Youth Quality of Life Instrument-Research version (YQOL-R): psychometric properties in a community sample

Abstracts

OBJETIVO: Testar algumas propriedades psicométricas da versão brasileira do Youth Quality of Life Instrument-Research (YQOL-R) em uma amostra comunitária de adolescentes brasileiros. MÉTODOS: Este é um estudo transversal comunitário realizado em seis escolas localizadas na área de abrangência de uma unidade de saúde da família. De uma população original de 2.754 estudantes com idade entre 10 e 17 anos, selecionamos aleatoriamente 419 para responder à versão brasileira do YQOL-R. Testamos a confiabilidade, as diferenças já conhecidas entre os grupos (utilizando sintomas de ansiedade e envolvimento em episódios de bullying) e a estrutura fatorial. RESULTADOS: O YQOL-R apresentou boa consistência interna, e as diferenças já conhecidas entre os grupos mostraram-se adequadas e estiveram dentro do esperado tanto com relação ao bullying quanto à ansiedade. A estrutura fatorial do modelo conceitual foi em parte fundamentada por nossa análise. CONCLUSÕES: A versão brasileira do YQOL-R apresentou propriedades psicométricas suficientemente boas. São necessários mais estudos a fim de melhor investigar configurações alternativas da estrutura fatorial.

Qualidade de vida relacionada à saúde; resultados relatados pelo paciente; qualidade de vida; ansiedade; bullying; validade

ORIGINAL ARTICLE

Youth Quality of Life Instrument-Research version (YQOL-R): psychometric properties in a community sample

Giovanni Abrahão Salum^I; Donald L. Patrick^II; Luciano Rassier Isolan^III; Gisele Gus Manfro^IV; Marcelo P. de Almeida Fleck^V

^IMD. Anxiety disorders program for Child and Adolescent Psychiatry (PROTAIA), Hospital de Clínicas de Porto Alegre (HCPA), Universidade Federal do Rio Grande do Sul (UFRGS), Porto Alegre, RS, Brazil. National Science and Technology Institute for Child and Adolescent Psychiatry (INPD), Brazil. Post-graduate Program in Medical Sciences, Psychiatry, HCPA, UFRGS, Porto Alegre, RS, Brazil.

^IIPhD. Department of Health Services, University of Washington, Seattle, Washington, USA.

^IIIMD. PROTAIA, HCPA, UFRGS, Porto Alegre, RS, Brazil. Post-graduate Program in Medical Sciences, Psychiatry, HCPA, UFRGS, Porto Alegre, RS, Brazil.

^IVMD, PhD. PROTAIA, HCPA, UFRGS, Porto Alegre, RS, Brazil. INPD, Brazil. Post-graduate Program in Medical Sciences, Psychiatry, HCPA, UFRGS, Porto Alegre, RS, Brazil.

^VMD, PhD. Post-graduate Program in Medical Sciences, Psychiatry, HCPA, UFRGS, Porto Alegre, RS, Brazil.

ABSTRACT

OBJECTIVE: To test some psychometric properties of the Brazilian-Portuguese version of the Youth Quality of Life Instrument-Research (YQOL-R) in a community sample of Brazilian adolescents.

METHODS: This is a cross-sectional community study conducted in six schools of the catchment area of a family health unit. From an original population of 2,754 students from 10 to 17 years old, we randomly selected 419 to answer the Brazilian-Portuguese version of the YQOL-R. We tested reliability, known group differences (using anxiety symptoms and bullying involvement), and factor structure.

RESULTS: The YQOL-R showed a good internal consistency and had an adequate and expected known group differences with both bullying and anxiety. The factor structure of the conceptual model was partially supported by our analysis.

CONCLUSIONS: The Brazilian-Portuguese version of the YQOL-R showed sufficiently good psychometric properties. Further studies are needed in order to better investigate alternative configurations of the factor structure.

Keywords: Health-related quality of life, patient reported outcome, quality of life, anxiety, bullying, validity.

Introduction

The validation of quality of life (QoL) and health-related quality of life (HRQoL) measures is being considered a research priority in child and adolescent mental health.¹ The Youth Quality of Life Instrument – Research Version (YQOL-R)² is an instrument that takes into consideration what adolescents themselves say that is important to their QoL. The instrument development has involved youth-centered models of qualitative research based upon subjective self-report and which are developmentally appropriate, using grounded theory approach.² In addition, quantitative validity data have shown preliminary psychometric properties encouraging further research,³ and the instrument is already being used to measure HRQoL in pediatric samples.^4-6 In spite of that, there are no studies aiming to demonstrate its psychometric proprieties in Brazilian samples.

The main objective of this study was to test some psychometric properties of the Brazilian Portuguese version of the YQOL-R in a community sample of Brazilian adolescents.

Methods

This is a cross-sectional community study conducted in six public schools from the catchment area of the family health unit of Hospital de Clínicas de Porto Alegre, Brazil. We invited all 10- to 17-year-old students to participate in the study. The inclusion criterion involved being a student from one of the included schools regularly attending classes at that school in the period of the research. The original population of these schools comprised 2,754 adolescents; 217 (7.9%) were not approached by the research team to answer the questionnaires due to school dropout, transference to another school or missing classes in the days when questionnaire responses were collected and when the rescue visit was performed, and 80 (3.2%) refused participation and/or signed a dissent form, which was distributed for all parents and adolescents 2 weeks before the administration of the questionnaires. From the remaining 2,457 students, 419 were randomly selected to participate in this research, and there were no additional refuses. Scales were administered in the classroom, with careful supervision of the research team. Further details about sampling procedures can be found elsewhere.7 The study was approved by the Institutional Review Board and Ethics Committee of Hospital de Clínicas de Porto Alegre.

The YQOL-R is a self-rated 41-item scale developed to evaluate self-perceived QoL in adolescents and comprises four domains: self (14 items), relationships (14 items), environment (10 items), and general QoL (three items).^2,3 The response scale is an 11-point scale with anchors at zero ("Not at all") and 10 ("A great deal or completely"). Higher scores represent better QoL. All domains have shown good internal consistency (Cronbach's alpha 0.77-0.96) and test-retest reliability (intraclass correlation coefficients = 0.74-0.85).³ Discriminant validity with measures of depression, disability and attention deficit/hyperactivity disorder has shown to be adequate.⁵

The SCARED is a 41-item self-report measure of child and adolescent anxiety during the past 3 months. Items are scored using a three-point scale. Final scores ranges from 0 to 82, and higher scores reflect higher levels of anxiety. This scale has shown good psychometric properties.⁸

As in previous studies,⁹ the frequency of involvement in bullying and victimization during the last year was assessed after a previous definition of bullying,⁹ followed by examples of behaviors that are considered bullying and by the question: "Taking that definition and those examples into consideration, have you ever felt like you are being bullied at school"?

Sample size was defined using a subject-to-variable ratio of approximately 10:1.¹⁰ We used the Kaiser-Meyer-Olkin (KMO) to evaluate sampling adequacy and the Bartlett's test for sphericity to observe if the variables were highly correlated, in order to provide a reasonable basis for factor analysis. We tested internal consistency using the Cronbach's alpha, and the construct validity using known group validity across SCARED percentiles and frequencies of bullying involvement. Principal component analysis with varimax rotation was used for exploratory factor solution, using the scree plot to define graphically the number of factors. We also used confirmatory factor analysis (CFA) for the four-factor solution of the original conceptual scale, using chi-square goodness-of-fit, root mean square error of approximation (RMSEA) and one of the baseline fit measures, and the Tucker-Lewis Index (TLI)¹¹ as fit indices. A chi-square > 0.5, a RMSEA < 0.6, and a TLI > 0.8-0.9 were suggested as cutoffs for these indices, indicating good model fit. The analysis of variance (ANOVA) was performed to compare YQOL-R scores between quartiles of anxiety symptoms according to SCARED and categories of bullying involvement. All tests were two-tailed, with an alpha value of 0.05 and a 95% confidence interval.

Results

From the 419 students, 209 (49.9%) were female. The mean age was 13.9 years (standard deviation = 2.45). Factor analysis with non-missing questions was available for 412 subjects. Frequencies and ranges indicated that all response choices were used and followed a normal distribution. Floor/ceiling effects were lesser than 50% for all items.

KMO and Bartlett's tests satisfied the assumption for factor analysis. Cronbach's alpha for the 41 items was 0.931 (0.803 for the self domain, 0.885 for the relationship domain, 0.851 for the environmental domain, and 0.779 for the general QoL domain of the YQOL-R scale). Principal component analysis with varimax rotation was conducted to assess the underlying structure of the 41 items of the YQOL-R. Four factors were requested, based on graphical analysis of the scree plot. Results of the factor loadings by item and variance explanation are depicted in Table 1.

CFA of the original scale reveals mixed results: whereas chi-square test indicated lack of fit (chi-square = 1,834.8, degrees of freedom = 768; p < 0.001), other indices seem to show an acceptable model (RMSEA = 0.058, TLI = 0.834).

In addition, a single factor principal component analysis was run on the four domain scores in order to assess whether the data supported the use of an overall score. The results of this analysis showed that a single factor explained 74.8% of the total variation in the domain scores with an eigenvalue of 2.99, hence supporting the use of an overall YQOL-R score.

Comparison between known groups of anxiety symptoms percentiles and bullying behaviors can be observed in Table 2. We can see a linear dose-response effect between both anxious symptoms and bullying involvement and YQOL-R scores in all domains evaluated.

Discussion

In the present study, we were able to investigate some psychometric properties of the YQOL-R in a community sample of 10- to 17-year-old adolescents. The YQOL-R has shown a good internal consistency and a linear dose-response effect with known group differences (for both anxiety and bullying). The factor structure has shown mixed results, and further studies with larger samples are needed.

There is a trend towards greater consideration of patient reported outcomes, including HRQoL. QoL measures can be used for several reasons: service planning, estimating cost-effectiveness of treatments, and measuring how well a healthcare system is meeting the health needs of a given population. They can also provide a better picture for service managers about the differences between patient groups, resulting in evidence-based resource allocation, and, for some researchers, are considered the gold standard of clinical decision making.¹

Several instruments are being constructed to measure QoL in children and adolescents. A recent study has found at least 14 measures of general QoL for children and adolescents,¹² but very few are fully validated in different cultures, and there are few validated scales to measure the QoL of children and adolescents in Brazil.^13-15 We decided to use the YQOL-R and to validate it for Brazilian-Portuguese because of its well-conducted process of conceptual development² and quantitative research showing good psychometric properties.³

In our study, factor structure achieves only partial support for the original conceptual four-factor solution (environmental, self, relationship and general QoL domains). Both exploratory and confirmatory analysis demonstrated mixed results. Our exploratory analysis has shown that items related to friendship have higher loadings in the environmental domain rather than in the relationship domain. This misfit is reasonable for schoolchildren, since the environmental domain has a lot of items related to school environment. In addition, as expected, the general QoL domain did not load into a separate domain.

Regarding CFA, the lack of fit of the chi-square test may be discounted, since a significant chi-square itself is not a reason to modify the model.¹¹ Therefore, further research is needed to confirm factor structure and to investigate the use of CFA alternative proposed factors.

Our study has some limitations. First, we have assessed some psychometric properties of YQOL-R, but not all of them. Second, the lack of an independent sample did not allow us to evaluate a CFA model for our findings from the exploratory analysis. Finally, our age range is large compared to our sample size and therefore we were not able to evaluate if psychometric properties are stable over different age ranges.

The concept and measures of QoL in adolescents still needs further research. Studies investigating psychometric properties are needed in order to provide valid and reliable instruments to better integrate QoL data into clinical research. The YQOL-R has shown sufficiently good psychometric properties to encourage further research investigations and clinical use.

Acknowledgements

Giovanni Abrahão Salum Junior receives a CAPES doctoral scholarship. Luciano Isolan receives a CNPq doctoral scholarship. Gisele Gus Manfro and Marcelo Fleck receive a senior CNPq researcher scholarship. This paper was partially funded by the Brazilian government agencies FIPE-HCPA (08-017), FAPERGS (PRONEX 2009 – FAPERGS/CNPq 10/0018.3) and CNPq (483032/2007-7).

References

• 1. Coghill D, Danckaerts M, Sonuga-Barke E, Sergeant J; ADHD European Guidelines Group. Practitioner review: quality of life in child mental health conceptual challenges and practical choices. J Child Psychol Psychiatry. 2009;50:544-61.
• 2. Edwards TC, Huebner CE, Connell FA, Patrick DL. Adolescent quality of life, part I: conceptual and measurement model. J Adolesc. 2002;25:275-86.
• 3. Patrick DL, Edwards TC, Topolski TD. Adolescent quality of life, part II: initial validation of a new instrument. J Adolesc. 2002;25:287-300.
• 4. Findling RL, Childress AC, Cutler AJ, Gasior M, Hamdani M, Ferreira-Cornwell MC, et al. Efficacy and safety of lisdexamfetamine dimesylate in adolescents with attention-deficit/hyperactivity disorder. J Am Acad Child Adolesc Psychiatry. 2011;50:395-405.
• 5. Topolski TD, Edwards TC, Patrick DL, Varley P, Way ME, Buesching DP. Quality of life of adolescent males with attention-deficit hyperactivity disorder. J Atten Disord. 2004;7:163-73.
• 6. Isolan L, Pheula G, Salum GA Jr, Oswald S, Rohde LA, Manfro GG. An open-label trial of escitalopram in children and adolescents with social anxiety disorder. J Child Adolesc Psychopharmacol. 2007;17:751-60.
• 7. Salum GA, Isolan LR, Bosa VL, Tocchetto AG, Teche SP, Schuch I, et al. The multidimensional evaluation and treatment of anxiety in children and adolescents: rationale, design, methods and preliminary findings. Rev Bras Psiquiatr. 2011;33:181-95.
• 8. Isolan L, Salum GA, Osowski AT, Amaro E, Manfro GG. Psychometric properties of the Screen for Child Anxiety Related Emotional Disorders (SCARED) in Brazilian children and adolescents. J Anxiety Disord. 2011;25:741-8.
• 9. Nansel TR, Overpeck M, Pilla RS, Ruan WJ, Simons-Morton B, Scheidt P. Bullying behaviors among US youth: prevalence and association with psychosocial adjustment. JAMA. 2001;285:2094-100.
• 10. Bryant FB, Yarnold PR. Principal components analysis and exploratory and confirmatory factor analysis. In: Grimm LG, Yarnold PR, eds. Reading and understanding multivariate analysis. Washington, DC: American Psychological Association Books; 1995. p. 99-136.
• 11. Garson GD. Structural Equation Modeling. In: Statnotes: Topics in Multivariate Analysis; 2009. http://faculty.chass.ncsu.edu/garson/pa765/statnote.htm Acesso: 22/03/2012.
• 12. Janssens L, Gorter JW, Ketelaar M, Kramer WL, Holtslag HR. Health-related quality-of-life measures for long-term follow-up in children after major trauma. Qual Life Res. 2008;17:701-13.
• 13. Assumpção FB Jr, Kuczynski E, Sprovieri MH, Aranha EM. Quality of life evaluation scale (AUQEI - Autoquestionnaire Qualité de Vie Enfant Imagé). Validity and reliability of a quality of life scale for children 4 to 12 years-old. Arq Neuropsiquiatr. 2000;58:119-27.
• 14. de Souza FM, Molina J, Terreri MT, Hilário MO, Len CA. Reliability of the Pediatric Quality of Life Inventory Healthcare Satisfaction Generic Module 3.0 version for the assessment of the quality of care of children with chronic diseases. J Pediatr (Rio J). 2012;88:54-60.
• 15. Klatchoian DA, Len CA, Terreri MT, Silva M, Itamoto C, Ciconelli RM, et al. Quality of life of children and adolescents from São Paulo: reliability and validity of the Brazilian version of the Pediatric Quality of Life Inventory version 4.0 Generic Core Scales. J Pediatr (Rio J). 2008;84:308-15.

Correspondence

Publication Dates

• Publication in this collection
  08 Nov 2012
• Date of issue
  Oct 2012

History

• Received
  16 Jan 2012
• Accepted
  14 Mar 2012