Abstract
Purpose
Health-related quality of life (HRQOL) is a critical diabetes outcome, yet differences between youth and parent-proxy ratings can make interpretation difficult. This study aims to explore potential differences between self- and parent-reports of Pediatric Quality of Life Inventory (PedsQL) scores from youth with type 1 (T1D) or type 2 diabetes (T2D) and to evaluate associations between discrepancies, PedsQL scores, and glycemic control (HbA1c).
Methods
Youth and parents in the SEARCH for Diabetes in Youth Study (T1D: age 5–18, n = 3402; T2D: age 8–18, n = 353) completed the PedsQL Generic and Diabetes Modules, and youth provided a blood sample to assess HbA1c. Discrepancies (youth minus parent PedsQL ratings) were calculated and examined by age and diabetes type, and associations with youth PedsQL scores and HbA1c were evaluated.
Results
Discrepancies existed between youth and parent-proxy reports of generic and diabetes PedsQL scores in T1D and T2D (all p values < 0.01). Higher (more favorable) ratings were reported by youth except for those 5–7-years old, where parents’ scores were higher. When parent-proxy scores were higher, discrepancies were largest when the child reported low PedsQL scores. Higher HbA1c was associated with larger discrepancies (youth scores higher) for adolescents with T1D.
Conclusions
Discrepant PedsQL ratings suggest that parents may often underestimate youths’ HRQOL except in the youngest children. Although examining both reports is optimal, the youth report should be prioritized, particularly for young children with T1D and for adolescents with either T1D or T2D.
Keywords: Type 1 diabetes, Type 2 diabetes, Quality of life, Glycemic control
Introduction
Health-related quality of life (HRQOL) is an important construct assessing the impact of a medical condition on physical and mental well-being [1]. Research has demonstrated that poorer HRQOL is associated with higher HbA1c and greater depressive symptoms in the pediatric type 1 diabetes (T1D) and type 2 diabetes (T2D) populations [2, 3]. Clinical assessment and discussion of HRQOL is important to comprehensive management of diabetes spanning both physical and mental health, and interventions to improve HRQOL may promote optimal metabolic and psychosocial outcomes for this population [4].
Administration of the Pediatric Quality of Life Inventory (PedsQL) [5] as a marker of HRQOL in health research frequently assesses both the youth and parent perspectives (proxy) of the youth’s HRQOL [5, 6]. However, parent-proxy and youth self-reports often differ [7]. Little is known about what discrepant scores mean and what they may indicate in relation to child health outcomes. This “proxy problem” has been debated in the HRQOL literature, [7–9] but no conclusion has been reached on how to handle the apparent discrepancies.
Parents of children without chronic illness typically rate their child’s HRQOL better than the children themselves; in contrast, parents of children with a variety of chronic health conditions (including diabetes) typically rate their child’s HRQOL as worse than children themselves [2, 8–10]. However, across studies, the findings tend to be mixed, with discrepancies commonly reported in both directions [7, 11–14]. In a cross-sectional analysis of adolescents with chronic conditions and their parents, differences between adolescents and parent-proxy scores were found to be statistically significant, although absolute differences were small and observed in both directions (i.e., some with youth ratings higher, some with parent ratings higher) [7]. Further, analysis of a variety of demographic and disease-related factors such as age, gender, and education varied in terms of how they associated with the direction of discrepancies across the multiple studies, when reported [7,12–14]. In summary, the direction and magnitude of HRQOL discrepancies remain unclear, and there are no data on the potential importance of discrepant scores within families in relation to clinical outcomes.
Given that many studies, including large-scale clinical trials, collect both the patient and parent-proxy reports of HRQOL, understanding discrepancies and their associations with glycemic outcomes can inform and clarify the clinical utility and predictive value of these tools. We explored these questions with data from the SEARCH for Diabetes in Youth Study [15]. Our objectives were to explore the direction and magnitude of discrepancies between self- and parent-proxy reports of PedsQL scores in youth with T1D and T2D and to examine whether discrepancies were associated with youth-reported PedsQL scores or glycemic control. Based on the previous research cited above, we hypothesized that there would be discrepancies in both directions and that it would be more common for parental scores to be lower than youth scores. Given the minimal evidence in previous literature, our investigation of differences between T1D and T2D on discrepancies and demographic and clinical predictors of discrepancies was exploratory.
Methods and materials
Participants
SEARCH for Diabetes in Youth is a large multicenter study that conducts population-based ascertainment of youth with clinically diagnosed, non-gestational diabetes who are <20 at the time of diagnosis [16]. SEARCH has enrolled youth newly diagnosed with diabetes from 2002 through the present. Cases are ascertained from geographically defined populations in Ohio, Colorado, South Carolina and Washington, Indian Health Service beneficiaries from four American Indian populations, and enrollees in a managed health care plan in California. Youths whose diabetes is not secondary to other conditions are invited to a SEARCH study visit. After obtaining informed consent and assent, physical measurements were taken, and fasting blood samples are collected from metabolically stable children, and questionnaires are administered to the children and their accompanying parent. The study protocol was approved by the institutional review boards at all of the participating centers.
Youths whose diabetes was incident in 2002 through 2005 or prevalent in 2001 and who completed a baseline study visit were eligible for inclusion in these analyses. From this group, we selected children with T1D or T2D aged 5 through 18 at the time of their initial study visit and for whom we had completed child and the parent-proxy versions of the PedsQL questionnaires.
Measures
Demographic variables
Demographic variables included age at study visit, sex, race/ethnicity, highest level of parental education (either parent), and insurance status. Insurance was categorized as private, state-funded (Medicaid/Medicare, etc.), other (which included student health clinics, military, Indian Health Services), or none. Family composition was categorized as two-parent household, one-parent household, or other/unknown composition.
Clinical variables
Clinical variables included duration of diabetes at the time the survey was completed, insulin regimen, and glycemic control. Insulin regimen was categorized as (1) basal-bolus using the insulin pump, (2) basal-bolus with glargine plus rapid-acting insulin, (3) multiple daily injections (MDI) with ≥3 injections/day, using glargine plus more than/or other than rapid-acting insulin type, (4) MDI with ≥3 injections/day, using any insulin types excluding basal insulin, or (5) 1–2 injections/day, excluding glargine [17]. Detemir and glulisine insulin use was not reported by participants during the data collection period.
Glycemic control
Glycemic control was assessed using blood samples shipped to a central laboratory (Northwest Lipid Research Laboratories, Seattle, WA) for analysis. An ion exchange unit (Variant II; Bio-Rad Diagnostics, Hercules, CA) quantified the glycated hemoglobin (HbA1c) levels.
Health-related quality of life (HRQOL)
Health-related quality of life (HRQOL) was assessed using two modules of the Pediatric Quality of Life Inventory (PedsQL): the Generic Core (generic HRQOL) and Diabetes Module (diabetes HRQOL). All youths aged 5 years and older completed an age-appropriate self-report version of the two modules. For children unable to read, items were read out loud. A parent was also invited to complete a parent-proxy version of each PedsQL module for all youths who were aged <18 years at the time of their visit. Discrepancy scores were calculated as youth total scores minus the parent-proxy scores. Individual rating discrepancies were then categorized as “Youth score higher” for dyads where the youth scores were higher than parent score or “Parent-proxy score higher” for dyads where the parent-proxy scores were higher than the youth score.
Statistical analysis
Diabetes duration, age, sex, race/ethnicity, insulin regimen, health insurance, highest parental education, family income, and family composition were compared by diabetes type. Differences were assessed using ANOVA for continuous measures and the Chi-squared tests or Fisher’s exact tests for categorical responses. Multivariate linear regression models stratified by diabetes type were used to model absolute differences in generic and diabetes-specific PedsQL scores. Age, gender, race (non-Hispanic White vs. other), type of health insurance (private insured vs. other), parental education (less than high school vs. high school or more), family composition (one-parent household or other vs. two-parent household), and duration of diabetes were included in multivariate linear regression models.
Paired t tests were used to test for overall differences in parent and child scores. Discrepancies in parent-proxy and youth-reported PedsQL scores stratified by the direction of the discrepancy, age group, and diabetes type were examined. We explored the average differences in discrepancy magnitude across PedsQL tertiles representing the lowest one-third of youth-reported PedsQL scores, the middle one-third, and the highest one-third. p values were assessed using ANOVA to determine whether the discrepancy (for diabetes type and direction of discrepancy) differed by tertile.
T tests were used to compare HbA1c values across the two discrepancy direction categories (youth score higher vs. parent-proxy score higher), stratified by diabetes type and age group. Linear regression models were used to determine the relationship of discrepancy size and HbA1c. Models were additionally adjusted for discrepancy direction (youth score higher), gender, race, type of health insurance, family composition, number of household members, diabetes duration, and age group. We tested for interactions between discrepancy size and direction, discrepancy size and age group, and discrepancy direction and age group; no interactions were significant and thus were not retained in the presented models. For interpretability, a ten-point change in the discrepancy was used. All analyses were performed in SAS version 9.4 (Cary, NC), and a 0.05 significance level was used throughout this analysis.
Results
Participants
The study included responses from 3755 youth with diabetes of which 3402 had T1D and 353 had T2D. As expected, there were significant differences in age group, sex, race/ethnicity, diabetes duration, treatment regimen, health insurance, highest parental education, family income, and family composition (all p’s < 0.001) by diabetes type (Table 1). Therefore, we conducted the analysis stratified by diabetes type.
Table 1.
Demographic and clinical information characteristics of the 3755 participants in the study sample, by diabetes type: the SEARCH for diabetes in youth study
| Characteristic | T1D, N = 3402 |
T2D, N = 353* |
||
|---|---|---|---|---|
| N | % or M (SD) | N | % or M (SD) | |
| Age at study visit, years | ||||
| 5–7 | 581 | 17.1 | 0 | – |
| 8–11 | 1438 | 42.3 | 68 | 19.3 |
| 13–18 | 1383 | 40.7 | 285 | 80.7 |
| Age (mean, SD years) | 11.9 (3.4) | 14.9 (2.0) | ||
| Duration (months) | 38.3 (39.6) | 17.6 (16.5) | ||
| Gender (female) | 1702 | 50.0 | 222 | 62.9 |
| Insulin regimen | ||||
| Pump | 533 | 15.9 | 1 | 0.7 |
| Long + short/rapid insulin, three or more times a day | 833 | 24.8 | 16 | 11.6 |
| Long + any other combination, two or more times a day | 317 | 9.4 | 5 | 3.6 |
| Any combination of insulins excluding long, three or more times/day | 516 | 15.4 | 27 | 19.6 |
| Any insulin(s) taken 1x/day, or any insulin combination excluding long 2x/day | 1158 | 34.5 | 89 | 64.5 |
| Not on insulin | 44 | 1.3 | 211 | 60.5 |
| Race/ethnicity | ||||
| Black | 319 | 9.4 | 124 | 35.1 |
| Hispanic | 382 | 11.3 | 91 | 25.8 |
| Other | 103 | 3.0 | 55 | 15.6 |
| Non-hispanic White | 2598 | 76.4 | 83 | 23.5 |
| Type of health insurance | ||||
| None | 32 | 0.9 | 12 | 3.4 |
| Other | 59 | 1.7 | 15 | 4.3 |
| Medicaid | 577 | 17.0 | 134 | 38.4 |
| Private | 2719 | 80.3 | 188 | 53.9 |
| Highest level of education of either parent | ||||
| Bachelors degree or more | 1575 | 46.5 | 53 | 15.1 |
| Some college through associate degree | 1163 | 34.3 | 126 | 35.8 |
| High school graduate | 504 | 14.9 | 112 | 31.8 |
| Less than high school graduate | 145 | 4.3 | 61 | 17.3 |
| Family composition | ||||
| One-parent household | 939 | 27.7 | 169 | 48.0 |
| Two-parents/one household | 2350 | 69.3 | 160 | 45.5 |
| Other | 100 | 3.0 | 23 | 6.5 |
| Income | ||||
| $25–49 K | 747 | 22.1 | 93 | 26.4 |
| $50–74 K | 697 | 20.6 | 38 | 10.8 |
| $75 K+ | 1336 | 39.5 | 33 | 9.4 |
| < $25 K | 432 | 12.8 | 151 | 42.8 |
| Don’t know or refused | 173 | 5.1 | 38 | 10.76 |
All p’s < 0.001 for comparison by diabetes type
Size and direction of PedsQL rating discrepancies
Mean youth and parent-proxy scores on the generic and diabetes-specific PedsQL are displayed in Table 2. Youth with T1D and T2D reported higher scores than parent-proxy scores for both generic and diabetes PedsQL modules (all p’s < 0.01). The size and direction of the discrepancies, stratified by age, are shown in Table 2. Among youth with T1D between the ages of 8–12 and 13–18, it was more common (approximately 2:1) for them to rate their generic and their diabetes PedsQL higher than their parents’ proxy scores. Among the 5–7-year-olds with T1D, higher parental scores were more common.
Table 2.
Overall youth and parent scores, and direction and magnitude of discrepancy scores (absolute value), stratified by age, for generic and diabetes-specific PedsQL Modules
| Age group |
PedsQL score: M ± SD |
Youth score higher |
Parent-proxy score higher |
||||
|---|---|---|---|---|---|---|---|
| Youth | Parent-proxy | Percent of age group |
Discrepancy size, M ± SD |
Percent of age group |
Discrepancy size, M ± SD |
||
| Type 1 diabetesa,b | |||||||
| Generic | All | 80.3 ± 13.2 | 78.0 ± 14.08 | ||||
| 5–7 | 71.7 ± 14.6 | 78.4 ± 12.7 | 45 | 10.0 ± 8.6 | 55 | 15.5 ± 11.9 | |
| 8–12 | 81.7 ± 12.7 | 78.6 ± 13.6 | 62 | 11.4 ± 9.7 | 38 | 10.0 ± 8.4 | |
| 13–18 | 82.5 ± 11.6 | 77.1 ± 15.0 | 67 | 12.4 ± 11.0 | 33 | 9.0 ± 6.9 | |
| Diabetes | All | 74.7 ± 13.5 | 70.4 ± 13.1 | ||||
| 5–7 | 68.3 ± 14.3 | 71.1 ± 11.9 | 47 | 10.6 ± 9.5 | 53 | 14.5 ± 11.6 | |
| 8–12 | 77.1 ± 13.1 | 71.5 ± 12.5 | 68 | 13.2 ± 9.8 | 32 | 10.3 ± 8.9 | |
| 13–18 | 74.9 ± 12.7 | 69.0 ± 14.1 | 68 | 13.2 ± 10.0 | 32 | 9.8 ± 7.8 | |
| Type 2 diabetesa | |||||||
| Generic | All | 75.0 ± 15.4 | 67.4 ± 17.4 | ||||
| 8–12 | 73.1 ± 19.3 | 66.2 ± 17.7 | 71 | 15.5 ± 12.0 | 29 | 13.4 ± 15.3 | |
| 13–18 | 75.5 ± 14.4 | 67.7 ± 17.3 | 65 | 17.8 ± 12.9 | 35 | 10.7 ± 8.7 | |
| Diabetes | All | 71.5 ± 15.6 | 68.0 ± 15.9 | ||||
| 8–12 | 69.6 ± 19.1 | 68.9 ± 15.3 | 62 | 12.3 ± 9.9 | 38 | 18.0 ± 15.0 | |
| 13–18 | 72.0 ± 14.7 | 67.8 ± 16.0 | 61 | 14.3 ± 11.3 | 39 | 11.4 ± 8.4 | |
All p values < 0.01 for comparisons of youth versus parent-proxy scores
All p values < 0.001 for comparisons between age groups per type of discrepancy
For all ages of youth with T2D, it was more common for the youths’ scores to be higher than the parents’ scores for both the generic and diabetes PedsQL. All parental and child scores significantly differed (p < 0.01 for all comparisons).
PedsQL discrepancies and demographic variables
Age, sex, race, and insurance were associated with generic PedsQL score discrepancy size in youth with T1D. Among youth with T1D, for generic PedsQL, the youngest group (age 5–7) had larger discrepancies compared to adolescents (age 13–18; β = 2.27, p < 0.001). Males had larger discrepancies than females (β = −0.83, p = 0.01), and those of non-White race had larger discrepancies than White participants (β = 1.21, p < 0.01). Those with private insurance had smaller discrepancies than those with public insurance (β = −2.27, p < 0.01), and those with two-parent households had smaller discrepancies than those with one-parent households (β = 1.34, p < 0.01). For diabetes PedsQL scores in youth with T1D, only insurance was associated with discrepancy size, such that those with private insurance had smaller discrepancies (β = −1.24, p = 0.01). For youth with T2D, males had larger discrepancies than females in generic HRQOL scores (β = 3.79, p = 0.01). No other demographic variable was associated with discrepancy size.
PedsQL discrepancies and glycemic control
Mean HbA1c values for each type of discrepancy (youth score higher vs. parent-proxy score higher), stratified by age and diabetes type, are presented in Table 3. For the full sample of youth with T1D, HbA1c values were significantly higher among youth whose PedsQL scores (both generic and diabetes) were higher than the parent-proxy scores compared to those whose parent-proxy ratings were higher. Stratified by age group, this pattern held only for youth aged 13–18 (p < 0.01). There were no HbA1c differences by discrepancy direction for youth aged 5–7 or 8–12. There were no HbA1c differences by discrepancy direction for youth with T2D at any age.
Table 3.
T-tests comparing HbA1c values by discrepancy direction, stratified by diabetes type and age group
| Generic PedsQL |
Diabetes PedsQL |
|||||||
|---|---|---|---|---|---|---|---|---|
| HbA1c, M ± SD |
p value | Effect size | HbA1c, M ± SD |
p value | Effect size | |||
| Youth score higher |
Parent-proxy score higher |
Youth score higher |
Parent-proxy score higher |
|||||
| Type 1 diabetes | ||||||||
| All Ages | 8.2 ± 1.7 | 8.0 ± 1.4 | <0.0001 | 0.14 | 8.2 ± 1.6 | 8.0 ± 1.5 | 0.00 | 0.13 |
| 5–7 years | 7.9 ± 1.3 | 7.9 ± 1.1 | 0.82 | 0.02 | 7.9 ± 1.2 | 7.9 ± 1.1 | 0.81 | 0.02 |
| 8–12 years | 8.0 ± 1.4 | 8.0 ± 1.3 | 0.44 | 0.04 | 8.0 ± 1.4 | 7.9 ± 1.4 | 0.07 | 0.11 |
| 13–18 years | 8.4 ± 1.9 | 8.1 ± 1.7 | 0.00 | 0.18 | 8.4 ± 1.9 | 8.2 ± 1.8 | 0.04 | 0.12 |
| Type 2 diabetes | ||||||||
| All Ages | 7.8 ± 2.3 | 7.3 ± 2.2 | 0.06 | 0.22 | 7.7 ± 2.3 | 7.5 ± 2.1 | 0.29 | 0.12 |
| 8–12 years | 6.8 ± 1.6 | 7.5 ± 2.2 | 0.17 | 0.37 | 7.1 ± 2.1 | 7.5 ± 2.0 | 0.41 | 0.21 |
| 13–18 years | 7.4 ± 2.3 | 7.8 ± 2.3 | 0.12 | 0.20 | 7.5 ± 2.1 | 7.8 ± 2.4 | 0.42 | 0.10 |
Linear regression models were also used to quantify the relationship between discrepancy sizes and HbA1c values (Table 4). Larger discrepancies and the “Youth score higher” group were associated with higher HbA1c in T1D after adjusting for sex, insurance, parental education, household composition, number of household members, diabetes duration, and age group. A ten-unit increase in generic and diabetes-specific PedsQL discrepancy size was associated with a 0.07 % and 0.08 % increase, respectively, in HbA1c. Youths who scored higher than their parents were associated with an 0.11 (for generic PedsQL) and 0.12 (for diabetes PedsQL) percent increase in HbA1c.
Table 4.
Linear regression models of generic and diabetes HRQOL predicting HbA1C, stratified by diabetes type
| Parameter | HbA1c |
|||||||
|---|---|---|---|---|---|---|---|---|
| Version | Beta | SE | p value | Version | Beta | SE | p value | |
| Type 1 diabetes | ||||||||
| PedsQL: Absolute difference (youth–parent), 10-point difference |
Generic | 0.07 | 0.03 | 0.01 | Diabetes-Specific | 0.08 | 0.03 | 0.00 |
| Youth PedsQL score higher | 0.11 | 0.05 | 0.04 | 0.12 | 0.06 | 0.03 | ||
| Female sex | 0.14 | 0.05 | 0.01 | 0.14 | 0.05 | 0.01 | ||
| Non-White race | 0.43 | 0.07 | <.0001 | 0.43 | 0.07 | <.0001 | ||
| Private insurance | −0.28 | 0.07 | 0.00 | −0.28 | 0.07 | <.0001 | ||
| Parental education > HS | −0.22 | 0.07 | 0.00 | −0.23 | 0.07 | 0.00 | ||
| Not living in a two-parent household | 0.34 | 0.06 | <.0001 | 0.35 | 0.06 | <.0001 | ||
| Number of household members | 0.04 | 0.02 | 0.08 | 0.04 | 0.02 | 0.08 | ||
| Diabetes duration (months) | 0.01 | 0.00 | <.0001 | 0.01 | 0.00 | <.0001 | ||
| Age group | ||||||||
| Ages 5–7 | −0.1 | 0.08 | 0.21 | −0.1 | 0.08 | 0.2 | ||
| Ages 8–12 | −0.1 | 0.06 | 0.09 | −0.11 | 0.06 | 0.06 | ||
| Ages 13–18 | Ref | Ref | ||||||
| Type 2 diabetes | ||||||||
| PedsQL: Absolute Difference (youth–parent), 10-point difference |
Generic | 0.1 | 0.1 | 0.32 | Diabetes-Specific | 0.04 | 0.11 | 0.72 |
| Youth PedsQL score higher | 0.52 | 0.26 | 0.04 | 0.28 | 0.24 | 0.25 | ||
| Female sex | 0.22 | 0.25 | 0.38 | 0.18 | 0.25 | 0.48 | ||
| Non-White race | 0.58 | 0.29 | 0.04 | 0.64 | 0.29 | 0.03 | ||
| Private insurance | 0.1 | 0.25 | 0.69 | 0.13 | 0.25 | 0.6 | ||
| Parental education > HS | −0.03 | 0.24 | 0.9 | −0.09 | 0.24 | 0.71 | ||
| Not living in a two-parent household | −0.75 | 0.25 | 0.00 | −0.77 | 0.26 | 0.00 | ||
| Number of household members | 0.06 | 0.07 | 0.44 | 0.05 | 0.08 | 0.54 | ||
| Diabetes duration (months) | 0.04 | 0.01 | <.0001 | 0.04 | 0.01 | <.0001 | ||
| Age group | ||||||||
| Ages 8–12 | −0.08 | 0.3 | 0.79 | −0.07 | 0.3 | 0.81 | ||
| Ages 13–18 | Ref | Ref | ||||||
Discussion
This study investigated the magnitude and direction of discrepancies between youth and parent-proxy reports of both generic and disease-specific PedsQL scores in a large, national sample of youth with T1D and T2D. We found that the majority of youth with T1D and T2D reported higher generic and diabetes-specific PedsQL scores than their parent reported. The only exception was in the T1D 5–7-year-old age group, where the majority of parents rated their child’s generic and diabetes-specific HRQOL higher than the youth.
Previous studies of the “proxy problem” have not been conducted in such large samples nor have they been stratified by age or diabetes type. Our report suggests that parents may underestimate youths’ PedsQL scores, except in the youngest children with T1D. Also of note in the 5–7-year-old T1D group was the finding that for parents rating youth PedsQL higher than the child (55 % of all parents for generic and 53 % for diabetes-specific), the magnitude of discrepancy was also the largest. Clearly there is a disconnection between parents’ and youths’ perceptions of youth HRQOL in this young age group. Given that parents are doing much of the diabetes care at this stage, many may underestimate the effect diabetes has on their young child. Parents of young children with diabetes may have difficulty distinguishing between normative and diabetes-related behavior, so it may be difficult to accurately assess their child’s overall well-being [18].
When we examined discrepancies of PedsQL reports in relation to youths’ HbA1c levels at the time of the study visit, we found that among adolescents 13–18-years old with T1D, HbA1c values were higher when the youth reported higher PedsQL scores than their parent reported, for both the generic and diabetes-specific PedsQL modules. Further, larger discrepancies were associated with worse glycemic control. Given adolescents are more likely to have worsening treatment adherence and poor glycemic control [19], it is possible that parental perceptions of adolescent HRQOL may take adherence or glycemic control into consideration, thus rating HRQOL lower when HbA1c is poor; however, teens’ own perceptions of HRQOL may not be related to HbA1c. These results may also reflect differences between parent and youth perspectives on HRQOL that encompass broader family issues that were not measured in this study, such as family conflict or poor communication in the home, as both have been associated with HRQOL and HbA1c [15–17]. This finding was only observed among youth with T1D. There were no associations between discrepancies and HbA1c among the youth with T2D.
We found consistent differences between T1D and T2D groups in relation to discrepancy size and demographic and clinical correlates of discrepancies. The smaller HRQOL discrepancies in T1D may reflect a relatively higher level of involvement in diabetes management among caregivers of youth with T1D; however, this is speculative as we are not aware of research comparing parent involvement in T1D versus T2D. Additionally, larger discrepancies were associated with demographic variables such as race and insurance status for T1D. More research is needed on the relationships between parental involvement, reported discrepancies, and demographic variables, particularly ones that reflect socio-economic status that we found significant in the T1D sample as they may be another indicator of the negative effects of health disparities. The lack of association between demographic variables and HRQOL discrepancies in T2D may be related to less variability (i.e., lower SES overall, more racial/ethnic minorities) in the T2D sample.
Limitations of this study included the cross-sectional design, which prohibits our ability to establish causality, particularly in the analyses with HbA1c in which the directionality of the associations we found remains unclear. Further, our study was not able to provide clarity on the reasons for the differences between the discrepant reports. Future research would benefit in understanding more about the nature of these reasons and how to reduce discrepancies. Lastly, we were unable to answer the question of whose HRQOL is more important to collect, as that becomes more of a philosophical versus empirical question undoubtedly influenced by many factors beyond what we have collected here.
In summary, our study investigated the difference between child and parent-proxy scores on the generic and diabetes-specific PedsQL modules. Despite the limitations listed above, we found that indeed discrepancies not only exist, but they relate to glycemic outcomes among adolescents with T1D. Thus, in the context of clinical care it is important to consider discrepancies in HRQOL reports as much as possible, as large discrepancies, particularly when the parents report higher scores than their child, can indicate a worrisome lack of correspondence between the parent’s and the adolescent’s perspectives. If providers note large discrepancies, it may be beneficial to dedicate more time to discussing HRQOL issues with families. However, more research is needed to determine whether HRQOL interventions [19] for families with larger versus smaller discrepancies are impactful.
Taken together, our findings support prior recommendations [17, 18] to use the youth report (when available) in conjunction with the parent-proxy report when making determinations on youth HRQOL. However, when youth and parent-proxy scores are highly discrepant, they should be considered individually and not be averaged together or otherwise aggregated. This may be particularly important in adolescence, given the large discrepancies found in T2D and the association of larger discrepancies with HbA1c in adolescents with T1D. Although examining both reports is optimal, if it is unrealistic to collect both, the youth report should be prioritized.
Acknowledgments
The SEARCH for Diabetes in Youth Study is indebted to the many youth and their families, and their health care providers, whose participation made this study possible. The authors wish to acknowledge the involvement of General Clinical Research Centers (GCRC) at the South Carolina Clinical & Translational Research (SCTR) Institute, at the Medical University of South Carolina (NIH/NCRR Grant number UL1RR029882); Seattle Children’s Hospital (NIH CTSA Grant UL1 TR00423 of the University of Washington); University of Colorado Pediatric Clinical and Translational Research Center (CTRC) (Grant Number UL1 TR000154) and the Barbara Davis Center at the University of Colorado at Denver (DERC NIH P30 DK57516); and the National Center for Research Resources and the National Center for Advancing Translational Sciences, National Institutes of Health, through Grant 8 UL1 TR000077; and the Children with Medical Handicaps program managed by the Ohio Department of Health. The work of M.E.H. on this project was supported by K12 DK097696. The findings and conclusions in this report are those of the authors and do not necessarily represent the official position of the Centers for Disease Control and Prevention and the National Institute of Diabetes and Digestive and Kidney Diseases.
Funding SEARCH for Diabetes in Youth is funded by the Centers for Disease Control and Prevention (PA numbers 00097, DP-05-069, and DP-10-001) and supported by the National Institute of Diabetes and Digestive and Kidney Diseases.
Site contract numbers Kaiser Permanente Southern California (U48/CCU919219, U01 DP000246, and U18DP002714), University of Colorado Denver (U48/CCU819241-3, U01 DP000247, and U18DP000247-06A1), Kuakini Medical Center (U58CCU919256 and U01 DP000245), Children’s Hospital Medical Center (Cincinnati) (U48/CCU519239, U01 DP000248, and 1U18DP002709), University of North Carolina at Chapel Hill (U48/CCU419249, U01 DP000254, and U18DP002708), University of Washington School of Medicine (U58/CCU019235-4, U01 DP000244, and U18DP002710-01), Wake Forest University School of Medicine (U48/CCU919219, U01 DP000250, and 200-2010-35171).
Footnotes
A version of these results was presented at the International Society of Pediatric and Adolescent Diabetes meeting, on September 2014
Compliance with ethical standards
Conflict of interest No authors have any reported conflict of interest.
Ethical approval All procedures performed in studies involving human participants were in accordance with the ethical standards of the institutional and/or national research committee and with the 1964 Helsinki declaration and its later amendments or comparable ethical standards.
Informed consent Informed consent was obtained from all individual participants included in the study.
References
- 1.Cameron FJ. The impact of diabetes on health-related quality of life in children and adolescents. Pediatric Diabetes. 2003;4:132–136. doi: 10.1034/j.1399-5448.2003.00019.x. [DOI] [PubMed] [Google Scholar]
- 2.Allan CL, Flett B, Dean HJ. Quality of life in First Nation youth with Type 2 diabetes. Maternal and Child Health Journal. 2008;12:103–109. doi: 10.1007/s10995-008-0365-x. [DOI] [PubMed] [Google Scholar]
- 3.Lawrence JM, Yi-Frazier JP, Black MH, et al. Demographic and clinical correlates of diabetes-related quality of life among youth with type 1 diabetes. Journal of Pediatrics. 2012;161:201–207. doi: 10.1016/j.jpeds.2012.01.016. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.de Wit M, Winterdijk P, Aanstoot HJ, et al. Assessing diabetes-related quality of life of youth with type 1 diabetes in routine clinical care: The MIND Youth Questionnaire (MY-Q) Pediatric Diabetes. 2012;13:638–646. doi: 10.1111/j.1399-5448.2012.00872.x. [DOI] [PubMed] [Google Scholar]
- 5.Varni JW, Burwinkle TM, Jacobs JR, et al. The PedsQL in type 1 and type 2 diabetes: Reliability and validity of the pediatric quality of life inventory generic core scales and type 1 diabetes module. Diabetes Care. 2003;26:631–637. doi: 10.2337/diacare.26.3.631. [DOI] [PubMed] [Google Scholar]
- 6.Varni JW, Limbers CA, Burwinkle TM. Parent proxy-report of their children’s health-related quality of life: an analysis of 13,878 parents’ reliability and validity across age subgroups using the PedsQL 4.0 Generic Core Scales. Health & Quality of Life Outcomes. 2007;5:2. doi: 10.1186/1477-7525-5-2. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Sattoe JN, van Staa A, Moll HA. On your own feet research group: The proxy problem anatomized: child-parent disagreement in health related quality of life reports of chronically ill adolescents. Health & Quality of Life Outcomes. 2012;10:10. doi: 10.1186/1477-7525-10-10. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Eiser C, Varni JW. Health-related quality of life and symptom reporting: Similarities and differences between children and their parents. European Journal of Pediatrics. 2013;172(10):1299–1304. doi: 10.1007/s00431-013-2049-9. [DOI] [PubMed] [Google Scholar]
- 9.Theunissen NC, Vogels TG, Koopman HM, et al. The proxy problem: child report versus parent report in health-related quality of life research. Quality of Life Research. 1998;7(5):387–397. doi: 10.1023/a:1008801802877. [DOI] [PubMed] [Google Scholar]
- 10.Nansel TR, Weissberg-Benchell J, Wysocki T, et al. Quality of life in children with Type 1 diabetes: A comparison of general and diabetes-specific measures and support for a unitary diabetes quality-of-life construct. Diabetic Medicine. 2008;25:1316–1323. doi: 10.1111/j.1464-5491.2008.02574.x. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Hilliard ME, Mann KA, Peugh JL, Hood KK. How poorer quality of life in adolescence predicts subsequent type 1 diabetes management and control. Patient Education and Counseling. 2013;91:120–125. doi: 10.1016/j.pec.2012.10.014. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Silva N, Crespo C, Carona C, Bullinger M, Canavarro MC. Why the (dis)agreement? Family context and child-parent perspectives on health-related quality of life and psychological problems in paediatric asthma. Child: Care, Health and Development. 2015;41:112–121. doi: 10.1111/cch.12147. [DOI] [PubMed] [Google Scholar]
- 13.Pan R, Egberts MR, Nascimento LC, et al. Health-related quality of life in adolescent survivors of burns: Agreement on self-reported and mothers’ and fathers’ perspectives. Burns. 2015;41:1107–1113. doi: 10.1016/j.burns.2014.12.011. [DOI] [PubMed] [Google Scholar]
- 14.Murray CB, Holmbeck GN, Ros AM, et al. A longitudinal examination of health-related quality of life in children and adolescents with spina bifida. Journal of Pediatric Psychology. 2015;40:419–430. doi: 10.1093/jpepsy/jsu098. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Hamman RF, Bell RA, Dabelea D, et al. The SEARCH for diabetes in youth study: Rationale, findings, and future directions. Diabetes Care. 2014;37:3336–3344. doi: 10.2337/dc14-0574. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.SEARCH Study Group SEARCH for Diabetes in Youth: a multicenter study of the prevalence, incidence and classification of diabetes mellitus in youth. Controlled Clinical Trials. 2004;25:458–471. doi: 10.1016/j.cct.2004.08.002. [DOI] [PubMed] [Google Scholar]
- 17.Paris CA, Imperatore G, Klingensmith G, et al. Predictors of insulin regimens and impact on outcomes in youth with type 1 diabetes: The SEARCH for diabetes in youth study. Journal of Pediatrics. 2009;155:183–189. doi: 10.1016/j.jpeds.2009.01.063. [DOI] [PubMed] [Google Scholar]
- 18.Hilliard ME, Monaghan M, Cogen FR, Streisand R. Parent stress and child behaviour among young children with type 1 diabetes. Child: Care, Health and Development. 2011;37:224–232. doi: 10.1111/j.1365-2214.2010.01162.x. [DOI] [PubMed] [Google Scholar]
- 19.de Wit M, Delemarre-van de Waal HA, Bokma JA, et al. Monitoring and discussing health-related quality of life in adolescents with type 1 diabetes improve psychosocial well-being: a randomized controlled trial. Diabetes Care. 2008;31:1521–1526. doi: 10.2337/dc08-0394. [DOI] [PMC free article] [PubMed] [Google Scholar]