Adolescent proxies’ health-related quality of life agreement using the kidscreen-52 questionnaire
The purpose of this study was to examine (a) the level of agreement and the magnitude of discrepancies between the adolescents and their parents; and (b) whether these levels of agreement/discrepancy depend on the dimension of quality of life assessed, the level of HRoL reported, the age and the gender of youth, and the country of living in Latin American.
Both versions of the KIDSCREEN-52 questionnaire were completed by 857 youths aged 12 to 17 years and their parents in three countries (Argentina, Brazil, and Chile). With the use of intraclass correlation coefficient (ICC) and paired t-test the agreement/discrepancy for each dimension of the KIDSCREEN-52 was reported. The results indicate that the agreement is higher for the girls than for the boys and for the adolescents older for all of the 10 dimensions. The level of agreement between adolescents and their proxies varied in the three countries. The largest mean difference was found for Chile.
Three dimensions were significantly overestimated by the parents of adolescents of the whole sample: Autonomy, Parents Relations and Home Life and School Environment. Moods and Emotions, Social Support and Peers, Social Acceptance/Bullying and Financial Resources were the dimensions that mostly underestimated by the parents of adolescents in whole sample.
Only recently have health professionals focused on the important of quality-of-life assessment in young people. Health-related quality of life (HRQoL) measures can be useful in identifying subgroups of children and adolescents who are at risk for health problems, and can assist in determining the burden of a particular disease or disability (Eiser and Morse, 2001). The Centers for Disease Control and Prevention explicitly recommends the identification of subgroups of children and adolescents who are at risk for health problems with suitable HRQOL measures (CDC, 2000).
HRQoL as a multidimensional construct can be defined as an individual’s satisfaction or happiness in various life domains that affect or are affected by health (Evans, 1994). In this particular, the measurement of HRQoL in child and adolescent populations was for a long time under-investigated in comparison with adults. However, to detect impairments of well-being and functioning, it is essential to assess how children and adolescents perceive their own situation. Self-reported questionnaires are regarded as the primary method for assessing quality-of-life.
Consequently, a substantial number of generic and disease-specific self-report measures to detect HRQoL for younger respondents exist these days, and they are still rapidly growing in number (Pane et al, 2006). However, the European KIDSCREEN group was the first research group to develop a cross-cultural, standardized HRQoL-screening instrument for children and adolescents in general population (Ravens-Sieberer et al, 2003). Moreover, recently a number of parent-proxy questionnaires have been developed to measure the young’ HRQoL from both point of view (Theunissen et al, 1998; Jokovic et al, 2004). Parent-proxy ratings should be considered carefully as a potential substitute for self-report ratings (Jokovic et al, 2004), but also should be regarded providing complementary information about youth’s HRQoL.
The parent-proxy effect is a hypothesized bias in the measure of self-perceived HRQoL, but there is still a lack of knowledge about its directionality and magnitude with age, gender, country in young population. Nevertheless, the usefulness of parent-proxy measures has been shown when assessing HRQoL of children too immature or presenting cognitive deficiencies (Lowford et al, 2001). To better understand the parent-proxy effect is need providing both young’ self-reports and parent-proxy ratings of the young’ HRQoL using the corresponding versions (selfreported and parent-proxy version) of the same HRQoL questionnaire.
The KIDSCREEN project conducted the development of the first HRQoL instrument that have parallel parent-proxy and child self-report versions crossculturally validated for young population. Thus, researchers have commenced examining the degree of concordance between the two scores in different realities. To better understand the parent-proxy effect, this study examined: (a) the level of agreement and the magnitude of discrepancies between adolescents and theirs parents; and (b) whether these levels of agreement/discrepancy depend on the dimension of quality of life assessed, the level of HRQoL reported, the age and the gender of adolescent, and the country of living in Latin American.
Participants and sampling
Both versions of the questionnaire were completed by 857 youths aged 12 to 17 years (46.4% boys) and their parents in three countries Latin American: Argentina (18.0%), Brazil (30.5%), and Chile (51.5%). Proxies were 705 mothers (82.3%) and 152 fathers (17.7). The study used a cross-sectional design. Different approaches for sample selection were used across countries in order to obtain random samples. Fieldwork was carried out between May and September 2008.
The adolescents and theirs parents were told that the purpose of the study was to attain knowledge about the quality of life and health, in general, among teenagers. They were further informed that their responses would be treated anonymously and that there was no right or wrong answer. The scale instructions were given in written form, and the adolescents filled in the questionnaire at school and parents were asked to complete the 439 questionnaire at home.
The questionnaires were performed lasted 30 to 40 minutes. The rights of all participants were safeguarded through informed consent and confidentiality. Inclusion criteria for the adolescents were to be between 12 and 17 years old, to be able to read and complete the questionnaire themselves, and to consent to be involved in the study. Inclusion criteria for the parent-proxies were to live with the adolescent. Only one parent-proxy was involved for each adolescent included in study. The adolescent and the parent-proxy completed the questionnaire sequentially.
HRQoL was measured using the KIDSCREEN-52 questionnaire, a self-report measure applicable to healthy and chronically ill children and adolescents 8-18 years old. Various psychometric aspects of the KIDSCREEN-52 had been studied in European (Ravens-Sieberer et al., 2005; Robitail et al, 2006) and Latin American countries (Berra et al, 2009; Guedes and Guedes, 2010).
The KIDSCREEN-52 questionnaire measures HRQoL in ten dimensions: Physical Well-being (PHY, five items); Psychological Well-being (PWB, six items); Moods and Emotions (MOOD, seven items), Self Perceptions (SELF, five items), Autonomy (AUTO, five items), Parents Relations and Home Life (PAR, six items), Social Support and Peers (SOC, six items), School Environment (SCHO, six items), Social Acceptance (BULL, three items), Financial Resources (PFIN, three items). In table 1 all dimensions are characterized more in details.
Tabla 1. Adolescent proxies’ health-related quality of life agreement using the kidscreen-52 questionnaire
The KIDSCREEN-52 item uses five-point Likert-type scales to assess either the frequency (never, seldom, sometimes, often, always) or intensity (not at all, slightly, moderately, very, extremely) and the recall period is one week. The score of each dimension was then transformed linearly to a 0-100 point scale, with 100 indicating the best quality of life and 0 the worst (Ravens-Sieberer et al., 2003).
The proxy questionnaire was designed to be completed by the parents. The questions asked to the adolescents were reworded to be completed by the parents. The parents were not asked what they thought how their child was, but they should rate in their child’s place to answer the question how their child thought and felt.
The statistical analyses were performed using SPSS for Windows (Version 17.0; SPSS, Chicago, IL, USA). The results regarding HRQoL scores were first compared to a normal distribution by the Kolmogorov-Smirnov distance test. Since the data showed a normal distribution, parametric tests were used and the mean and standard deviation were calculated. The level of agreement and magnitude of discrepancies were assessed for the whole sample, by age group, gender, and country. To assess and compare agreement between adolescent and proxy report, intraclass correlation coefficient (ICC) was computed as recommended. To evaluate the magnitude of discrepancies between adolescent and proxy ratings, paired t-test of the mean differences were computed as usually recommended.
RESULTSBaseline scores of youths and proxies on the whole sample were presented in table 2. The results indicate that there was variation between countries in the HRQoL dimensions. Generally the Brazilian youths and their parents had higher mean scores.
Tabla 2. Adolescent proxies’ health-related quality of life agreement using the kidscreen-52 questionnaire
With regard to the level of agreement between adolescent and proxy report, the results indicate that the agreement is higher for the girls than for the boys and for the adolescents older for all of the 10 dimensions – table 3. For the whole sample, a moderate agreement (ICC > 0.40) was reported in only three out of 10
KIDSCREEN-52 dimensions: Psychological Well-being (ICC = 0.68), Moods and Emotions (ICC = 0.47), and Self Perceptions (ICC = 0.78). In the girls’ and boys’ samples, a moderate agreement between self-reports and proxies reports was found for the dimensions Psychological Well-being (ICC = 0.79 and ICC = 0.58, respectively) and Self Perceptions (ICC = 0.61 and ICC = 0.56, respectively). Further, an ICC > 0.40 is achieved for Physical Well-being (ICC = 0.53) for the girls’ sample and Moods and Emotions (ICC = 0.77) for the boys’ sample.
Tabla 3. Adolescent proxies’ health-related quality of life agreement using the kidscreen-52 questionnaire
With regard to the effect of country on the agreement between adolescents and their proxies, the level of agreement varied. Overall, the highest levels of agreement were found in the Brazil sample, where it was at least moderate for three dimensions: Psychological Well-being (ICC = 0.49), Social Support and Peers (ICC = 0.52), and Financial Resources (ICC = 0.55). No significant agreement was found in the Chile sample whatever the dimension considered.
As for the magnitude of discrepancies, the largest mean difference between proxy and youth score are found for Chile concerning Autonomy (t = 8.41; p0.001). Three dimensions were significantly overestimated by the parents of adolescents of the whole sample: Autonomy (t = 9.98; p0.001), Parents Rlations and Home Life (t = 8.44; p 0,001), and School Environment (t = 8.51; p0.001).Moods and Emotions (t =-2.25; p=0.05), Social Support and Peers (t = -3.71; p0.001), Social Acceptance/Bullying (t = -4.43; p0.001), and Financial Resources (t= -5.14; p0001) were the dimensions that mostly underetimated by the parents of adolescents in whole sample.
According to ICC results our study reported a higher agreement in girls for all the KIDSCREEN dimensions. Concordance of self-assessed and proxy-reported HRQoL might be expected to increase with the youth’s age, as a reflection of the youth’s increasing ability to describe his or her experiences and emotions to the parent/proxy (Eiser and Morse, 2001).
Irrespective of youth’s gender and country, the School Environment HRQoL was overestimated by the proxies. For this dimension, youth-proxies agreement has been previously described with higher level for sick youth (Eiser et al, 1999). This is supported by the finding that a chronic condition of a youth has a significant impact on the level of agreement between the child and proxy measure. On the other hand, results of other study reported poor agreement for cognitive functioning (Vogels et al, 1998). These results highlighted that the proxy bias is not clearly explored and understood.
For the dimensions Self-Perception and Psychological Well-being, a high level of agreement between the adolescent and proxy report was found with low magnitude of discrepancies. Concerning a sample of children suffering from cancer, the dimension Self-Esteem presented a satisfactory agreement with proxy report made at home, and poor agreement with reports at clinic (Glaser et al, 1997). This is supported by the finding of this study, chronic condition had a significant impact on the youth/proxy agreement.
For the dimensions Autonomy, Parents Relations and Home Life, Social Support and Peers, Social Acceptance/ Bullying and Financial Resources, a poor agreement and a high magnitude of discrepancies between the self-assessed and proxy-report were found. Studies available in literature also indicate that generally proxies over-or-underestimated the levels of HRQoL reported by their children (Davis et al, 2007; Robitail et al, 2007; Theunissen et al, 1998). However, in the dimension Social Acceptance/Bullying a satisfactory agreement was show for ill children (Eiser et al, 1999).
Our results showed a lower score for the proxies in five out of 10 KIDSCREEN-52 dimensions (Physical Well-being, Moods and Emotions, Social Support and Peers, Social Acceptance/Bullying, Financial Resources). In a study of samples from European countries was observed lower score for the proxies only for three dimensions: Psychological Well-being, Social Support and Peers and Financial Resources (Robitail et al., 2007). Still, Theunissen et al (1998) reported that children scores were lower than proxy score for five out of the seven dimensions using the TACQoL questionnaire (physical complaints, motor functioning, autonomy, cognitive functioning and positive emotions).
Considering two others studies (Graham et al., 1997; Bruil, 1999), results were not consistent with each other, no conclusions should be drawn given the question of under- or overestimation by the proxies. For the 10 KIDSCREEN-52 dimensions the mean difference between proxy and youth score decreases as the HRQoL level increases. The greatest discrepancies have been found for intermediate level of HRQoL. The range of parent scores is narrower than those of adolescents. For a low youth report, proxy ratings were higher.
Conversely, for a high youth report of children’s HRQol, proxy report was lower. The level of agreement was depending on the level of HRQoL report. These results are consistent with those reported by Theunissen et al. (1998) and Sneuuw et al. (1998). The results showed that level of HRQoL report of proxy varies according to country. For example, the proxies overestimated significantly the level of their child’s HRQOL in the Moods and Emotions, Self Perceptions and Social Acceptance/Bullying in the Brazil, but underestimated it in Argentina and Chile.
In the present study, the country impact on the level of agreement and magnitude of discrepancy should be divided in two parts: a study design impact and a cultural impact. The study design impact should rely on cultural discrepancies or by the nonrepresentativity of the different national samples used. Nevertheless, we cannot exclude a potential cultural impact. This study is one of the few that assess cultural differences on a large sample and the first one that include different Latin American countries.
This is the first study in Latin American countries with the purpose of examining the concordance between parent proxy reports and child self-reports of the level of HRQoL using the KIDSCREEN-52 questionnaire. Physical Well-being and Self Perceptions dimensions showed the highest level of agreement and minor magnitude of discrepancies between youth and proxy measure. Autonomy and Social Support and Peers dimensions presented the main discrepancies. Gender and country had a significant and consistent impact on the level of agreement and magnitude of discrepancies.This impact depended on the KIDSCREEN dimensions studied and of the level of HRQoL reported. Therefore, our results suggested to take into account country of living, gender, and age when comparing children and proxy scores.
Berra S., Bustingorry V., Henze C., Díaz M.P., Rajmil L., Butinof M. (2009): Adaptación transcultural del cuestionario KIDSCREEN para medir calidad de vida relacionada con la salud en población argentina de 8 a 18 años. Archives Argentino de Pediatria, 107(4):307-314.
Bruil J. (1999): Development of a quality of life instrument for children with chronic illness. Leiden: Health Psychology, Leiden University.
Centers for Disease Control and Prevention. (2000): Measuring healthy days: Population assessment of healthrelated quality of life. Atlanta: U.S. Department of Health and Human Services. Davis E., Nicolas C., Waters E., Cook K., Gibbs L., Gosh A. et al. (2007): Parent-proxy and child self-reported health-related quality of life: using qualitative methods to explain the discordance. Quality of Life Research, 16:863-871.
Eiser C., Cotter I., Oades P., Seamark D., Smith R. (1999): Health-related quality of life measures for children. International Journal of Cancer, 12(Suppl):87-90.
Eiser C. & Morse R. (2001): Quality of life measures in chronic diseases of childhood. Health Technology Assessment, 5:1-157.
Evans D.R. (1994): Enhancing quality of life in the population at large. Social Indicators Research, 33(1):47-88. Glasser A.W., Davies K., Walker D., Brazier D. (1997): Influence of proxy respondents and mode of administration on health status assessment following central nervous system tumors in childhood. Quality of Life Research, 6(1):43-53.
Graham P., Stevenson J., Flynn D. (1997): A new measure of health-related quality of life for children: preliminary findings. Psychology and Health, 12:655-665.
Guedes D.P. & Guedes J.E.R.P. (2010): Tradução, adaptação transcultural e propriedades psicométricas do KIDSCREEN-52 para população brasileira. Revista Paulista de Pediatria. No prelo.
Jokovic A., Locker D., Guyatt G. (2004): How well do parents know their children? Implications for proxy reporting of child health-related quality of life. Quality of Life Research, 13(7):1297-1307.
Lowford J., Volavka N., Eiser C. (2001): A generic measure of quality of life for children aged 3-8 years: results of two preliminary studies. Pediatric Rehabilitation, 4(4):197-207.
Pane S., Solans M., Gaife L., Serra-Sutton V., Estrada M.D., Rajnil L. (2006). Instrumentos de calidad de vida relacionada con la en la edad pediátrica. Revisión sistemática de la literatura: actualización. Barcelona: Agencia de Evaluación de Tecnología E investigación Médicas.
Ravens-Sieberer U., Erhart M., Power M., Auquier P., Cloetta B., Hagquist C. (2003): Item-response-theory analyses of child and adolescent self-report quality of life data: the European cross-cultural research instrument KIDSCREEN. Quality of Life Research, 12:722.
Ravens-Sieberer U., Gosch A., Rajmil L., Erhart M., Bruil J., Duer W. et al. (2005): KIDSCREEN-52 quality of life measure for children and adolescents. Expert Review Phamacoeconomics & Outcomes Research, 5:353-364.
Robitail S., Simeoni M.C., Erhart M., Ravens-Sieberer U., Bruil J., Auquier P. et al. (2006): Validation of the European Proxy KIDSCREEN-52 Pilot Test Health-Related Quality of Life Questionnaire: First results. Journal of Adolescent Health, 39:596.e1-596.e10.
Robitail S., Siméoni M-C., Ravens-Sieberer U., Bruil J., Auquier O., KIDSCREEN Group. (2007): Children proxies’ quality-of-life agreement depended on the country using the European KIDSCREEN-52 questionnaire. Journal of Clinical Epidemiology, 60:469-478.
Sneuuw K.C., Aaronson N.K., Sprangers M.A., Detmar S., Wever L.D., Shornagel J.H. Comparison of patient and proxy EORTC QLQ-C30 ratings in assessing the quality of life of cancer patients. Journal of Clinical Epidemiologic, 51:617-631.
Theunissen N.C., Vogels T., Koopman H.M., Verrips G.H., Zwinderman K.A., Verloove-Vanhorick S.P. et al. (1998): The proxy problem: child report versus parent report in health-related quality of life research. Quality of Life Research, 7:387-397.
Vogels T., Verrips G.H., Verloove-Vanhorick S.P., Fekkes M., Kamphuis R.P., Koopman H.H. et al. (1998): Measuring health-related quality of life in children: the development of the TACQOL parent form. Quality of Life Research, 7:457-465.