Colombian Rasch validation of KIDSCREEN-27 quality of life questionnaire

Health and Quality of Life Outcomes

Research

Luz-Helena Lugo-Agudelo¹  Héctor-Iván García-García¹  Gilma-Norela Hernández-Herrera²  Claudia-Marcela Vélez³ 

[1] Academic Group of Clinical Epidemiology (GRAEPIC) and Group Health Rehabilitation, School of Medicine, University of Antioquia, Carrera 51 D # 62-29, Medellín, Colombia;Group of Clinical Epidemiology (GRAEPIC), School of Medicine, University of Antioquia, Medellín, Colombia;Group of Clinical Epidemiology (GRAEPIC), School of Medicine, University of Antioquia, Medellín, Colombia;Department of Clinical Epidemiology and Biostatistics, McMaster University, 1280, Main, Hamilton, Ontario, Canada;

关键词: Health-related quality of life;    KIDSCREEN;    Psychometric properties;    Rasch analysis;    Calidad de vida relacionada con la salud;    Kidscreen;    Psicometría;    análisis Rasch;   

DOI  :  10.1186/s12955-016-0472-0

 received in 2015-05-29, accepted in 2016-04-23,  发布年份 2016

来源: Springer

【 摘 要 】

BackgroundThe family of KIDSCREEN instruments is the only one with trans-cultural adaptation and validation in Colombia. These validations have been performed from the classical test theory approach, which has evidenced satisfactory psychometric properties. The aim of this study was to evaluate psychometric properties of KIDSCREEN-27 children and parent-proxy versions, through Rasch analysis.MethodsThe participants in the present study were two different sets of populations, 321 kids with a mean age of 12.3 (SD 2.6), 41 % 8 to 11 years old and 59 % 12 to 18 years old; and 1150 parent-proxy with an average age of 45.5 (SD 18.9). Psychometric properties were assessed using the partial credits model in the Rasch approach. Unidimensionality, fitting of person and item, response form, and differential item functioning (DIF) were measured.ResultsThe Infit MNSQ in child self-reported version that ranges between 0.71–1.76, and 0.69–1.31 in the parent-proxy version. Scores gathered on Likert forms of 5-response options, person separation was 2.08 for child self-reported version and 2.40 for parent-proxy; reliability was 0.81 and 0.85, respectively. Items reliability was 0.99 on both versions, with separations of 11.92 for child self-reported and 10.83 for parent-proxy. There was not DIF according to the variables sex and age but was present according to socioeconomic status.ConclusionThere was a good fit for items and individuals to the Rasch model. Item separation was adecuate, and person separation improved when the response form was re-codified to four options. The presence of DIF according to socioeconomic status implies a scale’s bias in the measure of HRQoL of Colombian children.

【 授权许可】

CC BY   
© Vélez et al. 2016