Development Of A Short Version Of A Weight-Specific Quality Of Life Instrument(YOOL-W-SF)

Objectives To develop a short form of the Youth Quality of Life Instrument–Weight Module(YQOL-W),the YQOL-W-SF,with the aim of preserving the conceptual model and,as far as possible,the content and the psychometric properties of the original instrument.Methods Step 1-Shortening Phase An expert panel with 10 professionals and a youth group with 30 teenagers were organized to conduct qualitative research on the original instrument,using a consensus Delphi method and ranking method to collect relevant opinions about item screening;Based on the field survey data of the original instrument,a quantitative psychometric analysis was performed for each item using methods of item response theory(IRT)and classic test theory(CTT);Given the results of qualitative research and quantitative statistics,the final version of the YQOL-W-SF was determined,favoring content validity as the most important.Step 2-Validation Phase The psychometric properties of the YQOL-W SF were assessed in an independent population-based sample of 630 subjects with different weight states(normal weight,overweight,and obesity).Using the quota method,children of 11-18 years old from 9 schools in Hangzhou,Zhejiang Province of China in 2019 were recruited.All the participants were asked to fill the self-assessment questionnaires and then their height and weight were measured.The field questionnaires include YQOL-W-SF,Peds QL 4.0,and a demographic characteristics questionnaire.After the baseline survey was completed,a retest survey of 90 subjects was conducted within 7-10 days.Data entry was performed by Epi Data 3.1.Statistical analyses were performed using SPSS 21.0,Amos 21.0 and Rumm 2030.Results Item selection The shortened YQOL-W scale retained 8 items of the original instrument with at least two or more items per dimension.For the self dimension,the three remaining items were item 1 " Feel depressed",item 4 " Hide my body" and item 5 " Feel unattractive ";for the social dimension,the three items finally retained were item 8 " Avoid being noticed",item 9 " Worry what people saying" and item 11 " Feel like a loser";For the environmental dimension,Item 17 " Difficult wearing cloths" and the item 21 " Difficult getting a good grade in physical education class" were kept.Psychometric properties of the YQOL-W-SF 1.Validation of the factorial structure The results of CFA analysis showed that the factor loading value of the three-factor model was 0.66-0.91,and the RMSEA of the approximate error was 0.10,which was slightly higher than the ideal value,but the GFI(0.95),CFI(0.96)and Tucker-Lewis index TLI(0.94)were both greater than 0.9,indicating that the YQOL-W-SF still maintains the three-factor structure of the original scale model.2.Reliability The dimensions and overall Cronbach’s α of the YQOL-W-SF were 0.71-0.92,suggesting good internal consistency reliability of the short instrument.The ICCs of each dimension were 0.90-0.96,indicating acceptable test-retest reliability.3.Validity The content validity of the original YQOL-W had been evaluated in previous qualitative study.The wording of each item was not changed,therefore the concepts of the items to be tested were the same as those of the target population.The higher Pearson correlation coefficients of the comparable dimensions between YQOL-W-SF and the Peds QL 4.0 scale showed good convergent and discriminant validity of the short version.Known groups validity was demonstrated using T-test and one-way analysis of variance(ANOVA),as their results showed significant differences in subscale and total scores across age,gender and BMI groups.4.Rasch model analysis For the overall model fitting effect of the YQOL-W-SF,the self dimension and social dimension were evaluated as "reasonable",but the environmental dimension was evaluated as "too low".The targeting analysis showed that the average location in each dimension were greater than 0,which indicated different level of ceiling effect.All the item residuals were between-2.5 and +2.5 with no statistical significance,indicating good item fit to the model.Subjects had the most difficulty obtaining high quality of life scores on item 9 as it had the highest average threshold.Conversely,subjects were most likely to obtain high quality of life scores on item 5 as it had the lowest average threshold.In addition,there was an out-of-order threshold for each item of the YQOL-W-SF,suggesting that the option settings for the items need to be improved.Conclusions The 8-item YQOL-W-SF,which can be improved in terms of item option setting for better,has good psychometric properties and can be used for the measurement of weight related quality of life for children and adolescents in China.