced psychometric methods on site. To avoid invalid measures of HL using the HLS-EU-Q47, we suggest using the HLS-Q12. Valid measures are particularly important in studies aiming to explain the variance in the latent trait HL, and explore the relation between HL and health outcomes with the purpose of informing policy makers.


Background
Health literacy (HL) is believed to have a vital impact on public health, including access to and the use of health services [1][2][3]. Low HL is associated with poorer health conditions [3,4], more frequent use of health services, longer hospitalisations [2,3] and higher mortality [4,5]. Further, some studies relate low HL to unhealthy behaviours, such as smoking [1,[6][7][8][9], low physical activity [7][8][9][10][11][12] and less use of preventive services [1,4,7]. Several instruments have been developed to measure individuals' HL and explore how health outcomes relate to HL. Some instruments assess basic reading and writing skills (such as the Test of Functional HL in Adults [TOFHLA] [13] and the Rapid Estimate of Adult Literacy in Medicine [REALM] [14]), while others assess specific health and cognitive domains (such as the European Health Literacy Survey Questionnaire [HLS-EU-Q47] [11,15]). The HLS-EU-Q47 was developed from a conceptual model covering four cognitive domains within three health domains, which constitutes a 12-cell matrix for item development [16]. The cognitive domains comprise the proficiency to access (A), understand (B), appraise (C) and apply (D) health information. These competencies are considered necessary to handle health information within health care (HC), disease prevention (DP) and health promotion (HP) settings.
The psychometric properties of the HLS-EU-Q47 have been investigated using principal component analysis (PCA) [15] and confirmatory factor analysis (CFA) [17][18][19]. When applying CFA in these studies, the health domains were treated as three uncorrelated or orthogonal subscales confirming a four-component structure reflecting the four cognitive domains. A Norwegian study among individuals with type 2 diabetes [20], in which the HLS-EU-Q47 was validated using Rasch-modelling in combination with CFA, revealed that the three health domains were highly correlated and that a 12-dimensional model (reflecting the four cognitive domains across the three health domains in the conceptual framework) obtained the best data-model fit. These results were later confirmed in a Taiwanese study among stroke patients [21]. A 12-dimensional scale requires a multidimensional approach and provides 12 different scores for each individual, which is impractical in clinical settings. Despite the theoretically identified and empirically confirmed multidimensionality of the HLS-EU-Q47, several studies assume a unidimensional latent variable and report one overall HL sum score [3,9,11,[17][18][19][22][23][24]. Beside multidimensionality, researchers have observed items with poor fit, redundant items, items displaying unordered response categories and subscales with low precision or reliability [20,21]. Therefore, we aim to clarify these issues based on a large sample of randomly selected individuals.
A unidimensional short version of the HLS-EU-Q47 will provide clinically useful data in a fast, reliable and accurate manner without applying advanced psychometry. Moreover, measurement scales should be anchored in theory and conceptual frameworks [25]. Hence, it is desirable that the short version reflects the conceptual HL model of Sørensen et al. [16], and that the 12 domains are equally represented. In addition, nurses and other health professionals should be able to use the short version to map patients' HL without applying advanced psychometric approaches. Several short versions of the HLS-EU-Q47 have been suggested, such as the HLS-EU-Q16 [11] and HL-SF12 [26]. However, we did not find any peer-reviewed publications describing the basis on which the items were selected, whether these short versions were validated using Rasch modelling, or whether they are sufficiently unidimensional. Therefore, based on the findings of Finbråten et al. [20] and the validation procedure in the present study, Rasch modelling and CFA are used to establish a unidimensional 12 item version of the HLS-EU-Q47.
Rasch models meet the assumptions and requirements of fundamental measurement, such as additivity [27], specific objectivity [28], sufficiency [29] and invariance [30]. Both Rasch modelling and CFA can be considered appropriate methods for assessing dimensionality [31,32]. CFA is used at an overall level to assess discriminant validity (dimensionality) and factorial validity (loadings or common variance), whereas Rasch modelling in addition provide detailed information at the item level, in addition to assessing local independency.
Against this background, the aims of our paper are to: (1) evaluate the psychometric properties of the HLS-EU-Q47 and associated short versions in the Norwegian population and (2) establish a parsimonious unidimensional short version (HLS-Q12) with sufficient psychometric properties. More specifically, we will test the following falsifiable hypothesis: When applied to Norwegian adults, the HLS-EU-Q47 represents a unidimensional, well-targeted scale with acceptable person separation (reliability), consisting of independent and invariant items at the ordinal level (i.e., ordered response categories) each displaying sufficient fit to the partial-credit parameterisation of the unidimensional polytomous Rasch-model.
The same hypothesis is tested for the HLS-Q12, and is used as a basis for comparing the psychometric properties of HLS-Q12 to the HL-SF12 and the HLS-EU-Q16.

Data collection
A random sample of 900 Norwegians aged 16 and over responded to the HLS-EU-Q47 during November 2014. For comparative purposes, we initially aimed at applying identical data collection procedures as the European HL survey [11], but owing to personal resources, it was not possible to perform face-to-face interviews. Hence, data were collected using computer-assisted telephone interviews. The international survey agency, Ipsos, with access to country representative samples, collected the data following detailed instructions offered by the researchers. The European HL survey sampled approximately 1000 respondents in each participating country [11] independent of population size. As there are significantly fewer inhabitants in Norway compared to most of the countries that participated in the European HL survey [33], 900 were sampled. The age limit was set to 16 years as this is the age of licence in relation to health in Norway.

Measures
The measures included the 47 items HLS-EU-Q47 HL survey questionnaire and additional person factors, such as gender, age and highest level of education completed.

The HLS-EU-Q47
Sørensen et al. [11,15] developed the HLS-EU-Q47 items to reflect their conceptual model, and suggested a 4-point rating scale with response categories ranging from very easy (1) to very difficult (4). In the present study, the rating scale was reversed to make a higher score indicate a higher proficiency at the latent trait. A 'don't know' category, which was later recoded to missing data, was added to record such spontaneous utterings from respondents during the telephone interviews.

Translation of the HLS-EU-Q47
The translation procedure, that involved forward and back translation, is thoroughly described in Finbråten et al. [20]. Cognitive interviews were used to explore item interpretation and clarify any ambiguities.

Data analysis
Rasch modelling was used to assess the psychometric properties of the HLS-EU-Q47 and as a basis for developing the HLS-Q12. Rasch modelling and CFA were used to study the psychometric properties of the short versions HLS-Q12 (our suggested short version), HL-SF12 [26] and the HLS-EU-Q16 [11].
The data were analysed up against the partial-credit parameterisation [34] of the unidimensional polytomous Rasch model (PRM) and the 'between-item' multidimensional PRM [35,36]. Below, the one-dimensional approach refers to unidimensional Rasch analysis. The two-and three-dimensional approaches refer to an oblique or unrestricted multidimensional Rasch analysis where the health domains are allowed to covary. The two-dimensional approach corresponds to the HC subscale and the combined DP and HP subscales, whereas the three-dimensional approach reflects the three health domains. The 12-dimensional approach refers to a similar analysis, where the four cognitive domains and the three health domains define 12 correlated subscales.
The RUMM2030 statistical software package [37] was used for the one-dimensional and consecutive Rasch analyses (treating the subscales as orthogonal or uncorrelated), and the ConQuest 4 statistical software package [38] was used for the one-dimensional and multidimensional Rasch analyses. CFA was performed using LISREL9.3 software [39].

Rasch modelling
The null hypothesis of a locally independent scale is weakened and might be rejected in favour of the alternative hypothesis, suggesting a multidimensional scale, in the following situations: When the proportion of individuals with significantly different person-location estimates on a pair of related subscales exceeds 5% (or the lower bound of the binominal 95% confidence interval (CI) exceeds 5%) [40][41][42][43]; when the subtest analysis in RUMM points to low common subscale variance and high unique subscale variance (i.e. low fractal index A < 0.8 and high fractal index c close to 0.5, respectively) [44]; when a multidimensional Rasch model obtains a better data-model fit than a one-dimensional Rasch model (i.e. significantly lower deviance or -2log-likelihoods using a likelihood-ratio test for nested models and AIC for non-nested models) [45]; and/or when absolute values of Rasch-model residual correlations among items exceed 0.3 [46].
Likelihood-ratio tests (LRT) were used to compare data-model fit for nested models, such as the one-, two-, three-and 12-dimensional Rasch models, where the first implies parallel subscales (correlations between all theoretically defined subscales were fixed at 1) and the latter implies freely correlated subscales. The LRT statistic "difference in deviance" is asymptotically χ2 distributed with degrees of freedom (df ) equal to the difference in the number of estimated model parameters comparing two nested models [47]. Akaike's information criterion (AIC) [48] was used to compare the data-model fit for non-nested models, such as the various short versions. Similar to Allen and Wilson [49], both deviance and AIC are reported when applying the consecutive approach to the HLS-EU-Q47 (calculated by adding the values for each of the three subscales). Lower values of deviance and AIC indicate better data-model fit. In addition to deviance and AIC, total item chi-square with p-values was used to assess the data-model fit of the short versions.
Items were interpreted as under−/over-discriminating relative to the Rasch model when the infit MNSQ were significantly different from the expected value of 1 (indicating perfect fit to model) or, more precisely, when the infit value was above/below the 95% CI and the corresponding t value was > 1.96 [45,50].
In addition to infit, chi-square probability was used to assess item fit for the various short versions. Items were considered as misfitting at high chi-square values and probability values lower than a Bonferroni-adjusted 5% [51]. Item difficulty was described using item-location estimates.
Applying the consecutive approach to the HLS-EU-Q47 and the one-dimensional approach to the short versions, differential item functioning (DIF) was explored using analysis of variance (ANOVA) [40] for the dichotomized person factors of gender, age and level of highest completed education: age was dichotomised as above or below the sample mean (47 years), and highest level of completed education was split into "compulsory and upper secondary school" versus "university level".
The ordering of the response categories was examined applying the one-dimensional and the consecutive approaches. Significantly different thresholds in "correct" order [52,53] were used as evidence for ordered response categories. The targeting of the (sub)scales was evaluated by comparing the distribution of the item threshold estimates to the distribution of the person estimates [51]. For more detailed descriptions of the analyses performed, see for example, Andrich [30], Hagquist [53], Hagquist, Bruce and Gustavsson [54] or Finbråten et al. [20].

Confirmatory factor analysis
The null hypothesis of a one-dimensional scale is weakened when overall fit indexes for multifactorial CFA are superior to a one-factorial CFA [55]. Owing to categorical raw data, we defined the variables as ordinal in the "lsf" file and estimated the polychoric item correlations and their asymptotic covariance matrix using the PRELIS application to LISREL [56]. As target values for goodness-of-fit (GOF) indexes is based on maximum likelihood estimation (ML), we used robust ML to obtain the Satorra-Bentler scaled chi-square GOF index (SB scaled χ 2 ) [57,58]. The following overall GOF indexes for model comparison based on the null model were used: SB scaled χ 2 , standardized root mean square residual (SRMR), root mean square error of approximation (RMSEA), comparative fit index (CFI) and non-normed fit index (NNFI, or Tucker-Lewis index [TLI]). Schumacker and Lomax [59] recommend SRMR < .05, RMSEA ≤ .05, and CFI and NNFI ≥ .95 as GOF index target values, whereas Hu and Bentler [60] claim that SRMR and RMSEA values close to or below 0.8 and 0.6, respectively, indicate sufficient overall fit. The AIC and the Bayesian information criterion (BIC) were also taken into account. AIC and GOF indexes were used to compare the various short versions and to compare the one-factor to the three-factor approaches for the various short versions. Item communalities (squared factor loadings for completely standardized solution) describe the shared variance among the items that is accounted for by the latent variable, HL [61].

Reliability
Person separation index (PSI), person separation reliability (PSR) and the H coefficient [62] (obtained from RUMM, ConQuest and calculated manually based on communalities from CFA) were used as measures of reliability. Values exceeding 0.85 (0.65) are sufficient if conclusions are to be drawn at the individual (group) level [63].

Developing a one-dimensional short version
The suggested short version in the present study, the HLS-Q12, was developed from analyses of the HLS-EU-Q47 in two Norwegian samples; people with type 2 diabetes [20] and the general population. The suggested version, HLS-Q12, was developed by stepwise exclusion of poorly fitting items, items displaying DIF and items collecting redundant information, together with a qualitative evaluation of the item content in light of the conceptual framework. For dependent items, the item with the most essential information was retained, while the other item was discarded. We ensured that the items in our final scale (specified in Table 5) were distributed across the HLS-EU 12-cell matrix [11,16].

Missing data
The dataset we analysed was complete, but there was, on average, 29 (3%) 'don't know' responses per item, which we recoded as missing values. Item 2 had the highest number of 'don't know' responses (69 or 8%). Rasch modelling was performed on incomplete data (except in the subtest analyses), whereas LISREL used list-wise deletion as the default method, reducing the sample size from 900 to effective sample sizes of 699, 680 and 670 for the HLS-Q12, HL-SF12 and the HLS-EU-Q16, respectively.

Results
The sample (n = 900) distributions for gender and age reflected the population distributions, while individuals with higher education were somewhat over-sampled (Table 1).

Psychometric properties of the HLS-EU-Q47 Overall analyses of HLS-EU-Q47
Overall analyses included those for dimensionality, data-model fit, response dependence and reliability. The initial analyses of dimensionality indicated that neither the entire HLS-EU-Q47 nor the three health domains, HC, DP and HP, were sufficiently unidimensional (the respective proportions of significant t-tests were 21, 9, 13 and 10% [not reported in Table 2]).
Using LRT, the observed data were more likely under a 12-dimensional Rasch model than a one-, two-or three-dimensional Rasch model (Fig. 1). As expected, the data-model fit improved significantly in each step from the one-dimensional through the 12-dimensional model. For example, the drop in deviance from the three-to the 12-dimensional model was ΔD = 3878 ≈ χ 2 [Δdf = Δestimated parameters (ep) = 72], p < 0.01, critical value = 103. The three-dimensional model had a lower AIC and a better fit than the consecutive approach of the three health domains, HC, DP and HP (ΔAIC = 1129). The correlations between HC and DP, HC and HP, and DP and HP were r = 0.81, 0.71 and 0.83, respectively.
Applying the one-dimensional approach, we observed nine pairs of dependent items (1 and 2, 22 and 23, 41 and 42, 44 and 46, 45 and 47, 3 and 7, 12 and 28, 17 and 21, 30 and 37), of which the latter four pairs surprisingly appeared across the theoretically defined subscales. Applying the consecutive approach to the three health domains, only one pair of dependent items was observed within each health domain (1 and 2, 17 and 21, and 41 and 42, respectively).
Applying the 12-dimensional approach to the HLS-EU-Q47, we observed relatively low subscale reliability indexes, most likely owing to few items per scale-on average 3.9 items per subscale (see PSR values in Fig. 1). Consequently, we observed acceptable reliability indexes for the one-dimensional approach and for the subscales of the two-and the three-dimensional approaches.

Analyses of HLS-EU-Q47 at item level
Analyses at the item level included analyses of item fit, item discrimination, DIF and ordering of response categories. Table 2 shows the item fit when applying the one-, three-and 12-dimensional approaches.
Unordered response categories were observed for items 8, 15 and 21 when applying the one-dimensional approach, as well as when analysing the HC, DP and HC health domains consecutively (Table 2).    Note. This table reports the item-fit indexes after applying the one-dimensional (subscale correlation fixed to 1), three-dimensional and 12-dimensional approaches (treating the subscales as correlated). The analyses of the ordering of response categories (ordered) and differential item functioning (DIF) are based on the one-dimensional and consecutive approaches (treating the subscales as orthogonal or uncorrelated) for the three health domains and the 12 subscales. Ordered: According to the t-tests, none of the short versions could be deemed sufficiently unidimensional (the proportions of significant t-tests were above 5% and the lower binominal 95% CI proportion was 0.07 for both the HLS-Q12 and the HL-SF12, whereas the lower binominal 95% CI-proportion was 0.09 for the HLS-EU-Q16 (Table 3). Nevertheless, the fractal indexes indicated that the HLS-Q12 and HL-SF12 had relatively high amounts of common variance (A = 0.90 and 0.91, respectively), high subscale correlation (r = 0.80 and 0.87, respectively), and relatively high unique subscale variance (c = 0.50 and 0.39, respectively).
Using LRT, the data fit better to the three-dimensional models compared to the one-dimensional models of the various short versions (three-dimensional models had significantly lower deviance, except for the HL-SF12, where the difference was insignificant). The GOF indexes obtained via CFA were also better for the three-factor models as compared to the one-factor models of the various short versions (Table 4).
Comparing the short versions, the HLS-Q12 obtained best fit to the Rasch model (had the lowest total item chi-square value and the lowest AIC; Table 3) whereas the HL-SF12 obtained best fit when performing CFA (had the lowest likelihood estimates and the best GOF indexes). The HLS-Q12 also had acceptable GOF indexes ( Table 4). The HLS-EU-Q16 displayed the weakest overall fit indexes (highest total item χ 2 value and highest log likelihood estimates).
All the one-dimensional short versions had acceptable reliability values (PSI and PSR values above 0.75 and the H coefficient was above 0.82). The HLS-EU-Q16 had the highest reliability indexes (PSI = 0.830, PSR = 0.826 and H = 0.882). The reliability indexes were slightly higher for the HLS-Q12 than for the HL-SF12 (Tables 3 and 4). Applying the three-dimensional approach, the PSR and H values of the correlated subscales were below the recommended values (Tables 3 and 4). The HLS-Q12 was the best-targeted, with a mean person location of 0.759. The mean person locations for the HL-SF12 and Fig. 1 Model fit of the HLS-EU-Q47 after applying various analysis approaches. Figure 1 shows the overall fit statistics for the one-dimensional approach (all subscale correlations fixed to 1), the consecutive approach (treating the three health domains as orthogonal or uncorrelated) and the two-, three-and 12-dimensional approaches (treating the theoretical subscales as correlated). A: access, B: understand, C: appraise, D: apply (cognitive domains). HC: health care, DP: disease prevention, HP: health promotion (health domains). Δ: change in parameter, AIC: Akaike's information criterion, cv: critical value, D: deviance, df: degrees of freedom, ep: number of estimated parameters, LRT: likelihood ratio test. PSR: person separation reliability based on marginal maximum likelihood estimate/Warm's mean weighted likelihood estimate HLS-EU-Q16 were 0.816 and 1.020, respectively (not shown in the table).

Analyses of HLS-Q12, HL-SF12 and HLS-EU-Q16 at the item level
Compared with the HL-SF12 and the HLS-EU-Q16, the HLS-Q12 showed best fit at the item level as there were no occurrences of misfitting items, items displaying DIF with regard to available person factors or items with unordered response categories. For the HL-SF12, items were within-item biased (items 15, 26 and 30 displayed DIF), and one item with unordered response categories was observed (item 15). Problems with regard to DIF (items 5, 11, 13, 21 and 37) and unordered response categories (items 8 and 21) were also observed for the HLS-EU-Q16. In addition, item 11 under-discriminated and showed significant misfit (chi-square probability = 0.001; Table 5). A high level of unique variance was observed for the items of the various short versions (relatively low communality values).
Investigating the item-location estimates of the HLS-Q12, items 28 (judge if the information on health risks in the media is reliable) and 10 (judge the advantages and disadvantages of different treatment options) had the highest location estimates (1.068 and 0.892, respectively) and were thus the hardest to endorse. Items 23 (understand why you need health screenings) and 32 (find information on healthy activities such as exercise, healthy food and nutrition) were the easiest to endorse (item-location estimates of − 1.158 and − 1.135, respectively).

Discussion
Based on our national random sample (n = 900), we found that empirical evidence did not support our null hypothesis associated with the psychometric properties of the HLS-EU-Q47. By excluding poorly fitting items, items displaying DIF and items violating local independence, we succeeded in establishing a psychometrically sound parsimonious 12-item version (HLS-Q12). From a measurement point of view, we found that the HLS-Q12 outperformed other available short versions of the HLS-EU-Q47, such as the HL-SF12 and the HLS-EU-Q16.

Psychometric properties of the HLS-EU-Q47 Overall analyses of HLS-EU-Q47
A 12-dimensional model described the HLS-EU-Q47 data best. This result is perfectly in line with prior research [20,21]. Hence, applying the HLS-EU-Q47 Table 3 Unidimensionality, data-model fit and reliability applying Rasch modelling of the various short versions HLS-Q12 a HL-SF12 [26] HLS-EU-Q16 [11] One-dimensional Threedimensional  recommends a complex 12-multidimensional approach returning 12 subscale-scores for each individual. This is of little practical use, especially as proficiency cannot be compared across the different subscales owing to relative points of zero. Owing to few items in each subscale, few of the 12 subscales were sufficiently reliable. Contrary to common practice when using the HLS-EU-Q47, we cannot recommend reporting either total or health domain subscale scores. The three-dimensional approach, reflecting the three health domains, obtained better data-model fit than the one-dimensional approaches. The three health domain subscales each returned sufficiently large reliability indexes, but lower indexes than those reported by the HLS-EU Consortium [11], Nakayama et al. [18] and Duong et al. [19].
Prior research on people with chronic disease (type 2 diabetes) [20] indicates that the HC subscale brings notable multidimensionality into the data. It is very interesting that this not was supported in our recent analyses of a national sample, because it indicates that the "health care" subscale actually work differently for patients with chronic disease than in the population as a whole. This seems reasonable, as people with chronic disease have more experiences with using health care facilities than the general population.

Analyses of HLS-EU-Q47 at item-level
Like previous studies [20,21] we observed items violating local independence, items displaying DIF, and poorly fitting items. Poorly fitting items which over-or under-discriminate tap into constructs other than the latent trait [64]. Over-discriminating items measure "too much of something else" that is positively correlated with the latent trait and are therefore viewed as less problematic than under-discriminating items. Contrary to Finbråten et al. [20], who found that few items displayed DIF in people with chronic disease (type 2 diabetes), we found several items displaying DIF for age and education, which means that people with higher versus lower levels of education, as well as younger versus older people, probably perceive these items differently. Consequently, comparisons of HL across age groups or educational levels would be invalid [54]. The increased number of items displaying DIF in our study might be expected as the sample of people with type 2 diabetes in Finbråten et al. [20] mainly consisted of elderly persons.
Psychometric properties of the of HLS-Q12, HL-SF12 and HLS-EU-Q16 Overall analyses of the of the HLS-Q12, HL-SF12 and HLS-EU-Q16 Based on the PCA and t-test procedures in RUMM, none of the short versions stood out as sufficiently unidimensional. However, considering the lower binominal 95% CI proportions of 0.07, the two 12 item versions HLS-Q12 and HL-SF12 could be considered sufficiently unidimensional [54]. Hagell [41] claims that the width of  When investigating dimensionality in a larger sample, binominal 95% CI-proportions might represent acceptable values. Moreover, unidimensionality is a relative matter and depends on the level of precision. Evaluation of unidimensionality should also include an assessment of the purpose of the measurement, together with theoretical and practical considerations [30,41].
Comparing the one-and three-dimensional approaches to the short versions, the three-dimensional approach obtained lowest deviance and sufficient values for the GOF indexes. Like some previous studies [32,65] we found that Rasch modelling and CFA returned similar results. Thus, a three-dimensional approach could be recommended for the short versions. However, when applying a three-dimensional approach the three subscales obtained low reliability indexes. The difference in deviance between the one-and three-dimensional approaches was significant; but the differences were rather small for the two 12 item versions.
A one-dimensional approach to the HLS-Q12 and HL-SF12 could probably be defensible as the results point to minor violations of unidimensionality. Further, the one-dimensional approach yields higher reliability indexes than each of the subscales of the three-dimensional approach. Providing one HL score for each individual, the one-dimensional approach to the HLS-Q12 and the HL-SF12 is practical from a clinical point of view. In addition, the HLS-Q12 and HL-SF12 are in line with the conceptual model of Sørensen et al. [16], but dimensionality should be further explored in future studies.

Comparing various short versions
Comparing the psychometric properties of the various short versions, the HLS-Q12 obtained best fit to the Rasch model, whereas the HL-SF12 obtained best fit through CFA. The HL-SF12 was found to have several weaknesses, including items with unordered response categories (item 15) and several items displaying DIF (items 15, 26 and 30). Owing to a larger number of items, the HLS-EU-Q16 stood out as more reliable than the other short versions [63]. However, the HLS-EU-Q16 could not be considered sufficiently unidimensional, it had the highest AIC, and did not yield acceptable GOF indexes. In addition, problems regarding item misfit, DIF and unordered response categories were observed.
Altogether, the HLS-Q12 has better psychometric properties than the HL-SF12 and the HLS-EU-Q16. The HLS-Q12 can be deemed the best-targeted scale and free from under-discriminating items and DIF. The HLS-Q12 reflects the conceptual framework, and it is well balanced because it consists of one item from each of the 12 dimensions. The HLS-Q12 version could be well suited for use in HL screenings at both the individual and community levels. Nevertheless, in future studies, we recommend extending the number of response categories from four to six to increase the reliability of the scale [66]. However, item 38 represents a specific concern because this item under-discriminated when applying one-, three-and 12-dimensional approaches to the entire HLS-EU-Q47. In contrast, when we applied a one-dimensional approach to the HLS-Q12, the item showed good model fit. In future validation of the HLS-Q12, this item should be further investigated.

Item difficulty
Using the HLS-Q12, items 28 (judge if the information on health risks in the media is reliable) and 10 (judge the advantages and disadvantages of different treatment options) were the hardest items to endorse. Consequently, people may use information that is not evidence-based and potentially harmful to their health. Hence, health professionals should guide individuals in accessing valid and reliable health information. Further, health professionals should help individuals to develop the ability to critically assess health information from different sources. People also need guidance in judging the advantages and disadvantages of various treatment options. Low HL might cause difficulties with promoting and taking responsibility for one's own health, as well as participating in shared descision-making regarding health issues. According to Nutbeam [67], nurses and other health professionals must become aware of the effects of low HL. Hence, nurses and other health professionals should map HL in their patients and adapt health communications to the individual or target group.
Further research on HL among Norwegians is needed, especially research examining the correlation between HL and health behaviour and the correlation between access to and the use of health services. It would also be interesting to compare HL across cultures and nationalities. However, before comparing HL across nations and cultures, DIF analyses should be performed to investigate the effects of cultural differences on participants' interpretations of the content of the items.

Limitations
Ipsos, the agency that performed the sampling and data collection, guarantees representativeness. However, the educational level of this sample was higher than that of the average Norwegian population. In this study, more than half of participants had education at the university level, while this is true of only 33% of the general population [68].
Although around 1000 individuals were recruited from each of the included EU countries, 900 individuals were included in this study. In Rasch modelling, there is no exact recommendations for sample size. However, Linacre [69] recommends 250 individuals for polytomous data and 10 extra individuals per response category. Hair [70] recommends at least 300 individuals when performing CFA. According to these recommendations, a sample size of 900 would be sufficient.
The HLS-Q12 was developed based on the results of Rasch modelling of the HLS-EU-Q47 using two populations, people with type 2 diabetes [20] and the general Norwegian population. However, its further application to other populations will yield more evidence regarding the validity and reliability of the scale. Future validation should be performed using multinational data.