Psychometric properties of the Norwegian version of the Kidscreen-27 questionnaire

Background The Norwegian version of the Kidscreen-27, a measure of generic health-related quality of life, has not yet been validated. Thus, the aim of this study was to investigate the reliability and validity of the Norwegian Kidscreen-27, in 10 year-old children. Methods The Kidscreen-27 consists of five domains and was validated in a cross-sectional study of 1085 school children (52.5 % boys). In addition a subsample of 56 children also had repeated measures in order to study test-retest reliability. Results Cronbach's alpha values ranged from 0.73 to 0.83, while intraclass correlation values over time ranged from 0.71 to 0.81. The domains of physical well-being, psychological well-being and autonomy & parents improved over time (Ps < 0.05), while social support and school environment domains did not. Confirmatory factor analysis showed an acceptable overall model fit: X2 = 707; df = 310; P <0.001, root mean squared error of approximation = 0.037, the comparative fit index = 0.96 and the Tucker-Lewis index = 0.95. All factor loading were > 0.40. The Kidscreen-27 domains were significantly associated with general life satisfaction as measured with the Cantrils Ladder (Spearman rank correlations ranged from 0.29 to 0.59, Ps < 0.05). Conclusion The Norwegian version of Kidscreen-27 has good reliability and validity. Electronic supplementary material The online version of this article (doi:10.1186/s12955-016-0460-4) contains supplementary material, which is available to authorized users.


Background
The Kidscreen-27 is a well-validated, short, multidimensional measure of generic health-related quality of life (HRQoL) in children and adolescents which is available in many languages [1,2]. However, unlike the longer Kidscreen-52, the Norwegian version of the  has not yet been validated [3]. The Kidscreen-27 is a particularly useful instrument for assessing HRQoL in younger children as a consequence of its content, ease of completion and the volume of available data with which international comparisons can be made. As our research group has planned several studies on HRQoL in younger children, the aim of this study was to investigate the reliability and validity of the Norwegian Kidscreen-27 questionnaire in 10-year-old children.

Methods
We applied both a cross-sectional and a prospective design in this study, which was approved by the regional ethical research committee (2013/1893, 2012/1089). We recruited children aged 10 years from 52 schools in Western Norway for the cross-sectional study, whereas children from one school were recruited for the prospective study. Written informed consent was obtained from the parents or guardians of the participants. The children were given brief oral information about the questionnaire by their teacher (the information was identical with the standardised information on the first pages of the Kidscreen-27 questionnaire), and then completed it while they sat at their desks in the classroom. After doing so, they were told to carry on with their school work in order to minimize any noise until all the other children were finished. They were allowed to ask the teacher for help if they did not understand any of the questions. Children with reading difficulties were helped by a teaching assistant, as would be the case normally. For a subsample of the children the Kidscreen-27 was administered at three time points: at the beginning of the school day (Test 1); at the end of the same school day (Test 2); and during the middle of the following school day (Test 3).

Kidscreen-27
Items of the Kidscreen-27 are derived from the Kidscreen-52 questionnaire [2]. It has five domains: physical well-being (5 items); psychological well-being (7 items); autonomy & parents (7 items); social support & peers (4 items); and school environment (4 items). We used the methodology given in the developers manual to obtain the T-scores; mean (±SD) scores of 50 ± 10 define normality for children and adolescents aged 8-18 years across Europe [2]. Higher scores indicate a better HRQoL. The Kidscreen-27 is standardized so that a difference of <0.2 points is considered trivial, 2.0 -4.9 as small, 5.0-7.9 as moderate and ≥8 as large effects [2,4]. There are two official forms of written Norwegian, Bokmål and Nynorsk. They are very similar, and it is mandatory that they are both taught in Norwegian schools. The Kidscreen-27 items are available in Bokmål [3], and has been translated from English into Norwegian in accordance with respected guidelines [5]. We decided to use Nynorsk in this study. The linguistic skills required to produce a Nynorsk version of Kidscreen-27 from the Kidscreen-52 are minimal. The translation was undertaken by the first three authors and modified after discussion with a professor of Norwegian languages (see acknowledgements). The final version was based on consensus and no problematic or difficult issues were noted. This process was approved by the European Kidscreen Group.

Cantrils life satisfaction ladder
In order to study convergent validity we assessed general life satisfaction using the adapted version of the Cantrils Life Satisfaction Ladder. This measure has been used in World Health Organisation surveys of children and adolescents, including in Norway [6]. The child is presented with a picture of a ladder with steps ranging from 0 to 10. They are told that the top step (10) represents the best possible quality of life, while the bottom (0) represents the worst. They are asked to indicate where on the ladder they currently consider themselves to be. This question was assessed together with the Kidscreen-27 in a subsample of the children.

Statistical analysis
Internal consistency was assessed by calculating Cronbach's alpha values; values ≥0.7 were considered satisfactory [1,2]. Floor and ceiling effects were demonstrated by the percentages of children with the lowest and highest possible scores. Test-retest reliability (Tests 1-3) was assessed by calculating single measures intraclass correlation coefficients (ICC), using a two-way mixed model with an absolute agreement definition; ICC values ≥0.7 were considered satisfactory [1,2]. A linear mixed model based on restricted maximum likelihood estimation with random intercept for subjects was used for analysis of change in Kidscreen-27 scores over the three time points (Tests 1-3). Individual variability was described by presenting mean differences ± SDs and 95 % limits of agreement; Bland Altman plots were used to graphically display the variation (Tests 1-3).
The structure validity of the questionnaire was tested using confirmatory factor analysis (CFA). The overall model fit was assessed using the chi-square test statistic. However as this test is highly sensitive to sample size, we used alternative fit indices having the following cutsoff suggesting acceptable fit; the root mean squared error of approximation (RMSEA), (<0.08); the comparative fit index (CFI) (>90) and the Tucker-Lewis index (TLI) ((>90), while factor loadings should be > 0.40 [7].
For convergent validity, we used the Spearman rank correlation (r s ) to test whether the Kidscreen-27 scores were positively correlated with the Cantrils Ladder score. Based on previous research, we hypothesized that the Kidscreen-27 psychological well-being domain would be the one most strongly correlated with the Cantrils Ladder score [1,2]. Correlation coefficients <0.1 were considered trivial, 0.1 -0.29 as small, 0.30 -0.49 as moderate, and ≥0.5 as high [4].
The software Prism version 6.05 was used to calculate and display the results of the mean differences with 95 % limits of agreement and the Bland Altman plots. The CFA analyses were conducted with Stata version 14. Other statistical analyses were performed using IBM SPSS version 21. Two-sided P-values <0.05 were considered statistically significant.

Results
A total of 1085 children (52.5 % boys) participated in the study (85 % response rate). Table 1 Table 2. Physical wellbeing, psychological well-being and autonomy & parents significantly improved over time. The variability over time in the five Kidscreen-27 domain scores is shown in Table 3. By way of an illustrative example, Fig. 1 shows Bland Altman plots for the physical well-being domain.

Discussion
The findings of this study demonstrate that the reliability and validity of the Norwegian version of the Kidscreen-27 are good. A large body of research findings regarding the Kidscreen-27 enables a direct comparison with our results [1,2]. Cronbach alpha values in the current study ranged from 0.73 to 0.83. This compares to 0.80 -0.84 reported in the literature when the instrument was administered to children aged between 8 and 18 years [2]. We found almost no floor effects, but a moderate ceiling effect in the social support & peers domain. These results are similar to previously reported findings [2]. The ICC values in the current study ranged from 0.71 to 0.80, which were slightly higher than the range of 0.61 -0.66 reported previously [1]. This finding may be a consequence of different test-retest intervals (1 day vs. 2 weeks). Additionally, our study showed variability in Kidscreen-27 scores over time in some individuals. This has been reported with other HRQoL measures and in various populations [8,9]. Although it could be a consequence of measurement error, natural fluctuations in HRQoL could also be responsible [10]. Previous research has suggested that there is a small retest effect, in order that Kidscreen-27 scores tend to rise if assessed multiple times, even without any interventions [2]. Our study demonstrated this in three of the domains, even though the test-retest intervals were short. Whether this effect is still present if the test-retest interval is longer, is unknown. However, it suggests that studies using Kidscreen-27 to assess the effectiveness of interventions must have a control group.
The CFA showed an acceptable overall model fit, especially when the largest modification indices were taken into account. We think the added covariation terms make sense conceptually. Item 6 and 7 on the autonomy & parents scale are related to perceived family economy while the other items on that scale are not. The questions on item 4, 5 and 6 on the psychological well-being are worded "negatively" such as "Have you felt sad? "(item 4), while item 1, 2, 3 and 7 are worded "positively such as" Has your  The sizes of the correlation coefficients are similar to those that have been reported previously when assessing convergent validity using a range of instruments and in various populations [1,2]. This study was limited as only children aged 10 years were included. It is more usual to assess a wider age group. Our decision to do so was a pragmatic one, as we required a validated instrument to assess HRQoL in 10-year-olds, for the purpose of a randomized controlled trial. However, we propose that if good reliability and validity levels of an assessment instrument are demonstrated in large sample of younger children, it is possible that this will also be the case in older children and adolescents as they typically have better cognitive skills. The variability in Kidscreen-27 scores in children and adolescents are also quite similar [2]. However, because cognitive skills are not the only aspects to consider in the assessment of the psychometric properties of HRQoL instruments, especially in pediatric, pre-adolescent and adolescent individuals, the properties of this instrument should also be investigated in Norwegian children older than 10 years. Our short test-retest intervals might be criticized.
However, in addition to the points made above, it should be noted that a previous study found that testretest effects were similar whether the interval is 2 days or 2 weeks [11].