Factor Structure of a Spanish Translation of an Obesity-Specific Parent-Report Measure of Health-Related Quality of Life

Factor Structure of a Spanish Translation of an Obesity-Specific Parent-Report Measure of... Abstract Objective Latino youth are disproportionately affected by pediatric obesity and consequently experience impaired health-related quality of life (HRQOL). Although many caregivers of Latino youth do not speak English fluently, no validated Spanish translations of obesity-specific HRQOL measures exist for this population. Therefore, non-English-speaking Latino parents have typically been excluded from analyses related to HRQOL. This study assesses the factor structure of a Spanish translation of a parent-report measure of obesity-specific HRQOL, Sizing Them Up, in a treatment-seeking sample of children with obesity. Methods Structural equation modeling was used to assess the factor structure of the 6-subscale, 22-item Sizing Them Up measure in 154 parents of treatment-seeking Latino youth (5–18 years of age). Analyses exploring internal consistency and convergent validity were also conducted. Results Acceptable measurement fit was achieved for the six-factor solution. However, the higher-order model assessing Total HRQOL did not reach acceptable levels, as results found that the Positive Social Attributes (PSA) subscale was not representative of Total HRQOL; internal consistency and convergent validity results also supported this finding. Conclusions The current study provides support for the utility of a modified version of Sizing Them Up, excluding the PSA Scale, as a parent-report measure of obesity-specific HRQOL in treatment-seeking Latino youth with obesity. disparities, obesity, quality of life Childhood obesity remains a major public health concern in the United States, with approximately 17% of children at or above the 95th body mass index (BMI) percentile for age and sex (Ogden, Carroll, Kit, & Flegal, 2014). Although the obesity epidemic affects youth of all ethnic groups, the prevalence of obesity is disproportionately higher among Latino children (21.9% among Latino children compared with 14.7% among non-Latino, White Children; Ogden, Carroll, Fryar, & Flegal, 2015). Early childhood risk factors for this ethnic group include greater odds of rapid weight gain in infancy, elevated risk of maternal restrictive feeding practices, higher intake of sugar-sweetened beverages, and higher intake of fast food during early childhood (Taveras, Gillman, Kleinman, Rich-Edwards, & Rifas-Shiman, 2010). The high prevalence is concerning because childhood obesity is associated with a number of negative physical and psychosocial consequences, including cardiovascular disease, musculoskeletal problems, and behavioral and emotional difficulties throughout childhood and into adulthood (Kalarchian & Marcus, 2012; Reilly & Kelly, 2011; Sanchez-Villegas et al., 2013; Taylor et al., 2006). Furthermore, the association of childhood obesity with type 2 diabetes and fatty liver is disproportionately observed in Latino youth compared with non-Latino White youth (Dabelea et al., 2014; Schwimmer et al., 2006). Although BMI percentile provides some information about a child’s current health status and risk for illness, it does not provide any information about the functional impact of obesity on a child’s life (Modi & Zeller, 2008); therefore, a focus of recent research has been placed on measuring health-related quality of life (HRQOL) in youth with obesity (Zeller & Modi, 2009). HRQOL is a multidimensional construct that assesses a person’s perception of his or her physical, emotional, and psychosocial functioning in relation to his or her health and/or medical condition (Kuyken, 1995). Current research on HRQOL in youth with obesity shows that higher BMI percentile or BMI z-score (zBMI) is associated with lower HRQOL (Buttitta et al., 2014; Ul-Haq et al., 2013) and that HRQOL for children with obesity is similar to children with other chronic diseases (e.g., cancer; Schwimmer, Burwinkle, & Varni, 2003). Examining HRQOL in Latino youth with obesity is important to establish a complete picture of the direct impact of weight on social, emotional, physical, and academic functioning. In the past several decades, there has been a proliferation of various forms and versions of HRQOL measures (Quittner, Cruz, Modi, & Marciel, 2009). HRQOL measures vary based on the reporter (i.e., self-report and parent-proxy report), as well as whether they are more general or illness specific. When measuring HRQOL in children with obesity, parent-proxy reports are important in conjunction with self-report measures because parents’ perceptions of child functioning influence their motivation for initiating services for their children (Modi & Zeller, 2008). Further, including parent proxy measures gives clinicians a more comprehensive picture of the child’s functioning, given that parents are often more accurate observers and reporters of their child’s behavior than children are of their own behavior, and some children and adolescents may be reticent to report issues such as peer victimization related to their size. Specific versions of HRQOL measures (e.g., PedsQL Diabetes Module; Varni et al., 2013) make it possible to understand functioning related specifically to the illness and to measure change because of illness-specific interventions (Modi & Zeller, 2008). When examining physical, emotional, and psychosocial functioning in children with obesity, obesity-specific HRQOL measures can help determine the burden of obesity, responsiveness to change related to weight loss, and progress toward public health goals (CDC, 2011). An example of an obesity-specific parent-proxy HRQOL measure is Sizing Them Up (Modi & Zeller, 2008), which is a measure that has been validated in parents of youth with obesity of age between 5 and 18 years. Sizing Them Up measures parents’ perceptions about their child’s functioning in the context of the child’s weight/size/shape in a variety of domains. Using an exploratory factor analysis (EFA) approach, Modi and Zeller (2008) found that the model that best fit the measure was a model with six subscales (i.e., emotional, physical, teasing/marginalization, positive social attributes (PSA), mealtime challenges, and school functioning). The results of the initial validation showed that the measure had adequate internal consistency (coefficients ranged from 0.52 to 0.90), test–retest reliability (coefficients ranged 0.57 to 0.80), construct validity (the Physical Functioning scale was negatively associated with zBMI), and convergent validity with other HRQOL measures. Additionally, no gender differences were shown, but age was significantly correlated with the scales of Sizing Them Up. The majority of scales were negatively correlated with age, except for the Mealtime Challenges subscale, which was positively correlated with age (Modi & Zeller, 2008). Since its development, Sizing Them Up has been used in a limited number of studies with primarily Caucasian and African-American samples (Black et al., 2014; e.g., Guilfoyle, Zeller, & Modi, 2010; Zeller et al., 2015). Although the prevalence of obesity is disproportionately higher among Latino children, they may be underrepresented in HRQOL research because of the lack of availability of a validated obesity-specific parent-report HRQOL measure in Spanish. Many Latino parents do not speak English fluently. In 2011, 20.84% of individuals in the United States reported speaking a language other than English in the home; 62.22% of these individuals reported speaking Spanish (Ryan, 2013). These non-English-speaking Latino parents have been excluded from analyses in previous research, leaving a gap in knowledge about parent perception of HRQOL in a high-need population (Black et al., 2014). Furthermore, Modi and Zeller (2008) called for future research to examine the measure in ethnically diverse samples of youth with obesity, particularly Latino youth; however, to date, there are no studies identified that have validated a Spanish version of Sizing Them Up. Recently, Tripicchio and colleagues (2017) provided a psychometric evaluation of Sizing Me Up (Zeller & Modi, 2009), the self-report counterpart of Sizing Them Up, among children who are Latino. Results of this confirmatory factor analysis (CFA) demonstrated that Sizing Me Up achieved acceptable fit in a Latino population of treatment-seeking youth with obesity, providing evidence that the English version is appropriate for use with Latino children. However, the psychometric properties of Sizing Them Up among those adults who require written materials in Spanish remain unknown. To fill the gap in the literature on obesity-specific HRQOL measures, the primary aim of the current study was to examine the psychometric properties of the Spanish translation of Sizing Them Up. It was hypothesized that the previous factor structure as identified by Modi and Zeller (2008) would be replicated with the translated measure in this Spanish-speaking sample of Latino parents. Specifically, it was hypothesized that the Spanish translation of the Sizing Them Up measure would retain the six-factor structure. Further, it was hypothesized that the intercorrelations would be positive and range from 0.08 to 0.68, similar to the results found by Modi and Zeller (2008). Additionally, based on the previous study by Modi and Zeller (2008), it was also hypothesized that Cronbach’s α, a measure of internal consistency, for the Total HRQOL scale would be >0.8, but that Cronbach’s α for the subscales would range from 0.5 to 0.9. It was also hypothesized that the Physical Functioning subscale, but no other subscales, of Sizing Them Up would be negatively correlated with zBMI. Regarding child age, it was hypothesized that there would be a negative correlation between all of the scales of Sizing Them Up except for the Mealtime Challenges subscale, which was expected to be positively correlated with age. Finally, consistent with the results from Modi and Zeller (2008), it was expected that gender would not be associated with Sizing Them Up subscales. Methods Participants and Study Procedures Participants were 154 Spanish-speaking parents of children presenting at one of four comprehensive behavioral lifestyle interventions for pediatric obesity across two sites in the Midwest (greater Kansas City area; three programs) and South (Texas; one program). Participants across all sites presented for weight management treatment, including weight management clinics and family-based behavioral group treatments, and all participant data were collected at entry into the program. The current project was approved as a separate multisite IRB project, and all study procedures were approved by the local IRBs, though the data collection process varied based on the site/program. Two programs analyzed clinical data abstracted under an IRB-approved retrospective chart review (N = 72), one program collected the measure through an opt-out consent information sheet procedure (N = 43), and the final program did so through a traditional opt-in consent/assent procedure (N = 39). The interventions served a high percentage of families who self-identified as Latino (with sites ranging from 28.5 to 67%); therefore, parents were offered the opportunity to participate in treatment appointments/sessions and complete treatment measures in either English or Spanish. Parents who self-selected to complete the measure in Spanish were included in the current study. Country of origin and acculturation status of participating families was not assessed; however, providers from each program provided anecdotal evidence that the vast majority of families were of Mexican descent. Additionally, inclusion criteria included parents of children of age 5–18 years with a BMI equal to or above the 95th percentile to replicate the original sample from the initial validation study (Modi & Zeller, 2008). Parents completed the Spanish version of Sizing Them Up at an initial treatment session. Child height and weight were obtained according to standard clinical procedures by trained clinical staff. Spanish Translation Procedures The procedures for translation were aligned with guidelines put forth by the World Health Organization (e.g., forward translation, back translation, pretesting, cognitive interviewing, and final version; World Health Organization, 2016). Sizing Them Up was translated into Spanish by two staff members who were certified as qualified bilingual staff. Both staff members were born in Mexico and are native Spanish speakers. One staff member translated all items independently (forward translation), and the second member back-translated the items (back translation); they worked together to resolve any discrepancies and synthesize the two versions of the translated measure. The translated questionnaire was pilot tested with 64 treatment-seeking parents of children with obesity (pretesting). Concerns arose with regard to five items; these items had been the most difficult to resolve during the back-translation, and further analyses revealed that Spanish-speaking parents scored differently from English-speaking parents on these items. To address the concerns, two different bilingual staff members held a focus group with five Spanish-speaking families (cognitive interviewing; these families were not included in the final sample, as their feedback was used to adjust the translation of the measure). The families generated ideas about how the wording of the five questions could be modified to best reflect the key themes of the questions; this feedback was given to one of the original translators. Changes were then made to the measure after holding an expert panel (consisting of qualified bilingual interpreters, bilingual graduate students in psychology, and experts in weight management) discussion to develop the correct terminology that most accurately reflected the meaning of the original measure. Specifically, changes were made to two items (6 and 18) because the original translations did not capture the connotation of the English words “argue” and “upset”; different Spanish words or phrases were chosen to better capture the meaning (“no estuvo de acuerdo sobre” was chosen over “discutido” to best capture “argue” and “hizo berrinche” was chosen to best capture “upset.”). Similarly, Items 14 and 19 were altered to replace the verbs with those that most accurately captured the ideas of teasing (“se burlaron” was chosen over “molestado”) and physically keeping up with others (“mantaner el mismo nivel de actividad fisica” was chosen over “convivir”). The final item (Number 17) was not changed, as the recommendations were deemed stylistic for verb choice. The current analyses focused on the final version of the translated form (final version). Measures Obesity-Specific HRQOL The Spanish translation of Sizing Them Up (Modi & Zeller, 2008) was used to assess parent-report of their child’s obesity-specific HRQOL. The 22-item measure assessed for child functioning in a variety of domains including emotional functioning (e.g., “Felt sad because of their weight/shape/size.”), physical functioning (e.g., “Had difficulty keeping up with other children because of their weight/shape/size.”), school functioning (e.g., “Chose not to go to school because of their weight/shape/size.”), teasing/marginalization (e.g., “Was teased by others when physically active because of their weight/shape/size.”), positive attributes (e.g., “Seen as having a good sense of humor.”), and mealtime challenges (e.g., “Argued about when, what and how much to eat.”). In addition to domain scores, Sizing Them Up also yields a Total HRQOL score. Parents rated how much their child’s weight/shape/size impacted their child’s functioning in the past 1 month on a scale from 1 (“Never”) to 4 (“Always”); the PSA scale is reverse scored. Sizing Them Up has been shown to have adequate internal consistency (coefficients ranged from 0.52 to 0.90), test–retest reliability (coefficients ranged 0.57–0.80), and construct and convergent validity (Modi & Zeller, 2008; Quittner et al., 2009). Weight Status Child height (in cm) and weight (in kg) were collected by trained research staff, while the children were wearing light clothing and no shoes. BMI percentile and zBMI were calculated based on child age and sex as recommended by the Centers for Disease Control and Prevention (CDC, 2011). Demographics Parent (i.e., sex, caregiver status, weight) and child (i.e., age, sex, race/ethnicity) demographic information was abstracted from the medical or research record. Data Analysis Plan MPlus Version 7 (Muthén & Muthén, 1998-2012) was used for factor analyses. A CFA of the translated Sizing Them Up measure was performed to assess the goodness of fit of the original factor structure (Modi & Zeller, 2008). First, a single-factor model was specified with all 22 items representing a single factor of Total HRQOL (Model 1). Next, Model 2 adjusted for nonsignificant loadings and tested a two-factor model with the PSA subscale as one factor, and the remaining 18 items as a second factor. Model 3 tested the 6-factor solution derived from the original EFA1. Model 4 tested the applicability of a second order factor as a composite of the 6 first order factors. The final model (Model 5) adjusted for the nonsignificant loading of the PSA subscale, and tested PSA as a distinct factor, with the remaining five factors as first-order factors and a second-order Total HRQOL factor. In all models, a robust weighted least squares estimation with a diagonal weight matrix and a mean- and variance-adjusted chi-square test statistic was used to account for the ordinal scale of Sizing Them Up (i.e., four responses along a Likert scale). Other factor analysis methods (e.g., maximum likelihood; commonly used for continuous data) use Pearson correlation matrices and are inappropriate for use with ordinal data (Brown, 2006). The WLSMV estimator also allows for a nonnormal distribution. Also, the marker variable method of identification was used, in which the first indicator of each construct was set to a loading of 1. Of the 154 participants, 115 parents provided complete data, resulting in 25.32% of participants with at least one missing item. Data were determined to be missing completely at random because the Little’s (1988) Test of Missing Data (IBM SPSS) was nonsignificant (Schlomer, Bauman, & Card, 2010); therefore, missing data were handled through the use of multiple imputation, with 100 imputed data sets being used for all CFA analyses. Model fit statistics were assessed to determine the degree to which the hypothesized (implied) model fit the observed data. All models were evaluated by examining the χ2 test of significance, root mean squared error of approximation (RMSEA), comparative fit index (CFI), and Tucker–Lewis Index (TLI). Model fit was considered to be acceptable if the RMSEA values were at or below 0.08, and good/close if values were below 0.05 (Little, 2013). Model fit was considered to be acceptable for CFI and TLI values above 0.90, and good for values above 0.95 (Little, 2013). Individual item and factor loadings were evaluated and considered acceptable if loadings were equal to or greater than 0.4 (Brown, 2006). Scaled scoring of each subscale and the total score were calculated according to the manual put forth by the authors of the original Sizing Them Up (scales are scored on a range of 0–100; 100 = no impairment in functioning; Manual Scoring for Sizing Them Up). Cronbach’s α was calculated using IBM SPSS Statistics Version 22 to evaluate the internal consistency of the subscales and Total HRQOL scale of the Spanish translation of Sizing Them Up. Excellent, good, and acceptable internal consistency was determined using the cutoffs of 0.9, 0.8, and 0.7, respectively (George & Mallery, 2003). Additionally, bivariate correlations between the subscales/total scale of the Spanish translation of Sizing Them Up and zBMI, age, and gender were conducted using IBM SPSS Statistics Version 22. Results Spanish-speaking Latino youth participants (N = 154) were 52.2% female (M age = 10.84; SD = 2.47) with a mean BMI score of 29.11 (SD = 5.68), mean zBMI of 2.23 (SD = 0.31), and BMI percentile of 98.34 (SD = 1.28). Of those with insurance status identified (N = 80), 78.8% had Medicaid, 11.3% had commercial insurance, and 10.0% had no insurance or were self-pay. The vast majority of participating caregivers was female and mothers (97.2%). Of those with caregiver weight status data (N = 56), mean BMI was 33.67 (SD = 5.90). Confirmatory Factor Analyses With all 22 items included, the one-factor model (Model 1) produced a mediocre fit to the data (χ2 = 557.83, n = 154, p < .001, RMSEA = 0.104, CFI = 0.892, TLI = 0.880). Further inspection revealed that items 10, 15, 17, and 20 had factor loadings that were negative or <0.1, and were nonsignificant. These four items represented the originally conceptualized PSA scale, so a two-factor solution was then tested, with the original PSA scale as one factor, and the remaining 18 items as a second factor. Results demonstrated that the two-factor solution (Model 2) produced a mediocre to acceptable fit to the data (χ2 = 413.154, n = 154, p < .001, RMSEA = 0.080, CFI = 0.936, TLI = 0.929), and all indicators loaded significantly with acceptable levels on the respective two scales. Based on this information, Model 2 fit the data better than Model 1; however, additional models were tested to determine if good or close model fit could be achieved. To replicate the factor structure of Sizing Them Up that was identified in the original EFA, the translated 22-item six-factor model was tested next. The residual variance of the School Functioning subscale (consisting of one item) was set to 1, because the residual error variance of categorical variables cannot be estimated. With all 22 items included, the six-factor model (Model 3) produced an acceptable to very good fit to the data (χ2 = 287.112, n = 154, p < .001, RMSEA = 0.055, CFI = 0.971, TLI = 0.966). Further inspection of the indicators revealed no indicator loadings which were <0.4 or nonsignificant; therefore, this model was determined to be an acceptable fit to the data and a better fitting model than Models 1 and 2. Model 3 would be appropriate for use among researchers and clinicians interested in individual domains of HRQOL. To test the utility of a Total HRQOL score for Sizing Them Up, the six-factor solution was then tested under a higher-order framework, in which the six first-order factors were set to comprise a second-order factor, Total HRQOL. This model (Model 4) produced an acceptable to very good fit to the data (χ2 = 339.497, N = 154, p < .001, RMSEA = 0.066, CFI = 0.958, TLI = 0.952), but the PSA scale did not load significantly onto the second-order factor (p = .783); therefore, Model 4 is not recommended for use. A subsequent model with PSA removed from the second-order factor produced no changes in fit indices. Further inspection revealed that the Teasing/Marginalization scale had a standardized negative residual variance of −0.02. Because the negative residual variance was small and nonsignificant, a final model (Model 5) was tested with the residual variance of Teasing/Marginalization set to 0 and the PSA removed from the higher-order factor. This model produced acceptable and significant indicator loadings at the first-order and second-order levels and was an acceptable to very good fit to the data (χ2 = 339.829, n = 154, p < .001, RMSEA = 0.065, CFI = 0.958, TLI = 0.953). Although Model 3 resulted in slightly improved fit statistics compared with Model 5, Model 5 demonstrates that all subscales, except for PSA, are representative of a unified construct (i.e., Total Quality of Life). Given both the clinical utility of a total score and the theory that Total Quality of Life is composed of multiple domains (e.g., emotional, physical, school, teasing/marginalization, positive attributes, and mealtime challenges), we believe that Model 5 (with all subscales except PSA as first-order factors and total QOL as a second-order factor) is also a theoretically sound model and would be appropriate for use. Model fit statistics for Model 5 are presented in Table I, and factor loadings and standard errors are presented in Table II. Table I. Fit Statistics for CFA Model χ2 df p RMSEA CFI TLI One-factor (Model 1) 557.83 209 .00 0.10 0.89 0.88 Two-factor (Model 2) 413.15 208 .00 0.08 0.94 0.93 Six-factor (Model 3) 287.11 195 .00 0.06 0.97 0.97 Second-order factor (Model 4) 339.50 204 .00 0.07 0.96 0.95 Second-order factora (Model 5) 339.83 205 .00 0.07 0.96 0.95 Model χ2 df p RMSEA CFI TLI One-factor (Model 1) 557.83 209 .00 0.10 0.89 0.88 Two-factor (Model 2) 413.15 208 .00 0.08 0.94 0.93 Six-factor (Model 3) 287.11 195 .00 0.06 0.97 0.97 Second-order factor (Model 4) 339.50 204 .00 0.07 0.96 0.95 Second-order factora (Model 5) 339.83 205 .00 0.07 0.96 0.95 a Second-order factor structure excludes the Positive Social Attributes subscale; definitions: CFA = confirmatory factor analyses; CFI = comparative fit index; χ2 = chi-square; df = degrees of freedom; RMSEA = root mean squared error of approximation; TLI = Tucker–Lewis Index. Table I. Fit Statistics for CFA Model χ2 df p RMSEA CFI TLI One-factor (Model 1) 557.83 209 .00 0.10 0.89 0.88 Two-factor (Model 2) 413.15 208 .00 0.08 0.94 0.93 Six-factor (Model 3) 287.11 195 .00 0.06 0.97 0.97 Second-order factor (Model 4) 339.50 204 .00 0.07 0.96 0.95 Second-order factora (Model 5) 339.83 205 .00 0.07 0.96 0.95 Model χ2 df p RMSEA CFI TLI One-factor (Model 1) 557.83 209 .00 0.10 0.89 0.88 Two-factor (Model 2) 413.15 208 .00 0.08 0.94 0.93 Six-factor (Model 3) 287.11 195 .00 0.06 0.97 0.97 Second-order factor (Model 4) 339.50 204 .00 0.07 0.96 0.95 Second-order factora (Model 5) 339.83 205 .00 0.07 0.96 0.95 a Second-order factor structure excludes the Positive Social Attributes subscale; definitions: CFA = confirmatory factor analyses; CFI = comparative fit index; χ2 = chi-square; df = degrees of freedom; RMSEA = root mean squared error of approximation; TLI = Tucker–Lewis Index. Table II. Factor Loadings and Standard Errors for Indicators of Final Models Latent Factor and Indicators Loading Emotional functioning  Q4: Felt sad 0.89 (0.02)*  Q8: Felt frustrated 0.90 (0.03)*  Q9: Avoided dressing in front of others 0.69 (0.05)*  Q11: Felt worried 0.73 (0.04)*  Q13: Felt mad 0.91 (0.02)*  Q16: Felt concerned 0.76 (0.04)*  Q22: Had low self-esteem 0.82 (0.04)* Physical functioning  Q1: Difficulty participating in physical activities 0.90 (0.03)*  Q5: Changes to physical surroundings 0.76 (0.06)*  Q7: Chose not to participate in gym 0.77 (0.08)*  Q19: Difficulty keeping up with other childrena 0.68 (0.07)*  Q21: Became out of breath 0.68 (0.06)* Teasing/marginalization  Q2: Teased by others 0.86 (0.03)*  Q12: Felt left out 0.88 (0.04)*  Q14: Teased when physically activea 0.86 (0.04)* Positive Social Attributes  Q10: Kept body clean 0.46 (0.11)*  Q15: Good sense of humor 0.67 (0.08)*  Q17: Perceived as healthya 0.66 (0.09)*  Q20: Felt successful in daily activities 0.62 (0.09)* Mealtime challenges  Q6: Argued about eatinga 0.61 (0.09)*  Q18: Upset at mealtimesa 0.64 (0.09)* School functioning  Q3: Chose not to go to school 1.00 (0.00) Higher-order  Emotional Functioning 0.86 (0.03)*  Physical Functioning 0.84 (.03)  Teasing/Marginalization 1.0 (0.0)b  Positive Social Attributes R  Mealtime Challenges 0.79 (0.11)*  School Functioning 0.60 (0.03)* Latent Factor and Indicators Loading Emotional functioning  Q4: Felt sad 0.89 (0.02)*  Q8: Felt frustrated 0.90 (0.03)*  Q9: Avoided dressing in front of others 0.69 (0.05)*  Q11: Felt worried 0.73 (0.04)*  Q13: Felt mad 0.91 (0.02)*  Q16: Felt concerned 0.76 (0.04)*  Q22: Had low self-esteem 0.82 (0.04)* Physical functioning  Q1: Difficulty participating in physical activities 0.90 (0.03)*  Q5: Changes to physical surroundings 0.76 (0.06)*  Q7: Chose not to participate in gym 0.77 (0.08)*  Q19: Difficulty keeping up with other childrena 0.68 (0.07)*  Q21: Became out of breath 0.68 (0.06)* Teasing/marginalization  Q2: Teased by others 0.86 (0.03)*  Q12: Felt left out 0.88 (0.04)*  Q14: Teased when physically activea 0.86 (0.04)* Positive Social Attributes  Q10: Kept body clean 0.46 (0.11)*  Q15: Good sense of humor 0.67 (0.08)*  Q17: Perceived as healthya 0.66 (0.09)*  Q20: Felt successful in daily activities 0.62 (0.09)* Mealtime challenges  Q6: Argued about eatinga 0.61 (0.09)*  Q18: Upset at mealtimesa 0.64 (0.09)* School functioning  Q3: Chose not to go to school 1.00 (0.00) Higher-order  Emotional Functioning 0.86 (0.03)*  Physical Functioning 0.84 (.03)  Teasing/Marginalization 1.0 (0.0)b  Positive Social Attributes R  Mealtime Challenges 0.79 (0.11)*  School Functioning 0.60 (0.03)* a Discrepancies during translation; bResidual variance set to 0; R = Removed; *p ≤ .001. Table II. Factor Loadings and Standard Errors for Indicators of Final Models Latent Factor and Indicators Loading Emotional functioning  Q4: Felt sad 0.89 (0.02)*  Q8: Felt frustrated 0.90 (0.03)*  Q9: Avoided dressing in front of others 0.69 (0.05)*  Q11: Felt worried 0.73 (0.04)*  Q13: Felt mad 0.91 (0.02)*  Q16: Felt concerned 0.76 (0.04)*  Q22: Had low self-esteem 0.82 (0.04)* Physical functioning  Q1: Difficulty participating in physical activities 0.90 (0.03)*  Q5: Changes to physical surroundings 0.76 (0.06)*  Q7: Chose not to participate in gym 0.77 (0.08)*  Q19: Difficulty keeping up with other childrena 0.68 (0.07)*  Q21: Became out of breath 0.68 (0.06)* Teasing/marginalization  Q2: Teased by others 0.86 (0.03)*  Q12: Felt left out 0.88 (0.04)*  Q14: Teased when physically activea 0.86 (0.04)* Positive Social Attributes  Q10: Kept body clean 0.46 (0.11)*  Q15: Good sense of humor 0.67 (0.08)*  Q17: Perceived as healthya 0.66 (0.09)*  Q20: Felt successful in daily activities 0.62 (0.09)* Mealtime challenges  Q6: Argued about eatinga 0.61 (0.09)*  Q18: Upset at mealtimesa 0.64 (0.09)* School functioning  Q3: Chose not to go to school 1.00 (0.00) Higher-order  Emotional Functioning 0.86 (0.03)*  Physical Functioning 0.84 (.03)  Teasing/Marginalization 1.0 (0.0)b  Positive Social Attributes R  Mealtime Challenges 0.79 (0.11)*  School Functioning 0.60 (0.03)* Latent Factor and Indicators Loading Emotional functioning  Q4: Felt sad 0.89 (0.02)*  Q8: Felt frustrated 0.90 (0.03)*  Q9: Avoided dressing in front of others 0.69 (0.05)*  Q11: Felt worried 0.73 (0.04)*  Q13: Felt mad 0.91 (0.02)*  Q16: Felt concerned 0.76 (0.04)*  Q22: Had low self-esteem 0.82 (0.04)* Physical functioning  Q1: Difficulty participating in physical activities 0.90 (0.03)*  Q5: Changes to physical surroundings 0.76 (0.06)*  Q7: Chose not to participate in gym 0.77 (0.08)*  Q19: Difficulty keeping up with other childrena 0.68 (0.07)*  Q21: Became out of breath 0.68 (0.06)* Teasing/marginalization  Q2: Teased by others 0.86 (0.03)*  Q12: Felt left out 0.88 (0.04)*  Q14: Teased when physically activea 0.86 (0.04)* Positive Social Attributes  Q10: Kept body clean 0.46 (0.11)*  Q15: Good sense of humor 0.67 (0.08)*  Q17: Perceived as healthya 0.66 (0.09)*  Q20: Felt successful in daily activities 0.62 (0.09)* Mealtime challenges  Q6: Argued about eatinga 0.61 (0.09)*  Q18: Upset at mealtimesa 0.64 (0.09)* School functioning  Q3: Chose not to go to school 1.00 (0.00) Higher-order  Emotional Functioning 0.86 (0.03)*  Physical Functioning 0.84 (.03)  Teasing/Marginalization 1.0 (0.0)b  Positive Social Attributes R  Mealtime Challenges 0.79 (0.11)*  School Functioning 0.60 (0.03)* a Discrepancies during translation; bResidual variance set to 0; R = Removed; *p ≤ .001. Latent Factor Intercorrelations Latent factor intercorrelations between individual subscale factors were also examined. The latent factors of Emotional Functioning, Physical Functioning, Teasing/Marginalization, Mealtime Behaviors, and School Functioning evidenced good latent intercorrelations (ranged from 0.39 to 0.89), with the majority of values >0.70. However, the PSA scale was not significantly correlated with Emotional Functioning, Teasing/Marginalization, or School Functioning, and evidenced significant correlations (at the p < .01 and p < .05 levels, respectively) with Physical Functioning and Mealtime Challenges in the unexpected direction (i.e., better PSA was associated with worse physical functioning and more mealtime challenges). Latent factor intercorrelations are presented in Table III. Table III. Standardized Interfactor Correlations of the Final Lower-Order Model Latent Factors Emotional Functioning Physical Functioning Teasing/Marginalization Positive Social Attributes Mealtime Challenges School Functioning Emotional Functioning 1.0 Physical Functioning 0.71 (0.05)*** 1.0 Teasing/Marginalization 0.85 (0.04)*** 0.85 (0.04)*** 1.0 Positive Social Attributes −0.14 (0.10)NS 0.19 (0.10)* 0.04 (0.10)NS 1.0 Mealtime Challenges 0.65 (0.08)*** 0.85 (0.09)*** 0.67 (0.10)*** 0.39 (0.13)** 1.0 School Functioning 0.71 (0.08)*** 0.45 (0.08)*** 0.84 (0.07)*** 0.02 (0.16)NS 0.39 (0.13)** 1.0 Latent Factors Emotional Functioning Physical Functioning Teasing/Marginalization Positive Social Attributes Mealtime Challenges School Functioning Emotional Functioning 1.0 Physical Functioning 0.71 (0.05)*** 1.0 Teasing/Marginalization 0.85 (0.04)*** 0.85 (0.04)*** 1.0 Positive Social Attributes −0.14 (0.10)NS 0.19 (0.10)* 0.04 (0.10)NS 1.0 Mealtime Challenges 0.65 (0.08)*** 0.85 (0.09)*** 0.67 (0.10)*** 0.39 (0.13)** 1.0 School Functioning 0.71 (0.08)*** 0.45 (0.08)*** 0.84 (0.07)*** 0.02 (0.16)NS 0.39 (0.13)** 1.0 *** p ≤ .001; **p ≤ .01; *p ≤ .05; NS = Not significant. Table III. Standardized Interfactor Correlations of the Final Lower-Order Model Latent Factors Emotional Functioning Physical Functioning Teasing/Marginalization Positive Social Attributes Mealtime Challenges School Functioning Emotional Functioning 1.0 Physical Functioning 0.71 (0.05)*** 1.0 Teasing/Marginalization 0.85 (0.04)*** 0.85 (0.04)*** 1.0 Positive Social Attributes −0.14 (0.10)NS 0.19 (0.10)* 0.04 (0.10)NS 1.0 Mealtime Challenges 0.65 (0.08)*** 0.85 (0.09)*** 0.67 (0.10)*** 0.39 (0.13)** 1.0 School Functioning 0.71 (0.08)*** 0.45 (0.08)*** 0.84 (0.07)*** 0.02 (0.16)NS 0.39 (0.13)** 1.0 Latent Factors Emotional Functioning Physical Functioning Teasing/Marginalization Positive Social Attributes Mealtime Challenges School Functioning Emotional Functioning 1.0 Physical Functioning 0.71 (0.05)*** 1.0 Teasing/Marginalization 0.85 (0.04)*** 0.85 (0.04)*** 1.0 Positive Social Attributes −0.14 (0.10)NS 0.19 (0.10)* 0.04 (0.10)NS 1.0 Mealtime Challenges 0.65 (0.08)*** 0.85 (0.09)*** 0.67 (0.10)*** 0.39 (0.13)** 1.0 School Functioning 0.71 (0.08)*** 0.45 (0.08)*** 0.84 (0.07)*** 0.02 (0.16)NS 0.39 (0.13)** 1.0 *** p ≤ .001; **p ≤ .01; *p ≤ .05; NS = Not significant. Internal Consistency Internal consistency results are presented in Table IV. The Emotional Functioning subscale and Total HRQOL evidenced good internal consistency, and Physical Functioning and Teasing/Marginalization scales produced acceptable values. The PSA scale demonstrated an internal consistency estimate (α = 0.61) similar to that documented in the original publication (α = 0.59, Modi & Zeller, 2008). The Mealtime challenges subscale evidenced poor internal consistency, with a value (α = 0.48) lower than that documented in the original publication. Table IV. Ms (SD) and Reliability Coefficients Scored Subscales and Total Score M (SD) Cronbach’s α Emotional Functioning 72.84 (22.82) 0.88 Physical Functioning 82.26 (17.90) 0.74 Teasing/Marginalization 80.00 (22.95) 0.77 Positive Social Attributes 61.88 (22.70) 0.61 Mealtime Challenges 68.84 (24.53) 0.48 School Functioninga 94.78 (16.28) – Total Quality of Life 74.60 (15.07) 0.87 Modified Total Quality of Life 77.42 (17.69) 0.91 Scored Subscales and Total Score M (SD) Cronbach’s α Emotional Functioning 72.84 (22.82) 0.88 Physical Functioning 82.26 (17.90) 0.74 Teasing/Marginalization 80.00 (22.95) 0.77 Positive Social Attributes 61.88 (22.70) 0.61 Mealtime Challenges 68.84 (24.53) 0.48 School Functioninga 94.78 (16.28) – Total Quality of Life 74.60 (15.07) 0.87 Modified Total Quality of Life 77.42 (17.69) 0.91 a Scale contains only one item. Table IV. Ms (SD) and Reliability Coefficients Scored Subscales and Total Score M (SD) Cronbach’s α Emotional Functioning 72.84 (22.82) 0.88 Physical Functioning 82.26 (17.90) 0.74 Teasing/Marginalization 80.00 (22.95) 0.77 Positive Social Attributes 61.88 (22.70) 0.61 Mealtime Challenges 68.84 (24.53) 0.48 School Functioninga 94.78 (16.28) – Total Quality of Life 74.60 (15.07) 0.87 Modified Total Quality of Life 77.42 (17.69) 0.91 Scored Subscales and Total Score M (SD) Cronbach’s α Emotional Functioning 72.84 (22.82) 0.88 Physical Functioning 82.26 (17.90) 0.74 Teasing/Marginalization 80.00 (22.95) 0.77 Positive Social Attributes 61.88 (22.70) 0.61 Mealtime Challenges 68.84 (24.53) 0.48 School Functioninga 94.78 (16.28) – Total Quality of Life 74.60 (15.07) 0.87 Modified Total Quality of Life 77.42 (17.69) 0.91 a Scale contains only one item. Correlations The correlations between zBMI and Sizing Them Up subscales and Total HRQOL score were conducted to determine convergent validity. Significant correlations were established between zBMI and the Physical Functioning scale (r = −.24, p < .05), the Teasing/Marginalization scale (r = −.22, p < .05), the Mealtime Challenges scale (r = −.22, p < .05), and Total HRQOL (r = −.23, p < .05). No significant correlations were found between zBMI and the Emotional Functioning, PSA, or the School Functioning subscales. Gender and age were not significantly associated with any scale scores or the Total HRQOL score. Discussion The purpose of the current study was to examine the psychometric properties of a Spanish translation of Sizing Them Up, a parent-report measure of obesity-specific HRQOL. Pediatric obesity disproportionally affects Latino youth, and previous research has indicated that HRQOL is a key measure in pediatric obesity (e.g., Buttitta et al., 2014; Schwimmer et al., 2003; Ul-Haq et al., 2013). Yet, there are no identified Spanish language parent-report measures of weight-related HRQOL. The current study extends the literature by examining the psychometric properties of the Spanish translation of Sizing Them Up. The first hypothesis, that the previously established six-factor structure would be replicated, was partially supported. The results support the use of the Spanish translation of Sizing Them Up among Spanish-speaking parents of treatment-seeking youth with obesity if the six factors are interpreted independently. However, under a structural equation modeling (SEM) framework, results do not support the use of a Total HRQOL score with this translated version of the measure, because the PSA subscale was not found to be representative of Total HRQOL, but instead was a separate factor. This finding was supported by the nonsignificant loading of the PSA factor in the model testing for a second-order factor. These findings suggest that the PSA scale from Sizing Them Up likely functions differently for the Latino population. This scale contains four items, including “kept their body clean and fresh,” “seen as having a good sense of humor,” “perceived as healthy by others,” and “felt successful in daily activities.” Recent research indicates that Latino families do not associate pediatric obesity with poor health in the same way that non-Latino Black and non-Latino Caucasian families do, and in fact interpret overweight to be an indicator of good health (Baker & Altman, 2015). For example, research suggests that Latino mothers perceive thinness to be an indicator of malnutrition, and thus see thinness as more dangerous than overweight/obesity for their children (Crawford et al., 2004). Thus, Latino families may perceive their children with obesity to display more positive attributes than do Black and Caucasian families. Second, it is possible that the overall construct of PSA is perceived differently among Latino families, above and beyond the specific items included in the subscale. More research is needed to learn more about how culture may affect this subscale. The second hypothesis, regarding the latent factor intercorrelations, was partially supported. As expected, the majority of the intercorrelations were medium to large positive correlations. However, the PSA scale was not significantly correlated with Emotional Functioning, Teasing/Marginalization, or School Functioning. The PSA subscale also had a negative correlation with the Physical Functioning and Mealtime Challenges subscales. The lack of significant correlations between PSA and several other subscales is similar to results from the psychometric evaluation of the self-report measure of obesity-specific HRQOL, Sizing Me Up, in which Tripicchio and colleagues (2017) found the PSA subscale to be uncorrelated with the Physical Functioning, Teasing/marginalization, and Social Avoidance Subscales. However, the negative correlations in the current study stand in contrast to results from the original validation study (Modi & Zeller, 2008), although in the original validation study, the PSA scale demonstrated a nonsignificant correlation with the school functioning subscale and loadings <0.3 with all other subscales. These findings provide further evidence that the PSA subscale was not found to be representative of Total HRQOL in this population, for both parent- and self-report. The third hypothesis, that Cronbach’s α for Total HRQOL would be >0.8 and for the subscales would range from 0.5 to 0.9, was partially supported. The findings from the current study were consistent with the previous study by Modi and Zeller (2008) for all scales except for the Mealtime Challenges subscale. The poor internal consistency for the Mealtime Challenges scale was not unexpected given the difficulties with the translation of both items, which comprise this scale. As discussed previously, the translation of the words for “argued” and “upset” proved challenging, due at least in part to regional variations on the proper translation of these words into Spanish. While our team was unable to evaluate country of origin for the parents’ Spanish influence, it is well recognized that there are regional variations in dialect and in word connotations, especially with emotionally laden terms. Future researchers would be wise to consider evaluating this subscale further, to identify more tailored phrases. Additionally, although the findings for the internal consistency for the PSA subscale in the current study were consistent with the previous study (α = 0.61 in current sample; α = 0.59 in Modi & Zeller, 2008), the Cronbach’s α is considered below acceptable (George & Mallery, 2003). Therefore, researchers and clinicians should use caution when using the Mealtime Challenges and PSA subscales of the Spanish translation of Sizing Them Up because the items on the subscales may not be adequately measuring the same underlying constructs. Although the Cronbach’s α for Total HRQOL reached acceptable levels, this is unexpected given the CFA results. The use of a Total HRQOL score that includes PSA is cautioned against, because SEM accounts for error and is a more acceptable estimator of reliability. The correlation hypotheses were also partially supported. Although it was only hypothesized that the Physical Functioning scale would be associated with zBMI based on the previous study, the current study found significant correlations between zBMI and the Physical Functioning scale, the Teasing/Marginalization scale, the Mealtime Challenges scale, and Total HRQOL; these correlations provide evidence for convergent validity. It should also be noted that although the Teasing/Marginalization scale was highly correlated with the School Functioning and Emotional Functioning scales, these do appear to be distinct constructs as only the Teasing/Marginalization scale was significantly associated with BMIz. Contrary to findings from Modi and Zeller (2008), age was not significantly associated with any scale scores. Modi and Zeller (2008) posited that self-esteem worsens as youth transition to adolescence, and self-esteem likely influenced the Emotional Functioning and PSA subscales. Given our findings, it is possible that weight status may not be as related to self-esteem among Latino youth as for those from other ethnic backgrounds. Consistent with previous findings, gender was not associated with the scale scores in the current sample. Thus, the Spanish translation may have utility for both sexes and all age groups. Limitations of the current study must be considered when interpreting findings. For example, the sample for the current analysis was relatively small, represented a predominantly low-income population, and included predominately female (mother) reporters, which may limit the generalizability of the results. Further, across all subscales and the Total HRQOL score, the mean scores for Sizing Them Up were higher in this population than in the original validation study. Reasons for these discrepancies may be related to the cultural differences in interpretation of overweight/obesity mentioned earlier, namely, the perception that excess weight is an indicator of health, and a perception that thinness is an indicator of malnutrition (Baker & Altman, 2015; Crawford et al., 2004). These perceptions could certainly impact reported HRQOL. These results may also be explained by a lower level of illness in this sample as compared with the original validation sample, as evidenced by participants presenting with a lower BMIz (2.23) than those from the original validation sample (2.60). It is likely that those in the current study experienced less impairment because of obesity given their relatively lower BMIz at baseline; therefore, the current results may not be generalizable to youth with higher BMIz. Future studies should examine psychometric properties in various Latino samples, particularly among Latino youth who are presenting to bariatric surgery, to comprehensively evaluate the psychometric properties of this scale, and to provide further evidence of the utility of the measure. Additionally, because of limited sample size, the current study did not conduct posttreatment analyses, and future research should evaluate additional measures of reliability and validity, including test–retest reliability and invariance testing. Multiple-group invariance testing between Spanish-speaking and English-speaking Latino parents is needed to completely determine whether the current translated measure is truly capturing the same constructs. Owing to insufficient sample size, the current study did not evaluate a translated version of the Adolescent Developmental Adaptation scale, which is a subscale of Sizing Them Up intended only for adolescent participants. Finally, the current study did not measure country of origin, acculturation status, or parent education level of participants, which would be have been beneficial in ensuring the measure’s applicability to Spanish speakers from a variety of backgrounds. Despite these limitations, the current study has many strengths. For example, the current study provides the first Spanish translation of a parent-report of obesity-specific HRQOL. The current study included the use of well-established translation procedures for creating the Spanish translation of the measure. Additionally, the current study drew from two geographical locations with high populations of Latino youth and Spanish-speaking parents, increasing generalizability of results. Overall, results support the use of the Spanish translation of Sizing Them Up using the six factors independently in an SEM framework, which may be appropriate for researchers interested in individual subscales. However, researchers and clinicians are cautioned against computing a Total HRQOL score with all six subscales included. Results do support the use of a Total HRQOL score of the remaining five factors (Emotional Functioning, Physical Functioning, Teasing/Marginalization, Mealtime Challenges, and School Functioning). We suggest that providers wishing to use a Total HRQOL score for this measure adopt a modified formula from the original formula for Sizing Them Up (Original Total HRQOL = ((Sum of 22 Items−22)/66) * 100; Manual for Scoring Sizing Them Up) for scoring without the PSA subscale, in which Modified Total HRQOL = ((Sum of 18 Items−18)/54) * 100. While the scores derived from this modified formula demonstrated a high correlation (r = .96) with those using the original 22-item formula, further testing of the validity of the formula is warranted for future research. Further, given the clear divergence between the PSA Scale and all other Sizing Them Up scales, it is important for future research to better understand the applicability of the PSA scale for all populations. Thus, it is recommended that future studies conduct a CFA of Sizing Them Up in a large representative sample of English-speaking parents of youth with obesity. Additionally, if a CFA confirms the appropriate inclusion of PSA in the English version of Sizing Them Up, additional field testing (e.g., focus groups) may be useful to better understand why the interpretation of PSA may be different in this population. Funding This work was supported by the Michael and Susan Dell Foundation; Health Care Foundation of Greater Kansas City; and the Junior League of Kansas City. Conflicts of interest: None declared. Footnotes 1 Although it is common practice for subscales to include at least three items, the authors chose to include the one-item School Functioning and the two-item Mealtime Challenges subscales to replicate the original English version of the measure validated through exploratory factor analysis. References Baker E. H. , Altman C. E. ( 2015 ). Maternal ratings of child health and child obesity, variations by mother’s race/ethnicity and nativity . Maternal and Child Health Journal , 19 , 1000 – 1009 . Google Scholar Crossref Search ADS PubMed Black W. R. , Davis A. M. , Gillette M. L. , Short M. B. , Wetterneck C. T. , He J. ( 2014 ). Health-related quality of life in obese and overweight, treatment-seeking youth . Ethnicity and Disease , 24 , 321 – 327 . Google Scholar PubMed Brown T. A. ( 2006 ). Confirmatory factor analysis for applied research . New York, NY : The Guilford Press . Buttitta M. , Iliescu C. , Rousseau A. , Guerrien A. ( 2014 ). Quality of life in overweight and obese children and adolescents: A literature review . Quality of Life Research , 23 , 1117 – 1139 . Google Scholar Crossref Search ADS PubMed Centers for Disease Control and Prevention (CDC) . ( 2011 ). Health-related quality of life . Retrieved from http://www.cdc.gov/hrqol/concept.htm Crawford P. B. , Gosliner W. , Anderson C. , Strode P. , Becerra-Jones Y. , Samuels S. , Carroll A. M. , Ritchie L. D. ( 2004 ). Counseling Latina mothers of preschool children about weight issues: Suggestions for a new framework . Journal of the American Dietetic Association , 104 , 387 – 394 . Google Scholar Crossref Search ADS PubMed Dabelea D. , Mayer-Davis E. J. , Saydah S. , Imperatore G. , Linder B. , Divers J. , Bell R. , Badaru A. , Talton J. W. , Crume T. , Liese A. D. , Merchant A. T. , Lawrence J. M. , Reynolds K. , Dolan L , Liu LL. , Hamman R. F. ; SEARCH for Diabetes in Youth Study . ( 2014 ). Prevalence of type 1 and type 2 diabetes among children and adolescents from 2001 to 2009 . Journal of the American Medical Association , 311 , 1778 – 1786 . Google Scholar Crossref Search ADS PubMed George D. , Mallery P. ( 2003 ). Reliability analysis . SPSS for Windows, step by step: A simple guide and reference, 14th ed. Boston : Allyn & Bacon , 222 – 232 . Guilfoyle S. M. , Zeller M. H. , Modi A. C. ( 2010 ). Parenting stress impacts obesity-specific health-related quality of life in a pediatric obesity treatment-seeking sample . Journal of Developmental and Behavioral Pediatrics , 31 , 17 . Google Scholar Crossref Search ADS PubMed Kalarchian M. A. , Marcus M. D. ( 2012 ). Psychiatric comorbidity of childhood obesity . International Review of Psychiatry , 24 , 241 – 246 . Google Scholar Crossref Search ADS PubMed Kuyken W. ( 1995 ). The World Health Organization Quality of Life Assessment (WHOQOL): Position paper from the world health organization . Social Science and Medicine , 41 , 1403 . Google Scholar Crossref Search ADS PubMed Little R. J. A. ( 1988 ). A test of missing completely at random for multivariate data with missing values . Journal of the American Statistical Association , 83 , 1198 – 1202 . Google Scholar Crossref Search ADS Little T. D. ( 2013 ). Longitudinal structural equation modeling . New York, NY : The Guilford Press . Muthén L. K. , Muthén B. O. ( 1998 -2012). Mplus user's guide ( 7th ed. ). Los Angeles, CA : Muthén & Muthén . Modi A. C. , Zeller M. H. ( 2008 ). Validation of a parent-proxy, obesity-specific quality-of-life measure: Sizing Them Up . Obesity , 16 , 2624 – 2633 . Google Scholar Crossref Search ADS PubMed Ogden C. L. , Carroll M. D. , Fryar C. D. , Flegal K. M. ( 2015 ). Prevalence of obesity among adults and youth: United States, 2011-2014 . NCHS Data Brief , 219 , 1 – 8 . Ogden C. L. , Carroll M. D. , Kit B. K. , Flegal K. M. ( 2014 ). Prevalence of childhood and adult obesity in the United States, 2011-2012 . Journal of the American Medical Association , 311 , 806 – 814 . Google Scholar Crossref Search ADS PubMed Quittner A. L. , Cruz I. , Modi A. C. , Marciel K. K. ( 2009 ). Health-related quality of life instruments for adolescents with chronic diseases. In O’Donohue W. T. , Woodward L. W. (Eds.), Behavioral approaches to chronic disease in adolescence: A guide to integrative care . (pp. 311 – 327 ). New York, NY : Springer Science + Business Media . Reilly J. J. , Kelly J. ( 2011 ). Long-term impact of overweight and obesity in childhood and adolescence on morbidity and premature mortality in adulthood: Systematic review . International Journal of Obesity , 35 , 891 – 898 . Google Scholar Crossref Search ADS PubMed Ryan C. ( 2013 ). Language use in the United States: 2011 . Washington, DC : U.S. Census Bureau . Sanchez-Villegas A. , Field A. E. , O'Reilly E. J. , Fava M. , Gortmaker S. , Kawachi I. , Ascherio A. ( 2013 ). Perceived and actual obesity in childhood and adolescence and risk of adult depression . Journal of Epidemiology and Community Health , 67 , 81 – 86 . Google Scholar Crossref Search ADS PubMed Schlomer G. L. , Bauman S. , Card N. A. ( 2010 ). Best practices for missing data management in counseling psychology . Journal of Counseling Psychology , 57 , 1 – 10 . Google Scholar Crossref Search ADS PubMed Schwimmer J. B. , Burwinkle T. M. , Varni J. W. ( 2003 ). Health-related quality of life of severely obese children and adolescents . JAMA: Journal of the American Medical Association , 289 , 1813 – 1819 . Google Scholar Crossref Search ADS Schwimmer J. B. , Deutsch R. , Kahen T. , Lavine J. E. , Stanley C. , Behling C. ( 2006 ). Prevalence of fatty liver in children and adolescents . Pediatrics , 118 , 1388 – 1393 . Google Scholar Crossref Search ADS PubMed Taveras E. M. , Gillman M. W. , Kleinman K. , Rich-Edwards J. W. , Rifas-Shiman S. L. ( 2010 ). Racial/ethnic differences in early-life risk factors for childhood obesity . Pediatrics , 125 , 686 – 695 . Google Scholar Crossref Search ADS PubMed Taylor E. D. , Theim K. R. , Mirch M. C. , Ghorbani S. , Tanofsky-Kraff M. , Adler-Wailes D. C. , Brady S. , Reynolds J. C. , Calis K. A. , Yanovski J. A. ( 2006 ). Orthopedic complications of overweight in children and adolescents . Pediatrics , 117 , 2167 – 2174 . Google Scholar Crossref Search ADS PubMed Tripicchio G. L. , Borner K. B. , Odar Stough C. , Poppert Cordts K. , Dreyer Gillette M. , Davis A. M. ( 2017 ). Confirmatory factor analysis of Sizing Me Up: Validation of an obesity-specific health-related quality of life measure in Latino youth . Journal of Pediatric Psychology , 42 , 457 – 465 . Google Scholar PubMed Ul-Haq Z. , Mackay D. F. , Fenwick E. , Pell J. P. ( 2013 ). Meta-analysis of the association between body mass index and health-related quality of life among children and adolescents, assessed using the pediatric quality of life inventory index . The Journal of Pediatrics , 162 , 280 – 286 . Google Scholar Crossref Search ADS PubMed Varni J. W. , Curtis B. H. , Abetz L. N. , Lasch K. E. , Piault E. C. , Zeytoonjian A. A. ( 2013 ). Content validity of the PedsQL™ 3.2 diabetes module in newly diagnosed patients with type 1 diabetes mellitus ages 8–45 . Quality of Life Research , 22 , 2169 – 2181 . Google Scholar Crossref Search ADS PubMed World Health Organization . ( 2016 ). Process of translation and adaptation of instruments . Retrieved from http://www.who.int/substance_abuse/research_tools/translation/en/ Zeller M. H. , Modi A. C. ( 2009 ). Development and initial validation of an obesity-specific quality-of-life measure for children: Sizing Me Up . Obesity , 17 , 1171 – 1177 . Google Scholar Crossref Search ADS PubMed Zeller M. H. , Noll J. G. , Sarwer D. B. , Reiter-Purtill J. , Rofey D. L. , Baughcum A. E. , Peugh J. , Courcoulas A. P. , Michalsky M. P. , Jenkins T. M. , Becnel J. N. ; TeenView Study Group and in Cooperation With Teen-LABS Consortium . ( 2015 ). Child maltreatment and the adolescent patient with severe obesity: Implications for clinical care . Journal of Pediatric Psychology , 40 , 640 – 648 . Google Scholar Crossref Search ADS PubMed © The Author(s) 2018. Published by Oxford University Press on behalf of the Society of Pediatric Psychology. All rights reserved. For permissions, please e-mail: journals.permissions@oup.com This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/about_us/legal/notices) http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Journal of Pediatric Psychology Oxford University Press

Factor Structure of a Spanish Translation of an Obesity-Specific Parent-Report Measure of Health-Related Quality of Life

Loading next page...
 
/lp/ou_press/factor-structure-of-a-spanish-translation-of-an-obesity-specific-cw0EEXmcvi
Publisher
Oxford University Press
Copyright
© The Author(s) 2018. Published by Oxford University Press on behalf of the Society of Pediatric Psychology. All rights reserved. For permissions, please e-mail: journals.permissions@oup.com
ISSN
0146-8693
eISSN
1465-735X
D.O.I.
10.1093/jpepsy/jsy030
Publisher site
See Article on Publisher Site

Abstract

Abstract Objective Latino youth are disproportionately affected by pediatric obesity and consequently experience impaired health-related quality of life (HRQOL). Although many caregivers of Latino youth do not speak English fluently, no validated Spanish translations of obesity-specific HRQOL measures exist for this population. Therefore, non-English-speaking Latino parents have typically been excluded from analyses related to HRQOL. This study assesses the factor structure of a Spanish translation of a parent-report measure of obesity-specific HRQOL, Sizing Them Up, in a treatment-seeking sample of children with obesity. Methods Structural equation modeling was used to assess the factor structure of the 6-subscale, 22-item Sizing Them Up measure in 154 parents of treatment-seeking Latino youth (5–18 years of age). Analyses exploring internal consistency and convergent validity were also conducted. Results Acceptable measurement fit was achieved for the six-factor solution. However, the higher-order model assessing Total HRQOL did not reach acceptable levels, as results found that the Positive Social Attributes (PSA) subscale was not representative of Total HRQOL; internal consistency and convergent validity results also supported this finding. Conclusions The current study provides support for the utility of a modified version of Sizing Them Up, excluding the PSA Scale, as a parent-report measure of obesity-specific HRQOL in treatment-seeking Latino youth with obesity. disparities, obesity, quality of life Childhood obesity remains a major public health concern in the United States, with approximately 17% of children at or above the 95th body mass index (BMI) percentile for age and sex (Ogden, Carroll, Kit, & Flegal, 2014). Although the obesity epidemic affects youth of all ethnic groups, the prevalence of obesity is disproportionately higher among Latino children (21.9% among Latino children compared with 14.7% among non-Latino, White Children; Ogden, Carroll, Fryar, & Flegal, 2015). Early childhood risk factors for this ethnic group include greater odds of rapid weight gain in infancy, elevated risk of maternal restrictive feeding practices, higher intake of sugar-sweetened beverages, and higher intake of fast food during early childhood (Taveras, Gillman, Kleinman, Rich-Edwards, & Rifas-Shiman, 2010). The high prevalence is concerning because childhood obesity is associated with a number of negative physical and psychosocial consequences, including cardiovascular disease, musculoskeletal problems, and behavioral and emotional difficulties throughout childhood and into adulthood (Kalarchian & Marcus, 2012; Reilly & Kelly, 2011; Sanchez-Villegas et al., 2013; Taylor et al., 2006). Furthermore, the association of childhood obesity with type 2 diabetes and fatty liver is disproportionately observed in Latino youth compared with non-Latino White youth (Dabelea et al., 2014; Schwimmer et al., 2006). Although BMI percentile provides some information about a child’s current health status and risk for illness, it does not provide any information about the functional impact of obesity on a child’s life (Modi & Zeller, 2008); therefore, a focus of recent research has been placed on measuring health-related quality of life (HRQOL) in youth with obesity (Zeller & Modi, 2009). HRQOL is a multidimensional construct that assesses a person’s perception of his or her physical, emotional, and psychosocial functioning in relation to his or her health and/or medical condition (Kuyken, 1995). Current research on HRQOL in youth with obesity shows that higher BMI percentile or BMI z-score (zBMI) is associated with lower HRQOL (Buttitta et al., 2014; Ul-Haq et al., 2013) and that HRQOL for children with obesity is similar to children with other chronic diseases (e.g., cancer; Schwimmer, Burwinkle, & Varni, 2003). Examining HRQOL in Latino youth with obesity is important to establish a complete picture of the direct impact of weight on social, emotional, physical, and academic functioning. In the past several decades, there has been a proliferation of various forms and versions of HRQOL measures (Quittner, Cruz, Modi, & Marciel, 2009). HRQOL measures vary based on the reporter (i.e., self-report and parent-proxy report), as well as whether they are more general or illness specific. When measuring HRQOL in children with obesity, parent-proxy reports are important in conjunction with self-report measures because parents’ perceptions of child functioning influence their motivation for initiating services for their children (Modi & Zeller, 2008). Further, including parent proxy measures gives clinicians a more comprehensive picture of the child’s functioning, given that parents are often more accurate observers and reporters of their child’s behavior than children are of their own behavior, and some children and adolescents may be reticent to report issues such as peer victimization related to their size. Specific versions of HRQOL measures (e.g., PedsQL Diabetes Module; Varni et al., 2013) make it possible to understand functioning related specifically to the illness and to measure change because of illness-specific interventions (Modi & Zeller, 2008). When examining physical, emotional, and psychosocial functioning in children with obesity, obesity-specific HRQOL measures can help determine the burden of obesity, responsiveness to change related to weight loss, and progress toward public health goals (CDC, 2011). An example of an obesity-specific parent-proxy HRQOL measure is Sizing Them Up (Modi & Zeller, 2008), which is a measure that has been validated in parents of youth with obesity of age between 5 and 18 years. Sizing Them Up measures parents’ perceptions about their child’s functioning in the context of the child’s weight/size/shape in a variety of domains. Using an exploratory factor analysis (EFA) approach, Modi and Zeller (2008) found that the model that best fit the measure was a model with six subscales (i.e., emotional, physical, teasing/marginalization, positive social attributes (PSA), mealtime challenges, and school functioning). The results of the initial validation showed that the measure had adequate internal consistency (coefficients ranged from 0.52 to 0.90), test–retest reliability (coefficients ranged 0.57 to 0.80), construct validity (the Physical Functioning scale was negatively associated with zBMI), and convergent validity with other HRQOL measures. Additionally, no gender differences were shown, but age was significantly correlated with the scales of Sizing Them Up. The majority of scales were negatively correlated with age, except for the Mealtime Challenges subscale, which was positively correlated with age (Modi & Zeller, 2008). Since its development, Sizing Them Up has been used in a limited number of studies with primarily Caucasian and African-American samples (Black et al., 2014; e.g., Guilfoyle, Zeller, & Modi, 2010; Zeller et al., 2015). Although the prevalence of obesity is disproportionately higher among Latino children, they may be underrepresented in HRQOL research because of the lack of availability of a validated obesity-specific parent-report HRQOL measure in Spanish. Many Latino parents do not speak English fluently. In 2011, 20.84% of individuals in the United States reported speaking a language other than English in the home; 62.22% of these individuals reported speaking Spanish (Ryan, 2013). These non-English-speaking Latino parents have been excluded from analyses in previous research, leaving a gap in knowledge about parent perception of HRQOL in a high-need population (Black et al., 2014). Furthermore, Modi and Zeller (2008) called for future research to examine the measure in ethnically diverse samples of youth with obesity, particularly Latino youth; however, to date, there are no studies identified that have validated a Spanish version of Sizing Them Up. Recently, Tripicchio and colleagues (2017) provided a psychometric evaluation of Sizing Me Up (Zeller & Modi, 2009), the self-report counterpart of Sizing Them Up, among children who are Latino. Results of this confirmatory factor analysis (CFA) demonstrated that Sizing Me Up achieved acceptable fit in a Latino population of treatment-seeking youth with obesity, providing evidence that the English version is appropriate for use with Latino children. However, the psychometric properties of Sizing Them Up among those adults who require written materials in Spanish remain unknown. To fill the gap in the literature on obesity-specific HRQOL measures, the primary aim of the current study was to examine the psychometric properties of the Spanish translation of Sizing Them Up. It was hypothesized that the previous factor structure as identified by Modi and Zeller (2008) would be replicated with the translated measure in this Spanish-speaking sample of Latino parents. Specifically, it was hypothesized that the Spanish translation of the Sizing Them Up measure would retain the six-factor structure. Further, it was hypothesized that the intercorrelations would be positive and range from 0.08 to 0.68, similar to the results found by Modi and Zeller (2008). Additionally, based on the previous study by Modi and Zeller (2008), it was also hypothesized that Cronbach’s α, a measure of internal consistency, for the Total HRQOL scale would be >0.8, but that Cronbach’s α for the subscales would range from 0.5 to 0.9. It was also hypothesized that the Physical Functioning subscale, but no other subscales, of Sizing Them Up would be negatively correlated with zBMI. Regarding child age, it was hypothesized that there would be a negative correlation between all of the scales of Sizing Them Up except for the Mealtime Challenges subscale, which was expected to be positively correlated with age. Finally, consistent with the results from Modi and Zeller (2008), it was expected that gender would not be associated with Sizing Them Up subscales. Methods Participants and Study Procedures Participants were 154 Spanish-speaking parents of children presenting at one of four comprehensive behavioral lifestyle interventions for pediatric obesity across two sites in the Midwest (greater Kansas City area; three programs) and South (Texas; one program). Participants across all sites presented for weight management treatment, including weight management clinics and family-based behavioral group treatments, and all participant data were collected at entry into the program. The current project was approved as a separate multisite IRB project, and all study procedures were approved by the local IRBs, though the data collection process varied based on the site/program. Two programs analyzed clinical data abstracted under an IRB-approved retrospective chart review (N = 72), one program collected the measure through an opt-out consent information sheet procedure (N = 43), and the final program did so through a traditional opt-in consent/assent procedure (N = 39). The interventions served a high percentage of families who self-identified as Latino (with sites ranging from 28.5 to 67%); therefore, parents were offered the opportunity to participate in treatment appointments/sessions and complete treatment measures in either English or Spanish. Parents who self-selected to complete the measure in Spanish were included in the current study. Country of origin and acculturation status of participating families was not assessed; however, providers from each program provided anecdotal evidence that the vast majority of families were of Mexican descent. Additionally, inclusion criteria included parents of children of age 5–18 years with a BMI equal to or above the 95th percentile to replicate the original sample from the initial validation study (Modi & Zeller, 2008). Parents completed the Spanish version of Sizing Them Up at an initial treatment session. Child height and weight were obtained according to standard clinical procedures by trained clinical staff. Spanish Translation Procedures The procedures for translation were aligned with guidelines put forth by the World Health Organization (e.g., forward translation, back translation, pretesting, cognitive interviewing, and final version; World Health Organization, 2016). Sizing Them Up was translated into Spanish by two staff members who were certified as qualified bilingual staff. Both staff members were born in Mexico and are native Spanish speakers. One staff member translated all items independently (forward translation), and the second member back-translated the items (back translation); they worked together to resolve any discrepancies and synthesize the two versions of the translated measure. The translated questionnaire was pilot tested with 64 treatment-seeking parents of children with obesity (pretesting). Concerns arose with regard to five items; these items had been the most difficult to resolve during the back-translation, and further analyses revealed that Spanish-speaking parents scored differently from English-speaking parents on these items. To address the concerns, two different bilingual staff members held a focus group with five Spanish-speaking families (cognitive interviewing; these families were not included in the final sample, as their feedback was used to adjust the translation of the measure). The families generated ideas about how the wording of the five questions could be modified to best reflect the key themes of the questions; this feedback was given to one of the original translators. Changes were then made to the measure after holding an expert panel (consisting of qualified bilingual interpreters, bilingual graduate students in psychology, and experts in weight management) discussion to develop the correct terminology that most accurately reflected the meaning of the original measure. Specifically, changes were made to two items (6 and 18) because the original translations did not capture the connotation of the English words “argue” and “upset”; different Spanish words or phrases were chosen to better capture the meaning (“no estuvo de acuerdo sobre” was chosen over “discutido” to best capture “argue” and “hizo berrinche” was chosen to best capture “upset.”). Similarly, Items 14 and 19 were altered to replace the verbs with those that most accurately captured the ideas of teasing (“se burlaron” was chosen over “molestado”) and physically keeping up with others (“mantaner el mismo nivel de actividad fisica” was chosen over “convivir”). The final item (Number 17) was not changed, as the recommendations were deemed stylistic for verb choice. The current analyses focused on the final version of the translated form (final version). Measures Obesity-Specific HRQOL The Spanish translation of Sizing Them Up (Modi & Zeller, 2008) was used to assess parent-report of their child’s obesity-specific HRQOL. The 22-item measure assessed for child functioning in a variety of domains including emotional functioning (e.g., “Felt sad because of their weight/shape/size.”), physical functioning (e.g., “Had difficulty keeping up with other children because of their weight/shape/size.”), school functioning (e.g., “Chose not to go to school because of their weight/shape/size.”), teasing/marginalization (e.g., “Was teased by others when physically active because of their weight/shape/size.”), positive attributes (e.g., “Seen as having a good sense of humor.”), and mealtime challenges (e.g., “Argued about when, what and how much to eat.”). In addition to domain scores, Sizing Them Up also yields a Total HRQOL score. Parents rated how much their child’s weight/shape/size impacted their child’s functioning in the past 1 month on a scale from 1 (“Never”) to 4 (“Always”); the PSA scale is reverse scored. Sizing Them Up has been shown to have adequate internal consistency (coefficients ranged from 0.52 to 0.90), test–retest reliability (coefficients ranged 0.57–0.80), and construct and convergent validity (Modi & Zeller, 2008; Quittner et al., 2009). Weight Status Child height (in cm) and weight (in kg) were collected by trained research staff, while the children were wearing light clothing and no shoes. BMI percentile and zBMI were calculated based on child age and sex as recommended by the Centers for Disease Control and Prevention (CDC, 2011). Demographics Parent (i.e., sex, caregiver status, weight) and child (i.e., age, sex, race/ethnicity) demographic information was abstracted from the medical or research record. Data Analysis Plan MPlus Version 7 (Muthén & Muthén, 1998-2012) was used for factor analyses. A CFA of the translated Sizing Them Up measure was performed to assess the goodness of fit of the original factor structure (Modi & Zeller, 2008). First, a single-factor model was specified with all 22 items representing a single factor of Total HRQOL (Model 1). Next, Model 2 adjusted for nonsignificant loadings and tested a two-factor model with the PSA subscale as one factor, and the remaining 18 items as a second factor. Model 3 tested the 6-factor solution derived from the original EFA1. Model 4 tested the applicability of a second order factor as a composite of the 6 first order factors. The final model (Model 5) adjusted for the nonsignificant loading of the PSA subscale, and tested PSA as a distinct factor, with the remaining five factors as first-order factors and a second-order Total HRQOL factor. In all models, a robust weighted least squares estimation with a diagonal weight matrix and a mean- and variance-adjusted chi-square test statistic was used to account for the ordinal scale of Sizing Them Up (i.e., four responses along a Likert scale). Other factor analysis methods (e.g., maximum likelihood; commonly used for continuous data) use Pearson correlation matrices and are inappropriate for use with ordinal data (Brown, 2006). The WLSMV estimator also allows for a nonnormal distribution. Also, the marker variable method of identification was used, in which the first indicator of each construct was set to a loading of 1. Of the 154 participants, 115 parents provided complete data, resulting in 25.32% of participants with at least one missing item. Data were determined to be missing completely at random because the Little’s (1988) Test of Missing Data (IBM SPSS) was nonsignificant (Schlomer, Bauman, & Card, 2010); therefore, missing data were handled through the use of multiple imputation, with 100 imputed data sets being used for all CFA analyses. Model fit statistics were assessed to determine the degree to which the hypothesized (implied) model fit the observed data. All models were evaluated by examining the χ2 test of significance, root mean squared error of approximation (RMSEA), comparative fit index (CFI), and Tucker–Lewis Index (TLI). Model fit was considered to be acceptable if the RMSEA values were at or below 0.08, and good/close if values were below 0.05 (Little, 2013). Model fit was considered to be acceptable for CFI and TLI values above 0.90, and good for values above 0.95 (Little, 2013). Individual item and factor loadings were evaluated and considered acceptable if loadings were equal to or greater than 0.4 (Brown, 2006). Scaled scoring of each subscale and the total score were calculated according to the manual put forth by the authors of the original Sizing Them Up (scales are scored on a range of 0–100; 100 = no impairment in functioning; Manual Scoring for Sizing Them Up). Cronbach’s α was calculated using IBM SPSS Statistics Version 22 to evaluate the internal consistency of the subscales and Total HRQOL scale of the Spanish translation of Sizing Them Up. Excellent, good, and acceptable internal consistency was determined using the cutoffs of 0.9, 0.8, and 0.7, respectively (George & Mallery, 2003). Additionally, bivariate correlations between the subscales/total scale of the Spanish translation of Sizing Them Up and zBMI, age, and gender were conducted using IBM SPSS Statistics Version 22. Results Spanish-speaking Latino youth participants (N = 154) were 52.2% female (M age = 10.84; SD = 2.47) with a mean BMI score of 29.11 (SD = 5.68), mean zBMI of 2.23 (SD = 0.31), and BMI percentile of 98.34 (SD = 1.28). Of those with insurance status identified (N = 80), 78.8% had Medicaid, 11.3% had commercial insurance, and 10.0% had no insurance or were self-pay. The vast majority of participating caregivers was female and mothers (97.2%). Of those with caregiver weight status data (N = 56), mean BMI was 33.67 (SD = 5.90). Confirmatory Factor Analyses With all 22 items included, the one-factor model (Model 1) produced a mediocre fit to the data (χ2 = 557.83, n = 154, p < .001, RMSEA = 0.104, CFI = 0.892, TLI = 0.880). Further inspection revealed that items 10, 15, 17, and 20 had factor loadings that were negative or <0.1, and were nonsignificant. These four items represented the originally conceptualized PSA scale, so a two-factor solution was then tested, with the original PSA scale as one factor, and the remaining 18 items as a second factor. Results demonstrated that the two-factor solution (Model 2) produced a mediocre to acceptable fit to the data (χ2 = 413.154, n = 154, p < .001, RMSEA = 0.080, CFI = 0.936, TLI = 0.929), and all indicators loaded significantly with acceptable levels on the respective two scales. Based on this information, Model 2 fit the data better than Model 1; however, additional models were tested to determine if good or close model fit could be achieved. To replicate the factor structure of Sizing Them Up that was identified in the original EFA, the translated 22-item six-factor model was tested next. The residual variance of the School Functioning subscale (consisting of one item) was set to 1, because the residual error variance of categorical variables cannot be estimated. With all 22 items included, the six-factor model (Model 3) produced an acceptable to very good fit to the data (χ2 = 287.112, n = 154, p < .001, RMSEA = 0.055, CFI = 0.971, TLI = 0.966). Further inspection of the indicators revealed no indicator loadings which were <0.4 or nonsignificant; therefore, this model was determined to be an acceptable fit to the data and a better fitting model than Models 1 and 2. Model 3 would be appropriate for use among researchers and clinicians interested in individual domains of HRQOL. To test the utility of a Total HRQOL score for Sizing Them Up, the six-factor solution was then tested under a higher-order framework, in which the six first-order factors were set to comprise a second-order factor, Total HRQOL. This model (Model 4) produced an acceptable to very good fit to the data (χ2 = 339.497, N = 154, p < .001, RMSEA = 0.066, CFI = 0.958, TLI = 0.952), but the PSA scale did not load significantly onto the second-order factor (p = .783); therefore, Model 4 is not recommended for use. A subsequent model with PSA removed from the second-order factor produced no changes in fit indices. Further inspection revealed that the Teasing/Marginalization scale had a standardized negative residual variance of −0.02. Because the negative residual variance was small and nonsignificant, a final model (Model 5) was tested with the residual variance of Teasing/Marginalization set to 0 and the PSA removed from the higher-order factor. This model produced acceptable and significant indicator loadings at the first-order and second-order levels and was an acceptable to very good fit to the data (χ2 = 339.829, n = 154, p < .001, RMSEA = 0.065, CFI = 0.958, TLI = 0.953). Although Model 3 resulted in slightly improved fit statistics compared with Model 5, Model 5 demonstrates that all subscales, except for PSA, are representative of a unified construct (i.e., Total Quality of Life). Given both the clinical utility of a total score and the theory that Total Quality of Life is composed of multiple domains (e.g., emotional, physical, school, teasing/marginalization, positive attributes, and mealtime challenges), we believe that Model 5 (with all subscales except PSA as first-order factors and total QOL as a second-order factor) is also a theoretically sound model and would be appropriate for use. Model fit statistics for Model 5 are presented in Table I, and factor loadings and standard errors are presented in Table II. Table I. Fit Statistics for CFA Model χ2 df p RMSEA CFI TLI One-factor (Model 1) 557.83 209 .00 0.10 0.89 0.88 Two-factor (Model 2) 413.15 208 .00 0.08 0.94 0.93 Six-factor (Model 3) 287.11 195 .00 0.06 0.97 0.97 Second-order factor (Model 4) 339.50 204 .00 0.07 0.96 0.95 Second-order factora (Model 5) 339.83 205 .00 0.07 0.96 0.95 Model χ2 df p RMSEA CFI TLI One-factor (Model 1) 557.83 209 .00 0.10 0.89 0.88 Two-factor (Model 2) 413.15 208 .00 0.08 0.94 0.93 Six-factor (Model 3) 287.11 195 .00 0.06 0.97 0.97 Second-order factor (Model 4) 339.50 204 .00 0.07 0.96 0.95 Second-order factora (Model 5) 339.83 205 .00 0.07 0.96 0.95 a Second-order factor structure excludes the Positive Social Attributes subscale; definitions: CFA = confirmatory factor analyses; CFI = comparative fit index; χ2 = chi-square; df = degrees of freedom; RMSEA = root mean squared error of approximation; TLI = Tucker–Lewis Index. Table I. Fit Statistics for CFA Model χ2 df p RMSEA CFI TLI One-factor (Model 1) 557.83 209 .00 0.10 0.89 0.88 Two-factor (Model 2) 413.15 208 .00 0.08 0.94 0.93 Six-factor (Model 3) 287.11 195 .00 0.06 0.97 0.97 Second-order factor (Model 4) 339.50 204 .00 0.07 0.96 0.95 Second-order factora (Model 5) 339.83 205 .00 0.07 0.96 0.95 Model χ2 df p RMSEA CFI TLI One-factor (Model 1) 557.83 209 .00 0.10 0.89 0.88 Two-factor (Model 2) 413.15 208 .00 0.08 0.94 0.93 Six-factor (Model 3) 287.11 195 .00 0.06 0.97 0.97 Second-order factor (Model 4) 339.50 204 .00 0.07 0.96 0.95 Second-order factora (Model 5) 339.83 205 .00 0.07 0.96 0.95 a Second-order factor structure excludes the Positive Social Attributes subscale; definitions: CFA = confirmatory factor analyses; CFI = comparative fit index; χ2 = chi-square; df = degrees of freedom; RMSEA = root mean squared error of approximation; TLI = Tucker–Lewis Index. Table II. Factor Loadings and Standard Errors for Indicators of Final Models Latent Factor and Indicators Loading Emotional functioning  Q4: Felt sad 0.89 (0.02)*  Q8: Felt frustrated 0.90 (0.03)*  Q9: Avoided dressing in front of others 0.69 (0.05)*  Q11: Felt worried 0.73 (0.04)*  Q13: Felt mad 0.91 (0.02)*  Q16: Felt concerned 0.76 (0.04)*  Q22: Had low self-esteem 0.82 (0.04)* Physical functioning  Q1: Difficulty participating in physical activities 0.90 (0.03)*  Q5: Changes to physical surroundings 0.76 (0.06)*  Q7: Chose not to participate in gym 0.77 (0.08)*  Q19: Difficulty keeping up with other childrena 0.68 (0.07)*  Q21: Became out of breath 0.68 (0.06)* Teasing/marginalization  Q2: Teased by others 0.86 (0.03)*  Q12: Felt left out 0.88 (0.04)*  Q14: Teased when physically activea 0.86 (0.04)* Positive Social Attributes  Q10: Kept body clean 0.46 (0.11)*  Q15: Good sense of humor 0.67 (0.08)*  Q17: Perceived as healthya 0.66 (0.09)*  Q20: Felt successful in daily activities 0.62 (0.09)* Mealtime challenges  Q6: Argued about eatinga 0.61 (0.09)*  Q18: Upset at mealtimesa 0.64 (0.09)* School functioning  Q3: Chose not to go to school 1.00 (0.00) Higher-order  Emotional Functioning 0.86 (0.03)*  Physical Functioning 0.84 (.03)  Teasing/Marginalization 1.0 (0.0)b  Positive Social Attributes R  Mealtime Challenges 0.79 (0.11)*  School Functioning 0.60 (0.03)* Latent Factor and Indicators Loading Emotional functioning  Q4: Felt sad 0.89 (0.02)*  Q8: Felt frustrated 0.90 (0.03)*  Q9: Avoided dressing in front of others 0.69 (0.05)*  Q11: Felt worried 0.73 (0.04)*  Q13: Felt mad 0.91 (0.02)*  Q16: Felt concerned 0.76 (0.04)*  Q22: Had low self-esteem 0.82 (0.04)* Physical functioning  Q1: Difficulty participating in physical activities 0.90 (0.03)*  Q5: Changes to physical surroundings 0.76 (0.06)*  Q7: Chose not to participate in gym 0.77 (0.08)*  Q19: Difficulty keeping up with other childrena 0.68 (0.07)*  Q21: Became out of breath 0.68 (0.06)* Teasing/marginalization  Q2: Teased by others 0.86 (0.03)*  Q12: Felt left out 0.88 (0.04)*  Q14: Teased when physically activea 0.86 (0.04)* Positive Social Attributes  Q10: Kept body clean 0.46 (0.11)*  Q15: Good sense of humor 0.67 (0.08)*  Q17: Perceived as healthya 0.66 (0.09)*  Q20: Felt successful in daily activities 0.62 (0.09)* Mealtime challenges  Q6: Argued about eatinga 0.61 (0.09)*  Q18: Upset at mealtimesa 0.64 (0.09)* School functioning  Q3: Chose not to go to school 1.00 (0.00) Higher-order  Emotional Functioning 0.86 (0.03)*  Physical Functioning 0.84 (.03)  Teasing/Marginalization 1.0 (0.0)b  Positive Social Attributes R  Mealtime Challenges 0.79 (0.11)*  School Functioning 0.60 (0.03)* a Discrepancies during translation; bResidual variance set to 0; R = Removed; *p ≤ .001. Table II. Factor Loadings and Standard Errors for Indicators of Final Models Latent Factor and Indicators Loading Emotional functioning  Q4: Felt sad 0.89 (0.02)*  Q8: Felt frustrated 0.90 (0.03)*  Q9: Avoided dressing in front of others 0.69 (0.05)*  Q11: Felt worried 0.73 (0.04)*  Q13: Felt mad 0.91 (0.02)*  Q16: Felt concerned 0.76 (0.04)*  Q22: Had low self-esteem 0.82 (0.04)* Physical functioning  Q1: Difficulty participating in physical activities 0.90 (0.03)*  Q5: Changes to physical surroundings 0.76 (0.06)*  Q7: Chose not to participate in gym 0.77 (0.08)*  Q19: Difficulty keeping up with other childrena 0.68 (0.07)*  Q21: Became out of breath 0.68 (0.06)* Teasing/marginalization  Q2: Teased by others 0.86 (0.03)*  Q12: Felt left out 0.88 (0.04)*  Q14: Teased when physically activea 0.86 (0.04)* Positive Social Attributes  Q10: Kept body clean 0.46 (0.11)*  Q15: Good sense of humor 0.67 (0.08)*  Q17: Perceived as healthya 0.66 (0.09)*  Q20: Felt successful in daily activities 0.62 (0.09)* Mealtime challenges  Q6: Argued about eatinga 0.61 (0.09)*  Q18: Upset at mealtimesa 0.64 (0.09)* School functioning  Q3: Chose not to go to school 1.00 (0.00) Higher-order  Emotional Functioning 0.86 (0.03)*  Physical Functioning 0.84 (.03)  Teasing/Marginalization 1.0 (0.0)b  Positive Social Attributes R  Mealtime Challenges 0.79 (0.11)*  School Functioning 0.60 (0.03)* Latent Factor and Indicators Loading Emotional functioning  Q4: Felt sad 0.89 (0.02)*  Q8: Felt frustrated 0.90 (0.03)*  Q9: Avoided dressing in front of others 0.69 (0.05)*  Q11: Felt worried 0.73 (0.04)*  Q13: Felt mad 0.91 (0.02)*  Q16: Felt concerned 0.76 (0.04)*  Q22: Had low self-esteem 0.82 (0.04)* Physical functioning  Q1: Difficulty participating in physical activities 0.90 (0.03)*  Q5: Changes to physical surroundings 0.76 (0.06)*  Q7: Chose not to participate in gym 0.77 (0.08)*  Q19: Difficulty keeping up with other childrena 0.68 (0.07)*  Q21: Became out of breath 0.68 (0.06)* Teasing/marginalization  Q2: Teased by others 0.86 (0.03)*  Q12: Felt left out 0.88 (0.04)*  Q14: Teased when physically activea 0.86 (0.04)* Positive Social Attributes  Q10: Kept body clean 0.46 (0.11)*  Q15: Good sense of humor 0.67 (0.08)*  Q17: Perceived as healthya 0.66 (0.09)*  Q20: Felt successful in daily activities 0.62 (0.09)* Mealtime challenges  Q6: Argued about eatinga 0.61 (0.09)*  Q18: Upset at mealtimesa 0.64 (0.09)* School functioning  Q3: Chose not to go to school 1.00 (0.00) Higher-order  Emotional Functioning 0.86 (0.03)*  Physical Functioning 0.84 (.03)  Teasing/Marginalization 1.0 (0.0)b  Positive Social Attributes R  Mealtime Challenges 0.79 (0.11)*  School Functioning 0.60 (0.03)* a Discrepancies during translation; bResidual variance set to 0; R = Removed; *p ≤ .001. Latent Factor Intercorrelations Latent factor intercorrelations between individual subscale factors were also examined. The latent factors of Emotional Functioning, Physical Functioning, Teasing/Marginalization, Mealtime Behaviors, and School Functioning evidenced good latent intercorrelations (ranged from 0.39 to 0.89), with the majority of values >0.70. However, the PSA scale was not significantly correlated with Emotional Functioning, Teasing/Marginalization, or School Functioning, and evidenced significant correlations (at the p < .01 and p < .05 levels, respectively) with Physical Functioning and Mealtime Challenges in the unexpected direction (i.e., better PSA was associated with worse physical functioning and more mealtime challenges). Latent factor intercorrelations are presented in Table III. Table III. Standardized Interfactor Correlations of the Final Lower-Order Model Latent Factors Emotional Functioning Physical Functioning Teasing/Marginalization Positive Social Attributes Mealtime Challenges School Functioning Emotional Functioning 1.0 Physical Functioning 0.71 (0.05)*** 1.0 Teasing/Marginalization 0.85 (0.04)*** 0.85 (0.04)*** 1.0 Positive Social Attributes −0.14 (0.10)NS 0.19 (0.10)* 0.04 (0.10)NS 1.0 Mealtime Challenges 0.65 (0.08)*** 0.85 (0.09)*** 0.67 (0.10)*** 0.39 (0.13)** 1.0 School Functioning 0.71 (0.08)*** 0.45 (0.08)*** 0.84 (0.07)*** 0.02 (0.16)NS 0.39 (0.13)** 1.0 Latent Factors Emotional Functioning Physical Functioning Teasing/Marginalization Positive Social Attributes Mealtime Challenges School Functioning Emotional Functioning 1.0 Physical Functioning 0.71 (0.05)*** 1.0 Teasing/Marginalization 0.85 (0.04)*** 0.85 (0.04)*** 1.0 Positive Social Attributes −0.14 (0.10)NS 0.19 (0.10)* 0.04 (0.10)NS 1.0 Mealtime Challenges 0.65 (0.08)*** 0.85 (0.09)*** 0.67 (0.10)*** 0.39 (0.13)** 1.0 School Functioning 0.71 (0.08)*** 0.45 (0.08)*** 0.84 (0.07)*** 0.02 (0.16)NS 0.39 (0.13)** 1.0 *** p ≤ .001; **p ≤ .01; *p ≤ .05; NS = Not significant. Table III. Standardized Interfactor Correlations of the Final Lower-Order Model Latent Factors Emotional Functioning Physical Functioning Teasing/Marginalization Positive Social Attributes Mealtime Challenges School Functioning Emotional Functioning 1.0 Physical Functioning 0.71 (0.05)*** 1.0 Teasing/Marginalization 0.85 (0.04)*** 0.85 (0.04)*** 1.0 Positive Social Attributes −0.14 (0.10)NS 0.19 (0.10)* 0.04 (0.10)NS 1.0 Mealtime Challenges 0.65 (0.08)*** 0.85 (0.09)*** 0.67 (0.10)*** 0.39 (0.13)** 1.0 School Functioning 0.71 (0.08)*** 0.45 (0.08)*** 0.84 (0.07)*** 0.02 (0.16)NS 0.39 (0.13)** 1.0 Latent Factors Emotional Functioning Physical Functioning Teasing/Marginalization Positive Social Attributes Mealtime Challenges School Functioning Emotional Functioning 1.0 Physical Functioning 0.71 (0.05)*** 1.0 Teasing/Marginalization 0.85 (0.04)*** 0.85 (0.04)*** 1.0 Positive Social Attributes −0.14 (0.10)NS 0.19 (0.10)* 0.04 (0.10)NS 1.0 Mealtime Challenges 0.65 (0.08)*** 0.85 (0.09)*** 0.67 (0.10)*** 0.39 (0.13)** 1.0 School Functioning 0.71 (0.08)*** 0.45 (0.08)*** 0.84 (0.07)*** 0.02 (0.16)NS 0.39 (0.13)** 1.0 *** p ≤ .001; **p ≤ .01; *p ≤ .05; NS = Not significant. Internal Consistency Internal consistency results are presented in Table IV. The Emotional Functioning subscale and Total HRQOL evidenced good internal consistency, and Physical Functioning and Teasing/Marginalization scales produced acceptable values. The PSA scale demonstrated an internal consistency estimate (α = 0.61) similar to that documented in the original publication (α = 0.59, Modi & Zeller, 2008). The Mealtime challenges subscale evidenced poor internal consistency, with a value (α = 0.48) lower than that documented in the original publication. Table IV. Ms (SD) and Reliability Coefficients Scored Subscales and Total Score M (SD) Cronbach’s α Emotional Functioning 72.84 (22.82) 0.88 Physical Functioning 82.26 (17.90) 0.74 Teasing/Marginalization 80.00 (22.95) 0.77 Positive Social Attributes 61.88 (22.70) 0.61 Mealtime Challenges 68.84 (24.53) 0.48 School Functioninga 94.78 (16.28) – Total Quality of Life 74.60 (15.07) 0.87 Modified Total Quality of Life 77.42 (17.69) 0.91 Scored Subscales and Total Score M (SD) Cronbach’s α Emotional Functioning 72.84 (22.82) 0.88 Physical Functioning 82.26 (17.90) 0.74 Teasing/Marginalization 80.00 (22.95) 0.77 Positive Social Attributes 61.88 (22.70) 0.61 Mealtime Challenges 68.84 (24.53) 0.48 School Functioninga 94.78 (16.28) – Total Quality of Life 74.60 (15.07) 0.87 Modified Total Quality of Life 77.42 (17.69) 0.91 a Scale contains only one item. Table IV. Ms (SD) and Reliability Coefficients Scored Subscales and Total Score M (SD) Cronbach’s α Emotional Functioning 72.84 (22.82) 0.88 Physical Functioning 82.26 (17.90) 0.74 Teasing/Marginalization 80.00 (22.95) 0.77 Positive Social Attributes 61.88 (22.70) 0.61 Mealtime Challenges 68.84 (24.53) 0.48 School Functioninga 94.78 (16.28) – Total Quality of Life 74.60 (15.07) 0.87 Modified Total Quality of Life 77.42 (17.69) 0.91 Scored Subscales and Total Score M (SD) Cronbach’s α Emotional Functioning 72.84 (22.82) 0.88 Physical Functioning 82.26 (17.90) 0.74 Teasing/Marginalization 80.00 (22.95) 0.77 Positive Social Attributes 61.88 (22.70) 0.61 Mealtime Challenges 68.84 (24.53) 0.48 School Functioninga 94.78 (16.28) – Total Quality of Life 74.60 (15.07) 0.87 Modified Total Quality of Life 77.42 (17.69) 0.91 a Scale contains only one item. Correlations The correlations between zBMI and Sizing Them Up subscales and Total HRQOL score were conducted to determine convergent validity. Significant correlations were established between zBMI and the Physical Functioning scale (r = −.24, p < .05), the Teasing/Marginalization scale (r = −.22, p < .05), the Mealtime Challenges scale (r = −.22, p < .05), and Total HRQOL (r = −.23, p < .05). No significant correlations were found between zBMI and the Emotional Functioning, PSA, or the School Functioning subscales. Gender and age were not significantly associated with any scale scores or the Total HRQOL score. Discussion The purpose of the current study was to examine the psychometric properties of a Spanish translation of Sizing Them Up, a parent-report measure of obesity-specific HRQOL. Pediatric obesity disproportionally affects Latino youth, and previous research has indicated that HRQOL is a key measure in pediatric obesity (e.g., Buttitta et al., 2014; Schwimmer et al., 2003; Ul-Haq et al., 2013). Yet, there are no identified Spanish language parent-report measures of weight-related HRQOL. The current study extends the literature by examining the psychometric properties of the Spanish translation of Sizing Them Up. The first hypothesis, that the previously established six-factor structure would be replicated, was partially supported. The results support the use of the Spanish translation of Sizing Them Up among Spanish-speaking parents of treatment-seeking youth with obesity if the six factors are interpreted independently. However, under a structural equation modeling (SEM) framework, results do not support the use of a Total HRQOL score with this translated version of the measure, because the PSA subscale was not found to be representative of Total HRQOL, but instead was a separate factor. This finding was supported by the nonsignificant loading of the PSA factor in the model testing for a second-order factor. These findings suggest that the PSA scale from Sizing Them Up likely functions differently for the Latino population. This scale contains four items, including “kept their body clean and fresh,” “seen as having a good sense of humor,” “perceived as healthy by others,” and “felt successful in daily activities.” Recent research indicates that Latino families do not associate pediatric obesity with poor health in the same way that non-Latino Black and non-Latino Caucasian families do, and in fact interpret overweight to be an indicator of good health (Baker & Altman, 2015). For example, research suggests that Latino mothers perceive thinness to be an indicator of malnutrition, and thus see thinness as more dangerous than overweight/obesity for their children (Crawford et al., 2004). Thus, Latino families may perceive their children with obesity to display more positive attributes than do Black and Caucasian families. Second, it is possible that the overall construct of PSA is perceived differently among Latino families, above and beyond the specific items included in the subscale. More research is needed to learn more about how culture may affect this subscale. The second hypothesis, regarding the latent factor intercorrelations, was partially supported. As expected, the majority of the intercorrelations were medium to large positive correlations. However, the PSA scale was not significantly correlated with Emotional Functioning, Teasing/Marginalization, or School Functioning. The PSA subscale also had a negative correlation with the Physical Functioning and Mealtime Challenges subscales. The lack of significant correlations between PSA and several other subscales is similar to results from the psychometric evaluation of the self-report measure of obesity-specific HRQOL, Sizing Me Up, in which Tripicchio and colleagues (2017) found the PSA subscale to be uncorrelated with the Physical Functioning, Teasing/marginalization, and Social Avoidance Subscales. However, the negative correlations in the current study stand in contrast to results from the original validation study (Modi & Zeller, 2008), although in the original validation study, the PSA scale demonstrated a nonsignificant correlation with the school functioning subscale and loadings <0.3 with all other subscales. These findings provide further evidence that the PSA subscale was not found to be representative of Total HRQOL in this population, for both parent- and self-report. The third hypothesis, that Cronbach’s α for Total HRQOL would be >0.8 and for the subscales would range from 0.5 to 0.9, was partially supported. The findings from the current study were consistent with the previous study by Modi and Zeller (2008) for all scales except for the Mealtime Challenges subscale. The poor internal consistency for the Mealtime Challenges scale was not unexpected given the difficulties with the translation of both items, which comprise this scale. As discussed previously, the translation of the words for “argued” and “upset” proved challenging, due at least in part to regional variations on the proper translation of these words into Spanish. While our team was unable to evaluate country of origin for the parents’ Spanish influence, it is well recognized that there are regional variations in dialect and in word connotations, especially with emotionally laden terms. Future researchers would be wise to consider evaluating this subscale further, to identify more tailored phrases. Additionally, although the findings for the internal consistency for the PSA subscale in the current study were consistent with the previous study (α = 0.61 in current sample; α = 0.59 in Modi & Zeller, 2008), the Cronbach’s α is considered below acceptable (George & Mallery, 2003). Therefore, researchers and clinicians should use caution when using the Mealtime Challenges and PSA subscales of the Spanish translation of Sizing Them Up because the items on the subscales may not be adequately measuring the same underlying constructs. Although the Cronbach’s α for Total HRQOL reached acceptable levels, this is unexpected given the CFA results. The use of a Total HRQOL score that includes PSA is cautioned against, because SEM accounts for error and is a more acceptable estimator of reliability. The correlation hypotheses were also partially supported. Although it was only hypothesized that the Physical Functioning scale would be associated with zBMI based on the previous study, the current study found significant correlations between zBMI and the Physical Functioning scale, the Teasing/Marginalization scale, the Mealtime Challenges scale, and Total HRQOL; these correlations provide evidence for convergent validity. It should also be noted that although the Teasing/Marginalization scale was highly correlated with the School Functioning and Emotional Functioning scales, these do appear to be distinct constructs as only the Teasing/Marginalization scale was significantly associated with BMIz. Contrary to findings from Modi and Zeller (2008), age was not significantly associated with any scale scores. Modi and Zeller (2008) posited that self-esteem worsens as youth transition to adolescence, and self-esteem likely influenced the Emotional Functioning and PSA subscales. Given our findings, it is possible that weight status may not be as related to self-esteem among Latino youth as for those from other ethnic backgrounds. Consistent with previous findings, gender was not associated with the scale scores in the current sample. Thus, the Spanish translation may have utility for both sexes and all age groups. Limitations of the current study must be considered when interpreting findings. For example, the sample for the current analysis was relatively small, represented a predominantly low-income population, and included predominately female (mother) reporters, which may limit the generalizability of the results. Further, across all subscales and the Total HRQOL score, the mean scores for Sizing Them Up were higher in this population than in the original validation study. Reasons for these discrepancies may be related to the cultural differences in interpretation of overweight/obesity mentioned earlier, namely, the perception that excess weight is an indicator of health, and a perception that thinness is an indicator of malnutrition (Baker & Altman, 2015; Crawford et al., 2004). These perceptions could certainly impact reported HRQOL. These results may also be explained by a lower level of illness in this sample as compared with the original validation sample, as evidenced by participants presenting with a lower BMIz (2.23) than those from the original validation sample (2.60). It is likely that those in the current study experienced less impairment because of obesity given their relatively lower BMIz at baseline; therefore, the current results may not be generalizable to youth with higher BMIz. Future studies should examine psychometric properties in various Latino samples, particularly among Latino youth who are presenting to bariatric surgery, to comprehensively evaluate the psychometric properties of this scale, and to provide further evidence of the utility of the measure. Additionally, because of limited sample size, the current study did not conduct posttreatment analyses, and future research should evaluate additional measures of reliability and validity, including test–retest reliability and invariance testing. Multiple-group invariance testing between Spanish-speaking and English-speaking Latino parents is needed to completely determine whether the current translated measure is truly capturing the same constructs. Owing to insufficient sample size, the current study did not evaluate a translated version of the Adolescent Developmental Adaptation scale, which is a subscale of Sizing Them Up intended only for adolescent participants. Finally, the current study did not measure country of origin, acculturation status, or parent education level of participants, which would be have been beneficial in ensuring the measure’s applicability to Spanish speakers from a variety of backgrounds. Despite these limitations, the current study has many strengths. For example, the current study provides the first Spanish translation of a parent-report of obesity-specific HRQOL. The current study included the use of well-established translation procedures for creating the Spanish translation of the measure. Additionally, the current study drew from two geographical locations with high populations of Latino youth and Spanish-speaking parents, increasing generalizability of results. Overall, results support the use of the Spanish translation of Sizing Them Up using the six factors independently in an SEM framework, which may be appropriate for researchers interested in individual subscales. However, researchers and clinicians are cautioned against computing a Total HRQOL score with all six subscales included. Results do support the use of a Total HRQOL score of the remaining five factors (Emotional Functioning, Physical Functioning, Teasing/Marginalization, Mealtime Challenges, and School Functioning). We suggest that providers wishing to use a Total HRQOL score for this measure adopt a modified formula from the original formula for Sizing Them Up (Original Total HRQOL = ((Sum of 22 Items−22)/66) * 100; Manual for Scoring Sizing Them Up) for scoring without the PSA subscale, in which Modified Total HRQOL = ((Sum of 18 Items−18)/54) * 100. While the scores derived from this modified formula demonstrated a high correlation (r = .96) with those using the original 22-item formula, further testing of the validity of the formula is warranted for future research. Further, given the clear divergence between the PSA Scale and all other Sizing Them Up scales, it is important for future research to better understand the applicability of the PSA scale for all populations. Thus, it is recommended that future studies conduct a CFA of Sizing Them Up in a large representative sample of English-speaking parents of youth with obesity. Additionally, if a CFA confirms the appropriate inclusion of PSA in the English version of Sizing Them Up, additional field testing (e.g., focus groups) may be useful to better understand why the interpretation of PSA may be different in this population. Funding This work was supported by the Michael and Susan Dell Foundation; Health Care Foundation of Greater Kansas City; and the Junior League of Kansas City. Conflicts of interest: None declared. Footnotes 1 Although it is common practice for subscales to include at least three items, the authors chose to include the one-item School Functioning and the two-item Mealtime Challenges subscales to replicate the original English version of the measure validated through exploratory factor analysis. References Baker E. H. , Altman C. E. ( 2015 ). Maternal ratings of child health and child obesity, variations by mother’s race/ethnicity and nativity . Maternal and Child Health Journal , 19 , 1000 – 1009 . Google Scholar Crossref Search ADS PubMed Black W. R. , Davis A. M. , Gillette M. L. , Short M. B. , Wetterneck C. T. , He J. ( 2014 ). Health-related quality of life in obese and overweight, treatment-seeking youth . Ethnicity and Disease , 24 , 321 – 327 . Google Scholar PubMed Brown T. A. ( 2006 ). Confirmatory factor analysis for applied research . New York, NY : The Guilford Press . Buttitta M. , Iliescu C. , Rousseau A. , Guerrien A. ( 2014 ). Quality of life in overweight and obese children and adolescents: A literature review . Quality of Life Research , 23 , 1117 – 1139 . Google Scholar Crossref Search ADS PubMed Centers for Disease Control and Prevention (CDC) . ( 2011 ). Health-related quality of life . Retrieved from http://www.cdc.gov/hrqol/concept.htm Crawford P. B. , Gosliner W. , Anderson C. , Strode P. , Becerra-Jones Y. , Samuels S. , Carroll A. M. , Ritchie L. D. ( 2004 ). Counseling Latina mothers of preschool children about weight issues: Suggestions for a new framework . Journal of the American Dietetic Association , 104 , 387 – 394 . Google Scholar Crossref Search ADS PubMed Dabelea D. , Mayer-Davis E. J. , Saydah S. , Imperatore G. , Linder B. , Divers J. , Bell R. , Badaru A. , Talton J. W. , Crume T. , Liese A. D. , Merchant A. T. , Lawrence J. M. , Reynolds K. , Dolan L , Liu LL. , Hamman R. F. ; SEARCH for Diabetes in Youth Study . ( 2014 ). Prevalence of type 1 and type 2 diabetes among children and adolescents from 2001 to 2009 . Journal of the American Medical Association , 311 , 1778 – 1786 . Google Scholar Crossref Search ADS PubMed George D. , Mallery P. ( 2003 ). Reliability analysis . SPSS for Windows, step by step: A simple guide and reference, 14th ed. Boston : Allyn & Bacon , 222 – 232 . Guilfoyle S. M. , Zeller M. H. , Modi A. C. ( 2010 ). Parenting stress impacts obesity-specific health-related quality of life in a pediatric obesity treatment-seeking sample . Journal of Developmental and Behavioral Pediatrics , 31 , 17 . Google Scholar Crossref Search ADS PubMed Kalarchian M. A. , Marcus M. D. ( 2012 ). Psychiatric comorbidity of childhood obesity . International Review of Psychiatry , 24 , 241 – 246 . Google Scholar Crossref Search ADS PubMed Kuyken W. ( 1995 ). The World Health Organization Quality of Life Assessment (WHOQOL): Position paper from the world health organization . Social Science and Medicine , 41 , 1403 . Google Scholar Crossref Search ADS PubMed Little R. J. A. ( 1988 ). A test of missing completely at random for multivariate data with missing values . Journal of the American Statistical Association , 83 , 1198 – 1202 . Google Scholar Crossref Search ADS Little T. D. ( 2013 ). Longitudinal structural equation modeling . New York, NY : The Guilford Press . Muthén L. K. , Muthén B. O. ( 1998 -2012). Mplus user's guide ( 7th ed. ). Los Angeles, CA : Muthén & Muthén . Modi A. C. , Zeller M. H. ( 2008 ). Validation of a parent-proxy, obesity-specific quality-of-life measure: Sizing Them Up . Obesity , 16 , 2624 – 2633 . Google Scholar Crossref Search ADS PubMed Ogden C. L. , Carroll M. D. , Fryar C. D. , Flegal K. M. ( 2015 ). Prevalence of obesity among adults and youth: United States, 2011-2014 . NCHS Data Brief , 219 , 1 – 8 . Ogden C. L. , Carroll M. D. , Kit B. K. , Flegal K. M. ( 2014 ). Prevalence of childhood and adult obesity in the United States, 2011-2012 . Journal of the American Medical Association , 311 , 806 – 814 . Google Scholar Crossref Search ADS PubMed Quittner A. L. , Cruz I. , Modi A. C. , Marciel K. K. ( 2009 ). Health-related quality of life instruments for adolescents with chronic diseases. In O’Donohue W. T. , Woodward L. W. (Eds.), Behavioral approaches to chronic disease in adolescence: A guide to integrative care . (pp. 311 – 327 ). New York, NY : Springer Science + Business Media . Reilly J. J. , Kelly J. ( 2011 ). Long-term impact of overweight and obesity in childhood and adolescence on morbidity and premature mortality in adulthood: Systematic review . International Journal of Obesity , 35 , 891 – 898 . Google Scholar Crossref Search ADS PubMed Ryan C. ( 2013 ). Language use in the United States: 2011 . Washington, DC : U.S. Census Bureau . Sanchez-Villegas A. , Field A. E. , O'Reilly E. J. , Fava M. , Gortmaker S. , Kawachi I. , Ascherio A. ( 2013 ). Perceived and actual obesity in childhood and adolescence and risk of adult depression . Journal of Epidemiology and Community Health , 67 , 81 – 86 . Google Scholar Crossref Search ADS PubMed Schlomer G. L. , Bauman S. , Card N. A. ( 2010 ). Best practices for missing data management in counseling psychology . Journal of Counseling Psychology , 57 , 1 – 10 . Google Scholar Crossref Search ADS PubMed Schwimmer J. B. , Burwinkle T. M. , Varni J. W. ( 2003 ). Health-related quality of life of severely obese children and adolescents . JAMA: Journal of the American Medical Association , 289 , 1813 – 1819 . Google Scholar Crossref Search ADS Schwimmer J. B. , Deutsch R. , Kahen T. , Lavine J. E. , Stanley C. , Behling C. ( 2006 ). Prevalence of fatty liver in children and adolescents . Pediatrics , 118 , 1388 – 1393 . Google Scholar Crossref Search ADS PubMed Taveras E. M. , Gillman M. W. , Kleinman K. , Rich-Edwards J. W. , Rifas-Shiman S. L. ( 2010 ). Racial/ethnic differences in early-life risk factors for childhood obesity . Pediatrics , 125 , 686 – 695 . Google Scholar Crossref Search ADS PubMed Taylor E. D. , Theim K. R. , Mirch M. C. , Ghorbani S. , Tanofsky-Kraff M. , Adler-Wailes D. C. , Brady S. , Reynolds J. C. , Calis K. A. , Yanovski J. A. ( 2006 ). Orthopedic complications of overweight in children and adolescents . Pediatrics , 117 , 2167 – 2174 . Google Scholar Crossref Search ADS PubMed Tripicchio G. L. , Borner K. B. , Odar Stough C. , Poppert Cordts K. , Dreyer Gillette M. , Davis A. M. ( 2017 ). Confirmatory factor analysis of Sizing Me Up: Validation of an obesity-specific health-related quality of life measure in Latino youth . Journal of Pediatric Psychology , 42 , 457 – 465 . Google Scholar PubMed Ul-Haq Z. , Mackay D. F. , Fenwick E. , Pell J. P. ( 2013 ). Meta-analysis of the association between body mass index and health-related quality of life among children and adolescents, assessed using the pediatric quality of life inventory index . The Journal of Pediatrics , 162 , 280 – 286 . Google Scholar Crossref Search ADS PubMed Varni J. W. , Curtis B. H. , Abetz L. N. , Lasch K. E. , Piault E. C. , Zeytoonjian A. A. ( 2013 ). Content validity of the PedsQL™ 3.2 diabetes module in newly diagnosed patients with type 1 diabetes mellitus ages 8–45 . Quality of Life Research , 22 , 2169 – 2181 . Google Scholar Crossref Search ADS PubMed World Health Organization . ( 2016 ). Process of translation and adaptation of instruments . Retrieved from http://www.who.int/substance_abuse/research_tools/translation/en/ Zeller M. H. , Modi A. C. ( 2009 ). Development and initial validation of an obesity-specific quality-of-life measure for children: Sizing Me Up . Obesity , 17 , 1171 – 1177 . Google Scholar Crossref Search ADS PubMed Zeller M. H. , Noll J. G. , Sarwer D. B. , Reiter-Purtill J. , Rofey D. L. , Baughcum A. E. , Peugh J. , Courcoulas A. P. , Michalsky M. P. , Jenkins T. M. , Becnel J. N. ; TeenView Study Group and in Cooperation With Teen-LABS Consortium . ( 2015 ). Child maltreatment and the adolescent patient with severe obesity: Implications for clinical care . Journal of Pediatric Psychology , 40 , 640 – 648 . Google Scholar Crossref Search ADS PubMed © The Author(s) 2018. Published by Oxford University Press on behalf of the Society of Pediatric Psychology. All rights reserved. For permissions, please e-mail: journals.permissions@oup.com This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/about_us/legal/notices)

Journal

Journal of Pediatric PsychologyOxford University Press

Published: Oct 1, 2018

References

You’re reading a free preview. Subscribe to read the entire article.


DeepDyve is your
personal research library

It’s your single place to instantly
discover and read the research
that matters to you.

Enjoy affordable access to
over 18 million articles from more than
15,000 peer-reviewed journals.

All for just $49/month

Explore the DeepDyve Library

Search

Query the DeepDyve database, plus search all of PubMed and Google Scholar seamlessly

Organize

Save any article or search result from DeepDyve, PubMed, and Google Scholar... all in one place.

Access

Get unlimited, online access to over 18 million full-text articles from more than 15,000 scientific journals.

Your journals are on DeepDyve

Read from thousands of the leading scholarly journals from SpringerNature, Elsevier, Wiley-Blackwell, Oxford University Press and more.

All the latest content is available, no embargo periods.

See the journals in your area

DeepDyve

Freelancer

DeepDyve

Pro

Price

FREE

$49/month
$360/year

Save searches from
Google Scholar,
PubMed

Create lists to
organize your research

Export lists, citations

Read DeepDyve articles

Abstract access only

Unlimited access to over
18 million full-text articles

Print

20 pages / month

PDF Discount

20% off