This study is an observational cross-validation of a HRQoL instrument involving 114 deaf children and adolescents aged 8-to-18 years. All of them completed the questionnaire twice with a 2-4 weeks interval. Participants were recruited from schools, colleges, high schools and associations of deaf people from the Regions of Castilla-La Mancha and Madrid. First we contacted with 79 schools and eight associations in Castilla-La Mancha. For the sake of a greater sample we contacted with seven schools and fourteen associations in Madrid. Finally, 61 schools and three associations of Castilla-La Mancha and four schools of Madrid participated in the study. Our final sample was composed of 75 subjects from Castilla-La Mancha and 39 subjects from Madrid. Fieldwork took place during the 2009-2010 and 2010-2011 school years.

The study protocol was approved by the Clinical Research Ethics Committee of the Virgen de la Luz Hospital in Cuenca, and also had the express support of the Provincial Office of Education in Cuenca and that of the Confederation of Deaf People. A letter was sent to parents informing them about the aims of the study and inviting them to accept the participation of their children by signing an informed consent document. The protocol follows the principles of the Declaration of Helsinki,24 and information management was conducted according to the provisions of the Organic Law of Protection of Personal Data (Law 15/1999)25 and its regulations (RD 1720/2007).26

As well as socio-demographic variables (age, sex, socioeconomic status), other variables were included: degree of hearing loss, type of deafness (congenital or acquired), time of onset of deafness (pre-lingual or post-lingual), preferred form of communication (sign language or spoken language), functionality of communication, communication problems, parental responsibilities and expectations of the deaf child's development, parental acceptance of the child's deafness, and education (mainstream or special school). HRQoL responses were obtained by self-administered versions of the KIDSCREEN-27 for children and adolescents adapted and translated into LSE.

The KIDSCREEN-27 questionnaire consisted of 27 items which are used to assess HRQoL across five dimensions: physical well-being (five items); psychological well-being (seven items); autonomy and parents (seven items); peers and social support (four items); and school environment (four items). The items assess either the frequency of behaviour/feelings or, in fewer cases, the intensity of an attitude and are answered on a five-point scale with a timeframe of one week. Higher scores indicate better HRQoL.27

Because many deaf children have difficulties with reading, the children's version of the KIDSCREEN-27 was translated and adapted into LSE.22 So that this could be self- administered and have utility for other professionals or other Spanish-speaking communities who use LSE. In addition, a web tool was designed in LSE for the questionnaire (http://www.cess.uclm.es/qd/) so that it is fully accessible to deaf children and adolescents who are users of this language.

Since we did not have a sufficient number of the sample's families with deafness to validate an adapted the adult version of the instrument, we have only adapted the children's version of the KIDSCREEN-27.

Back translation was used in order to ensure cultural alignment of the original (written Spanish) and new (LSE) versions. Three bilingual deaf adults (Spanish and LSE) with high levels of reading skills took part in the translation of the Spanish questionnaire to LSE (and the subsequent preparation of the carrier medium). They were supplied with a Spanish version of the KIDSCREEN-27 children and were asked to translate each item into LSE. Working alongside an LSE interpreter, a consensus was reached on the best version of each item and the first recording of the survey was made.

The visual material was delivered to another LSE interpreter and to a post-lingually deaf bilingual who had not taken part in the translation of the written version to LSE. They translated independently each item from LSE to written Spanish. The back-translations were then compared with the original Spanish written version to assess the degree of concordance.

In this first revision of the original instrument and of the translation to LSE, some items with unclear meaning (i.e. “Have you been in a good mood?”, “Have you been happy with the way you are?”, among others) were identified, and possible alternatives were discussed. These items were then reassessed by a bilingual deaf person and discussed with the interpreter who had taken part in the first adaptation. A second translation was then created by them with special emphasis on clarifying the ambiguous items. A second back-translation was then prepared by two additional translators: one was deaf, and one with a moderate hearing loss; both were bilingual and with a high level of reading skill.

Finally, a group of LSE experts and deaf people met in order to select the final LSE version of each item for video-recording. There were two signers for the video: a female Spanish/LSE interpreter and a hearing young male, whose father was deaf and who was bilingual in LSE and Spanish.

The software for delivering the questionnaire was specially developed to meet the characteristics and needs of deaf people. In order to enable the questionnaire to be self-administered, written Spanish and LSE are displayed for both the questions and the response options, and it is possible to reply the LSE video clips before responding. Each answer choice is marked with a colored dot as a visual aid for deaf children.21,28 Responses are completed on a paper version of the questionnaire, and the respondent answers each item before moving to the next.

A descriptive analysis of the sample characteristics was performed, the floor and ceiling effects (proportion of cases obtaining the minimum and maximum scores respectively) were calculated; and the overall score and scores for each of the dimensions were also calculated according to the KIDSCREEN-27 questionnaire instructions.

An exploratory factor analysis (EFA; main components analysis) was conducted to check if the dimension factor was saturated by each item in order to ensure that it constituted a valid indicator. The Kaiser-Meyer-Olkin (KMO) test and Bartlett's sphericity test were also used to assess whether the factor analysis was appropriate to test the scale structure. A factor analysis would be appropriate if the KMO test reaches values above 0.5 and the Barlett test obtains values lower than 0.05.

The adequacy of the LSE-adapted KIDSCREEN-27 was assessed by means of confirmatory factor analysis (CFA) using IBM SPSS Amos 19 software.29 The adjustment of the model was analyzed in various ways. A chi-square (χ2) was used to assess the fit to the data of the sample. The goodness of fit was analyzed using the criteria of Hu and Bentler,30 including the comparative fit index (CFI) and standardized root mean square residual (SRMR).

An analysis of multiple groups was performed to examine gender differences. A χ2 test of differences was used to determine whether differences in the factor loadings were statistically significant.

To test the factor structure of the KIDSCREEN-27, the sample was divided into two subsamples and EFA and CFA were conducted using these two sub-samples respectively.

The internal consistency of the scale was analyzed using Cronbach's alpha coefficient for both the total scale and for each dimension.

For the test-retest reliability the deaf children completed the questionnaire twice with a 2-4 weeks interval. This time gap was selected as it was considered long enough to ensure that participants would not remember their responses and that it was short enough to avoid significant changes in the quality of the participants’ lives.

The intra-class correlation coefficient (ICC) was used to estimate the correlation between baseline scores on the questionnaire and the scoring of their responses 2-4 weeks later.

Except for CFA analyses were performed with IBM SPSS Statistics 19 software.

An exploratory analysis identified six factors with an eigenvalue greater than 1, but in the screen plot it was observed that a solution with five factors would be appropriate (Fig. 1). This five-factor solution showed good sampling adequacy indexes (Kaiser-Meyer-Olkin 0.822; Barlett sphericity p<0.001), explained 59% of the variance, and was more comparable to the analysis of the original Spanish version. Some differences between the original Spanish KIDSCREEN-27 and the LSE version in the allocation of items after Varimax rotation were observed (Table 2).

Figure 1.

Sedimentation graph of factor components of KIDSCREEN-27 LSE version.

(0.08MB).

The factorial structure of the LSE version with the items arranged in the same way than the original Spanish version was tested by CFA. A five-factor model showed acceptable goodness of fit (χ2=551.8; CFI=0.808; normal fit index [NFI]=0.648; p=0.0001; SRMR=0.0832; root mean square error of approximation [RMSEA]=0.082) (Fig. 2). The model displayed good fit by sex and the factor loadings showed no differences between males and females (χ2 DIFF=1218 – 1195=23, df DIFF=650 – 628=22, p=0.29).

Figure 2.

Confirmatory factor analysis diagram. Total sample: n=114; χ2=551.8; CFI=0.805; NFI=0.648; p=0.0001; SRMR=0.0832; RMSEA=0.082.

(0.39MB).

The Cronbach's α coefficient for the KIDSCREEN-27 LSE version was 0.819, and did not increase with elimination of any of the items. This coefficient was higher than 0.75 in all dimensions. ICC for test-retest analysis of the different KIDSCREEN-27 dimensions in the LSE version ranged from 0.523 (Autonomy & Parents) to 0.795 (School Environment) (Table 3).

Table 4 show the mean scores of the KIDSCREEN-27 dimensions in the LSE version, by sex, and by level of hearing loss. Boys scored higher than girls in the Physical Well-Being dimension, and not significant differences by sex were found in the other dimensions. On the other hand, Peers and Social Support was the only dimension that showed differences by level of hearing loss.