A diagnosis of type 1 diabetes leads to a significant change in lifestyle for the affected child or teenager and their family. Considerable and unrelenting personal and family stress arise from the regular administration of subcutaneous insulin, frequent blood glucose testing, monitoring, dietary alterations and management of hypo- and hyperglycaemia. It is not surprising, therefore, that type 1 diabetes would be associated with a reduction in an individuals health-related quality of life (HRQOL).Quality of life has been defined as a broad range of human experience related to ones overall well-being\u2026 defined by subjective experiences, states and perceptions. 1 It is increasingly recognised as an important part of any assessment of health and well-being,2 and in the overall understanding of chronic illness. A number of different measurement tools have been validated and proven useful to detect hidden morbidities, improve patient-doctor communication, and improve clinical decision-making.3 Some advocate the assessment of HRQOL as part of best clinical practice, and quantifying HRQOL may predict future health-care costs.4 HRQOL tools can allow for comparisons between healthy populations and those struggling with chronic conditions, such as asthma or diabetes.5,6The impact of living with type 1 diabetes has been well described, and it is generally reported that children and adolescents with diabetes show sub-optimal quality of life compared with their peers.7-11 The concern is not only around negative effects on day-to-day functioning, but whether impairment in quality of life reflects long-term impacts on mental health. A number of studies in children with diabetes have found poorer HRQOL, and later higher psychiatric morbidity compared with a group of age- and gender-matched controls.12-14 However, a meta-analysis of 22 studies found only a mildly elevated overall risk of psychological difficulties, including depression, anxiety and behavioural problems in children with diabetes.15 HRQOL has not previously been studied in New Zealand children with diabetes.The primary objective of this study was to evaluate the quality of life of children and adolescents with type 1 diabetes in the Taranaki region (New Zealand) compared to a group of their siblings without diabetes or chronic disease.MethodsParticipantsAll children and adolescents with type 1 diabetes in the Taranaki region and their parents were approached during the multi-disciplinary diabetes clinics in May-July, 2013. At this time, there were 67 children and adolescents with diabetes in the region (background population 23,127 children/young people less than 15 years of age).16 All patients were managed by a team of paediatricians, a dietitian, and a diabetes nurse educator, who were able to refer patients to Child and Adolescent Mental Health Services (CAMHS), if deemed necessary. Study approval was obtained by the Taranaki Base Hospital Clinical Board, and the Mori Health Unit. The study was classed as an audit for study purposes by the New Zealand Health and Disability Ethics Committee.Exclusion criteria for diabetes group analysis included: those diagnosed within the last 6 months; those with type 2 diabetes or cystic fibrosis (CF)-related diabetes; and those with psychological or other medical conditions (active CAMHS involvement, complex congenital heart disease, and attention deficit hyperactivity disorder [ADHD]).The Pediatric Quality of Life Inventory (PedsQLTM) measurement model was specifically designed to evaluate HRQOL in children and adolescents, with both parent-proxy and child self-report versions. The 23-item Generic Core Scale questionnaire is divided into four areas, assessing problems over the preceding month related to physical, emotional, social and school functioning.17-19 The reliability and content validity of this instrument has been demonstrated.20,21 The developers of the PedsQLTM propose using one standard-deviation below the population mean as a meaningful cut-off for those at risk of impaired HRQOL compared to other children.20Parent-proxy questionnaires in all age groups (2-17 years) and patient self-report questionnaires in those aged 8-17 years were completed in the waiting room prior to, or after, clinical appointments. Parent-proxy questionnaires and self-reports for those aged 8-16 years were also collected for all siblings. The same parent answered the proxy reports for all participating offspring, so if there was more than one sibling, all were included. Questionnaires were also collected from diabetes patients who were without siblings, or whose siblings were outside the age range (2-17 years).Demographic information was collected and New Zealand Deprivation Index score was calculated based on home address, using 2006 census data on household income. The score is divided into deciles; one representing the least deprived area and 10 the most deprived.22 Diabetes-related information was collected, including clinical details, duration of disease, average glycated Haemoglobin (HbA1c) over past year, associated diagnoses, and hospital admissions (admissions at diagnosis were excluded in analyses). HbA1c was collected in clinic, using point-of-care testing with the DCA 2000 Analyzer (Siemens Medical Solutions Diagnostics, Puteaux, France).For those patients with diabetes included in analysis, average HbA1c were divided into three levels of control used in clinic, based on the International Society for Pediatric and Adolescent Diabetes (ISPAD) guidelines,23 with optimal control defined as <58mmol/mol (HbA1c <7.5% DCCT standardised), suboptimal control 58-75mmol/mol (7.5-9%) and high risk of metabolic complications>75mmol/mol (>9%).Quality of life measuresThe PedsQLTM 4.0 Generic Core Scale was used for both the sibling and diabetes groups. Item answers were reverse scored and linearly transformed to a 1-100 scale, with higher scores representing a better health-related quality of life (Generic Scaled Score). The Generic Core Total Score results were divided into two outcomes: the Psychosocial Health Summary Score and Physical Health Summary Score. The summary scores were then converted to Psychosocial Scaled Score and Physical Health Scaled Scores.Statistical analysesStatistical analyses were performed using SAS version 9.3 (SAS Institute Inc. Cary NC) and R version 2.15 (R Foundations for Statistical Computing). All statistical tests were two-sided at a 5% significance level. Correlation and agreement between the parent proxy- and child self-reports for those aged 8 years and older, were evaluated using the Spearmans correlation coefficients and the Bland-Altman plots. Any missing parent proxy-reports were replaced with the child self-report, if the information was available (ie, child was>8 years of age).Simple t-tests were first conducted to compare the Scaled Scores between the diabetes group and their siblings. Linear regression models were next fitted to explore the mean difference in scaled scores between the two groups adjusting for important confounding factors (age group, gender, ethnicity, and deprivation index). For the type 1 diabetes group, linear regression models were used to investigate the predictive effects of HbA1c (the mean and levels of control), duration of diabetes and any hospital admission for diabetes on the scaled scores, adjusting for the same confounding factors. Regression coefficients and 95% confidence intervals were estimated.ResultsThis study achieved an 84% response rate in our total diabetes population (n=56/67). Fourteen patients were subsequently excluded (those with type 2 diabetes [n=3], Cystic fibrosis [n=1], ADHD [n=2], CAMHS involvement [n=4], diagnosis within 6 months [n=3] and other medical conditions [n=1]).Results from 42 type 1 diabetes patients were analysed and compared with siblings aged 2-16 years without diabetes or chronic illness who participated in this study.Parent-proxy reports for the PedsQLTM Generic Core Scale were obtained for 93% of those who participated in the diabetes group (n=39/42; 3 adolescents attended clinic without a parent). Self-reports from 100% of older children and adolescents in the diabetes group (aged 8-17 years) were completed (n=35/35).Questionnaires were completed on the sibling comparisons, with parent-proxy reports returned on the whole group and self-reports obtained on 96% of the siblings aged 8-16 years (n=25/26).The demographic data of the participants are given in Table 1. There were no differences in demographic parameters between groups, apart from deprivation index, which was higher in the diabetes group. This arose from some diabetes patients having multiple siblings included, and some with no, unwilling or ineligible siblings (n=38). This was adjusted for in regression analysis.Table 1: Demographic characteristics of patients with type 1 diabetes and their siblings. Data are means \u00b1 SD (ranges) or n (%). \r\n \r\n \r\n \r\n \r\n \r\n Type 1 diabetes\r\n \r\n Siblings\r\n \r\n p-value\r\n \r\n \r\n \r\n N\r\n \r\n 42\r\n \r\n 35\r\n \r\n \r\n \r\n \r\n \r\n Gender (males)\r\n \r\n 25 (60%)\r\n \r\n 16 (46%)\r\n \r\n 0.23\r\n \r\n \r\n \r\n Ethnicity\r\n \r\n \r\n \r\n \r\n \r\n 0.67\r\n \r\n \r\n \r\n New Zealand European\r\n \r\n 30 (71%)\r\n \r\n 28 (80%)\r\n \r\n \r\n \r\n \r\n \r\n Mori\r\n \r\n 8 (19%)\r\n \r\n 5 (14%)\r\n \r\n \r\n \r\n \r\n \r\n Other\r\n \r\n 4 (10%)\r\n \r\n 2 (6%)\r\n \r\n \r\n \r\n \r\n \r\n Age (years)\r\n \r\n 11.5 \u00b1 3.8 (2-17)\r\n \r\n 10.2 \u00b1 3.7 (4-16)\r\n \r\n 0.14\r\n \r\n \r\n \r\n Age group\r\n \r\n \r\n \r\n \r\n \r\n 0.22\r\n \r\n \r\n \r\n 2-7 years\r\n \r\n 7 (16%)\r\n \r\n 9 (26%)\r\n \r\n \r\n \r\n \r\n \r\n 8-12 years\r\n \r\n 15 (36%)\r\n \r\n 16 (46%)\r\n \r\n \r\n \r\n \r\n \r\n 13-17 years\r\n \r\n 20 (48%)\r\n \r\n 10 (29%)\r\n \r\n \r\n \r\n \r\n \r\n Deprivation index\r\n \r\n 5.6 \u00b1 2.4 (1-10)\r\n \r\n 4.5 \u00b1 2.1 (1-10)\r\n \r\n 0.04\r\n \r\n \r\n \r\nThe duration of diabetes ranged from 9 months to 13 years, with a mean of 5.1 years (\u00b1 3.6SD). One third (n=14/42) had at least one diabetes-related hospital admission within the last year. The mean HbA1c in the past year was 72.5 \u00b1 18.5 mmol/mol (8.8% \u00b1 3.9 DCCT standard). Only 12% of patients (n=5/42) had an average HbA1c in the optimal control range, compared with 60% (n=25/42) in the suboptimal control, and 29% (12/42) in the high risk of metabolic complications groups (Table 2).Table 2: Clinical characteristics of patients with type 1 diabetes. Data are means \u00b1 SD (ranges) or n (%).\r\n \r\n \r\n \r\n N\r\n \r\n 42\r\n \r\n \r\n \r\n Diabetes duration (years)\r\n \r\n 5.1 \u00b1 3.6 (0.75-13.58)\r\n \r\n \r\n \r\n HbA1c (mmol/mol)\r\n \r\n 72.5 \u00b1 18.5 (36.8-130)\r\n \r\n \r\n \r\n Glycaemic control (HbA1c)\r\n \r\n \r\n \r\n \r\n \r\n High risk (>75 mmol/mol or>9%)\r\n \r\n 12 (29%)\r\n \r\n \r\n \r\n Suboptimal (58-75mmol/mol or 7.5-9%)\r\n \r\n 25 (60%)\r\n \r\n \r\n \r\n Optimal (<58mmol/mol or <7.5%)\r\n \r\n 5 (12%)\r\n \r\n \r\n \r\n Hospital admissions for diabetes in the past year\r\n \r\n \r\n \r\n \r\n \r\n 0\r\n \r\n 28 (67%)\r\n \r\n \r\n \r\n 1\r\n \r\n 11 (26%)\r\n \r\n \r\n \r\n \u22652\r\n \r\n 3 (7%)\r\n \r\n \r\n \r\nExcluding those diagnosed <6 months, type 2 diabetes, CF-related and associated diagnoses (congenital heart disease, ADHD, CAMHS involvement). Comparing parent-proxy and child self-report responses for those children aged 8 years and older showed significant correlations across all PedsQLTM Generic Scaled Scores using the total cohort. The Spearman correlation coefficient for the Physical Scaled Score was 0.55 (p<0.0001), and for the Psychosocial Scaled Score 0.48 (p<0.0001) respectively. The Bland-Altman plots indicated good level of agreement on all scores, with only small bias of 1.5 and 0.3 between the child and parent reports respectively. These findings have enabled the parent responses to be used for the entire analysis.Table 3 shows the unadjusted total Generic Scaled Scores for the sibling and diabetes patient groups, as assessed by the parent proxy reports.Table 3: Unadjusted quality of life Generic Scaled Scores (out of 100), as assessed by the parent proxy-reports. Data are means \u00b1 SD. \r\n \r\n \r\n \r\n \r\n \r\n Type 1 diabetes\r\n \r\n Siblings\r\n \r\n p-value\r\n \r\n \r\n \r\n n\r\n \r\n 42\r\n \r\n 35\r\n \r\n \r\n \r\n \r\n \r\n Total Generic Scaled Score\r\n \r\n 75.9 \u00b1 13.4\r\n \r\n 80.8 \u00b1 14.0\r\n \r\n 0.14\r\n \r\n \r\n \r\n Physical Scaled Score\r\n \r\n 80.0 \u00b1 17.9\r\n \r\n 82.7 \u00b1 14.2\r\n \r\n 0.52\r\n \r\n \r\n \r\n Psychosocial Scaled Score\r\n \r\n 73.7 \u00b1 13.1\r\n \r\n 79.8 \u00b1 15.0\r\n \r\n 0.08\r\n \r\n \r\n \r\n Adjusted regression analysis on the total cohort indicated no difference in the Scaled Scores between the siblings and the diabetes groups (Table 4).Table 4: Regression analysis on quality of life Generic Scaled Scores comparing patients with type 1 diabetes and siblings (n=77). Statistical analyses have adjusted for age group, gender, ethnicity, and New Zealand Deprivation Index. Model-adjusted mean estimate of difference in Scaled Scores between two groups, standard error of mean (SEM), and 95% confidence interval (CI) are reported with p-value. \r\n \r\n \r\n \r\n \r\n \r\n Estimate\r\n \r\n SEM\r\n \r\n 95% C.I.\r\n \r\n p-value\r\n \r\n \r\n \r\n Generic Scaled Score\r\n \r\n -4.37\r\n \r\n 3.16\r\n \r\n -10.67, 1.92\r\n \r\n 0.17\r\n \r\n \r\n \r\n Physical Scaled Score\r\n \r\n -1.55\r\n \r\n 3.72\r\n \r\n -8.97, 5.87\r\n \r\n 0.68\r\n \r\n \r\n \r\n Psychosocial Scaled Score\r\n \r\n -5.92\r\n \r\n 3.32\r\n \r\n -12.54, 0.69\r\n \r\n 0.08\r\n \r\n \r\n \r\nFor pre-defined confounding variables, the Deprivation Index was a significant confounding factor in all regression models on the total cohort, indicating a strong negative association with the Scaled Scores. That is, children living in more socially deprived areas had poorer HRQOL, regardless of having diabetes or not. The mean Generic Scaled Score was 1.82 lower (95% CI [-3.26, -0.39]; p-value 0.01) with per unit increase in Deprivation Index. Compared with male patients, female patients also had a significantly lower Psychosocial Scaled Score (mean -8.11, 95% CI [-14.47, -1.76]; p-value 0.01) and Generic Scaled Score (mean -6.91, 95% CI [-12.96, -0.86]; p-value 0.03). There was no association on HRQOL found with age group and ethnicity.For those type 1 diabetes patients, three potential predictors of interest were fitted in the adjusted regression model to see whether they had any predictive effect on the scaled scores. The mean HbA1c had a significant effect on the Physical Scaled Score (mean -0.61, 95% CI [-1.01, -0.21]; p=0.004), and the Generic Scaled Score (mean -0.32, 95% CI [-0.63, -0.01]; p-value 0.04). Any hospital admission for diabetes was a significant predictor of higher HRQOL in all domains (Table 5). Using the HbA1c ranges of control did not show any significant association with the scaled scores. None of the confounding factors were statistically significant in these models.Table 5: Regression analysis on quality of life Generic Scaled Scores among patients with type 1 diabetes (n=42), with the predictors of interest. Statistical analyses have adjusted for age group, gender, ethnicity, and New Zealand Deprivation Index. Model-adjusted mean estimate of change in Scaled Scores associated with each predictor, standard error of mean (SEM), and 95% confidence interval (CI) are reported with p-value. \r\n \r\n \r\n \r\n Predictors of quality of life scores\r\n \r\n Estimate\r\n \r\n SEM\r\n \r\n 95% C.I.\r\n \r\n p-value\r\n \r\n \r\n \r\n Generic Scaled Score\r\n \r\n \r\n \r\n HbA1c (mean)\r\n \r\n -0.32\r\n \r\n 0.15\r\n \r\n -0.63, -0.01\r\n \r\n 0.04\r\n \r\n \r\n \r\n Diabetes duration\r\n \r\n 0.78\r\n \r\n 0.63\r\n \r\n -0.51, 2.06\r\n \r\n 0.23\r\n \r\n \r\n \r\n Admission for diabetes\r\n \r\n 14.99\r\n \r\n 5.61\r\n \r\n 3.51, 26.5\r\n \r\n 0.01\r\n \r\n \r\n \r\n Physical Scaled Score\r\n \r\n \r\n \r\n HbA1c (mean)\r\n \r\n -0.61\r\n \r\n 0.19\r\n \r\n -1.01, -0.21\r\n \r\n 0.004\r\n \r\n \r\n \r\n Diabetes duration\r\n \r\n 1.17\r\n \r\n 0.82\r\n \r\n -0.50, 2.84\r\n \r\n 0.16\r\n \r\n \r\n \r\n Admission for diabetes\r\n \r\n 18.62\r\n \r\n 7.28\r\n \r\n 3.74, 33.5\r\n \r\n 0.02\r\n \r\n \r\n \r\n Psychosocial Scaled Score\r\n \r\n \r\n \r\n HbA1c (mean)\r\n \r\n -0.15\r\n \r\n 0.16\r\n \r\n -0.47, 0.17\r\n \r\n 0.35\r\n \r\n \r\n \r\n Diabetes duration\r\n \r\n 0.53\r\n \r\n 0.66\r\n \r\n -0.83, 1.88\r\n \r\n 0.44\r\n \r\n \r\n \r\n Admission for diabetes\r\n \r\n 12.84\r\n \r\n 5.93\r\n \r\n 0.72, 25.0\r\n \r\n 0.04\r\n \r\n \r\n \r\n DiscussionThis study represents the perceived HRQOL of Taranaki children and young people with diabetes. Overall, there was no difference in HRQOL found between the diabetes group and their siblings. Females however, had lower overall HRQOL and psychosocial quality of life compared to males in both the sibling and diabetes groups. The psychosocial component makes up the majority of the PedsQLTM Generic 4.0 questionnaire, with questions relating to emotions, fears, school and social functioning. The observed poorer HRQOL in females is found in many other studies7,11,24 and may reflect eventual higher rates of psychological diagnoses, such as depression, somatic complaints and anxiety in women.Higher deprivation scores in the diabetes group compared to sibling group likely reflects the greater number of households sampled in the diabetes group, as those diabetes patients without siblings were still included in the study.Poorer diabetes control was associated with significantly lower physical quality of life and overall HRQOL with increasing HbA1c. This finding is replicated in other larger studies.11,25,26 Persistently raised blood glucose levels contribute to adverse effects on mood and coordination, and possibly neurocognitive function, but long- term studies are minimal in children and adolescents.23 A Swiss study did find that in boys with type 1 diabetes, there was a significant decline in verbal intelligence quotient between age 7 and 16 years if diagnosed before age 6, and this was correlated with high long-term HbA1c and degree of metabolic deterioration at diagnosis. These findings were not replicated in girls, or boys diagnosed after 6 years.27An unexpected finding was that a diabetes-related hospital admission was associated with higher HRQOL scores. This finding has not been described elsewhere, and may be due to the small numbers in this study. Information on whether the admission was due to diabetic ketoacidosis, poorly-controlled blood glucose levels, or an episode of hypoglycaemia was outside the scope of the study. It may be that not all admissions for diabetes have equal impact. Intensive education and support gained while on the ward perhaps has a beneficial effect on those patients.Parent-proxy and child self-report responses were strongly correlated for both the diabetes and sibling groups. Self-reports were not collected on the younger children (2-7 years), but given the correlation found with our older children, it was assumed that parent responses in these cases were a representative reflection of their childs. Other studies have found that parents tend to report their childs HRQOL as being poorer and more restricted by the burden of chronic disease than do the children themselves.8The main limitations of this study were the small sample size, and the lack of a population-based control group. The raw scores of our sibling group (Generic Scaled Score mean 80.8 \u00b1 13.97), however, are comparable to a large population study of children using the same PedsQLTM 4.0 instrument (10,241 children aged 2-16 assessed at enrolment in Californias Childrens Health Insurance Program 2001-2003; Generic Scaled Score mean 81.3 \u00b1 15.9).28 Siblings, with the same family background and living in identical environments, may actually be more closely correlated to our patient group than a group of non-related controls. It is also difficult to take into account the impact on a siblings quality of life by having a brother or sister with a chronic illness, such as diabetes, in the family. We were unable to conduct a paired one-to-one comparison with diabetes patients and their siblings, as there were many with no, unwilling or ineligible siblings. Other potential limitations include undertaking the questionnaire before or after the clinic appointment, and the use of parent reports for analysis. With a relatively small regional population, the decision to use parent reports was deemed justified by the authors. However, while there was good correlation between child-parent reports, it is acknowledged these will not be identical.Diabetes is a life-long illness with potentially major effects on a childs physical and mental health. Impacts on psychological health may be difficult to quantify and HRQOL assessment is one useful tool in the evaluation of a patients well-being.Factors such as family dynamics, parental separation, sibling relationships, behavioural problems, and school performance should be explored. It may be a diagnosis of diabetes un-masks or worsens underlying psychosocial stressors. Early identification of these is important for families to be given support and coping strategies.In summary, children and adolescents with type 1 diabetes reported a quality of life surprisingly similar to their siblings. While siblings might be adversely affected by having a family member with diabetes, these results were similar to the limited background international population data available. These results are encouraging as type 1 diabetes may not adversely affect quality of life to the degree expected, but need to be interpreted with caution, given the lack of a population-based sample. The current strategies used in paediatric diabetes care may be effective in at least addressing some of the psychological challenges children with diabetes face. It does not minimise, however, the burden of psychological stress experienced in this population, and the need for access to appropriate psychological services. Further study of HRQOL with parent and child reports in a larger cohort of New Zealand children and adolescents with diabetes is warranted.\r\n

A diagnosis of type 1 diabetes leads to a significant change in lifestyle for the affected child or teenager and their family. Considerable and unrelenting personal and family stress arise from the regular administration of subcutaneous insulin, frequent blood glucose testing, monitoring, dietary alterations and management of hypo- and hyperglycaemia. It is not surprising, therefore, that type 1 diabetes would be associated with a reduction in an individuals health-related quality of life (HRQOL).Quality of life has been defined as a broad range of human experience related to ones overall well-being\u2026 defined by subjective experiences, states and perceptions. 1 It is increasingly recognised as an important part of any assessment of health and well-being,2 and in the overall understanding of chronic illness. A number of different measurement tools have been validated and proven useful to detect hidden morbidities, improve patient-doctor communication, and improve clinical decision-making.3 Some advocate the assessment of HRQOL as part of best clinical practice, and quantifying HRQOL may predict future health-care costs.4 HRQOL tools can allow for comparisons between healthy populations and those struggling with chronic conditions, such as asthma or diabetes.5,6The impact of living with type 1 diabetes has been well described, and it is generally reported that children and adolescents with diabetes show sub-optimal quality of life compared with their peers.7-11 The concern is not only around negative effects on day-to-day functioning, but whether impairment in quality of life reflects long-term impacts on mental health. A number of studies in children with diabetes have found poorer HRQOL, and later higher psychiatric morbidity compared with a group of age- and gender-matched controls.12-14 However, a meta-analysis of 22 studies found only a mildly elevated overall risk of psychological difficulties, including depression, anxiety and behavioural problems in children with diabetes.15 HRQOL has not previously been studied in New Zealand children with diabetes.The primary objective of this study was to evaluate the quality of life of children and adolescents with type 1 diabetes in the Taranaki region (New Zealand) compared to a group of their siblings without diabetes or chronic disease.MethodsParticipantsAll children and adolescents with type 1 diabetes in the Taranaki region and their parents were approached during the multi-disciplinary diabetes clinics in May-July, 2013. At this time, there were 67 children and adolescents with diabetes in the region (background population 23,127 children/young people less than 15 years of age).16 All patients were managed by a team of paediatricians, a dietitian, and a diabetes nurse educator, who were able to refer patients to Child and Adolescent Mental Health Services (CAMHS), if deemed necessary. Study approval was obtained by the Taranaki Base Hospital Clinical Board, and the Mori Health Unit. The study was classed as an audit for study purposes by the New Zealand Health and Disability Ethics Committee.Exclusion criteria for diabetes group analysis included: those diagnosed within the last 6 months; those with type 2 diabetes or cystic fibrosis (CF)-related diabetes; and those with psychological or other medical conditions (active CAMHS involvement, complex congenital heart disease, and attention deficit hyperactivity disorder [ADHD]).The Pediatric Quality of Life Inventory (PedsQLTM) measurement model was specifically designed to evaluate HRQOL in children and adolescents, with both parent-proxy and child self-report versions. The 23-item Generic Core Scale questionnaire is divided into four areas, assessing problems over the preceding month related to physical, emotional, social and school functioning.17-19 The reliability and content validity of this instrument has been demonstrated.20,21 The developers of the PedsQLTM propose using one standard-deviation below the population mean as a meaningful cut-off for those at risk of impaired HRQOL compared to other children.20Parent-proxy questionnaires in all age groups (2-17 years) and patient self-report questionnaires in those aged 8-17 years were completed in the waiting room prior to, or after, clinical appointments. Parent-proxy questionnaires and self-reports for those aged 8-16 years were also collected for all siblings. The same parent answered the proxy reports for all participating offspring, so if there was more than one sibling, all were included. Questionnaires were also collected from diabetes patients who were without siblings, or whose siblings were outside the age range (2-17 years).Demographic information was collected and New Zealand Deprivation Index score was calculated based on home address, using 2006 census data on household income. The score is divided into deciles; one representing the least deprived area and 10 the most deprived.22 Diabetes-related information was collected, including clinical details, duration of disease, average glycated Haemoglobin (HbA1c) over past year, associated diagnoses, and hospital admissions (admissions at diagnosis were excluded in analyses). HbA1c was collected in clinic, using point-of-care testing with the DCA 2000 Analyzer (Siemens Medical Solutions Diagnostics, Puteaux, France).For those patients with diabetes included in analysis, average HbA1c were divided into three levels of control used in clinic, based on the International Society for Pediatric and Adolescent Diabetes (ISPAD) guidelines,23 with optimal control defined as <58mmol/mol (HbA1c <7.5% DCCT standardised), suboptimal control 58-75mmol/mol (7.5-9%) and high risk of metabolic complications>75mmol/mol (>9%).Quality of life measuresThe PedsQLTM 4.0 Generic Core Scale was used for both the sibling and diabetes groups. Item answers were reverse scored and linearly transformed to a 1-100 scale, with higher scores representing a better health-related quality of life (Generic Scaled Score). The Generic Core Total Score results were divided into two outcomes: the Psychosocial Health Summary Score and Physical Health Summary Score. The summary scores were then converted to Psychosocial Scaled Score and Physical Health Scaled Scores.Statistical analysesStatistical analyses were performed using SAS version 9.3 (SAS Institute Inc. Cary NC) and R version 2.15 (R Foundations for Statistical Computing). All statistical tests were two-sided at a 5% significance level. Correlation and agreement between the parent proxy- and child self-reports for those aged 8 years and older, were evaluated using the Spearmans correlation coefficients and the Bland-Altman plots. Any missing parent proxy-reports were replaced with the child self-report, if the information was available (ie, child was>8 years of age).Simple t-tests were first conducted to compare the Scaled Scores between the diabetes group and their siblings. Linear regression models were next fitted to explore the mean difference in scaled scores between the two groups adjusting for important confounding factors (age group, gender, ethnicity, and deprivation index). For the type 1 diabetes group, linear regression models were used to investigate the predictive effects of HbA1c (the mean and levels of control), duration of diabetes and any hospital admission for diabetes on the scaled scores, adjusting for the same confounding factors. Regression coefficients and 95% confidence intervals were estimated.ResultsThis study achieved an 84% response rate in our total diabetes population (n=56/67). Fourteen patients were subsequently excluded (those with type 2 diabetes [n=3], Cystic fibrosis [n=1], ADHD [n=2], CAMHS involvement [n=4], diagnosis within 6 months [n=3] and other medical conditions [n=1]).Results from 42 type 1 diabetes patients were analysed and compared with siblings aged 2-16 years without diabetes or chronic illness who participated in this study.Parent-proxy reports for the PedsQLTM Generic Core Scale were obtained for 93% of those who participated in the diabetes group (n=39/42; 3 adolescents attended clinic without a parent). Self-reports from 100% of older children and adolescents in the diabetes group (aged 8-17 years) were completed (n=35/35).Questionnaires were completed on the sibling comparisons, with parent-proxy reports returned on the whole group and self-reports obtained on 96% of the siblings aged 8-16 years (n=25/26).The demographic data of the participants are given in Table 1. There were no differences in demographic parameters between groups, apart from deprivation index, which was higher in the diabetes group. This arose from some diabetes patients having multiple siblings included, and some with no, unwilling or ineligible siblings (n=38). This was adjusted for in regression analysis.Table 1: Demographic characteristics of patients with type 1 diabetes and their siblings. Data are means \u00b1 SD (ranges) or n (%). \r\n \r\n \r\n \r\n \r\n \r\n Type 1 diabetes\r\n \r\n Siblings\r\n \r\n p-value\r\n \r\n \r\n \r\n N\r\n \r\n 42\r\n \r\n 35\r\n \r\n \r\n \r\n \r\n \r\n Gender (males)\r\n \r\n 25 (60%)\r\n \r\n 16 (46%)\r\n \r\n 0.23\r\n \r\n \r\n \r\n Ethnicity\r\n \r\n \r\n \r\n \r\n \r\n 0.67\r\n \r\n \r\n \r\n New Zealand European\r\n \r\n 30 (71%)\r\n \r\n 28 (80%)\r\n \r\n \r\n \r\n \r\n \r\n Mori\r\n \r\n 8 (19%)\r\n \r\n 5 (14%)\r\n \r\n \r\n \r\n \r\n \r\n Other\r\n \r\n 4 (10%)\r\n \r\n 2 (6%)\r\n \r\n \r\n \r\n \r\n \r\n Age (years)\r\n \r\n 11.5 \u00b1 3.8 (2-17)\r\n \r\n 10.2 \u00b1 3.7 (4-16)\r\n \r\n 0.14\r\n \r\n \r\n \r\n Age group\r\n \r\n \r\n \r\n \r\n \r\n 0.22\r\n \r\n \r\n \r\n 2-7 years\r\n \r\n 7 (16%)\r\n \r\n 9 (26%)\r\n \r\n \r\n \r\n \r\n \r\n 8-12 years\r\n \r\n 15 (36%)\r\n \r\n 16 (46%)\r\n \r\n \r\n \r\n \r\n \r\n 13-17 years\r\n \r\n 20 (48%)\r\n \r\n 10 (29%)\r\n \r\n \r\n \r\n \r\n \r\n Deprivation index\r\n \r\n 5.6 \u00b1 2.4 (1-10)\r\n \r\n 4.5 \u00b1 2.1 (1-10)\r\n \r\n 0.04\r\n \r\n \r\n \r\nThe duration of diabetes ranged from 9 months to 13 years, with a mean of 5.1 years (\u00b1 3.6SD). One third (n=14/42) had at least one diabetes-related hospital admission within the last year. The mean HbA1c in the past year was 72.5 \u00b1 18.5 mmol/mol (8.8% \u00b1 3.9 DCCT standard). Only 12% of patients (n=5/42) had an average HbA1c in the optimal control range, compared with 60% (n=25/42) in the suboptimal control, and 29% (12/42) in the high risk of metabolic complications groups (Table 2).Table 2: Clinical characteristics of patients with type 1 diabetes. Data are means \u00b1 SD (ranges) or n (%).\r\n \r\n \r\n \r\n N\r\n \r\n 42\r\n \r\n \r\n \r\n Diabetes duration (years)\r\n \r\n 5.1 \u00b1 3.6 (0.75-13.58)\r\n \r\n \r\n \r\n HbA1c (mmol/mol)\r\n \r\n 72.5 \u00b1 18.5 (36.8-130)\r\n \r\n \r\n \r\n Glycaemic control (HbA1c)\r\n \r\n \r\n \r\n \r\n \r\n High risk (>75 mmol/mol or>9%)\r\n \r\n 12 (29%)\r\n \r\n \r\n \r\n Suboptimal (58-75mmol/mol or 7.5-9%)\r\n \r\n 25 (60%)\r\n \r\n \r\n \r\n Optimal (<58mmol/mol or <7.5%)\r\n \r\n 5 (12%)\r\n \r\n \r\n \r\n Hospital admissions for diabetes in the past year\r\n \r\n \r\n \r\n \r\n \r\n 0\r\n \r\n 28 (67%)\r\n \r\n \r\n \r\n 1\r\n \r\n 11 (26%)\r\n \r\n \r\n \r\n \u22652\r\n \r\n 3 (7%)\r\n \r\n \r\n \r\nExcluding those diagnosed <6 months, type 2 diabetes, CF-related and associated diagnoses (congenital heart disease, ADHD, CAMHS involvement). Comparing parent-proxy and child self-report responses for those children aged 8 years and older showed significant correlations across all PedsQLTM Generic Scaled Scores using the total cohort. The Spearman correlation coefficient for the Physical Scaled Score was 0.55 (p<0.0001), and for the Psychosocial Scaled Score 0.48 (p<0.0001) respectively. The Bland-Altman plots indicated good level of agreement on all scores, with only small bias of 1.5 and 0.3 between the child and parent reports respectively. These findings have enabled the parent responses to be used for the entire analysis.Table 3 shows the unadjusted total Generic Scaled Scores for the sibling and diabetes patient groups, as assessed by the parent proxy reports.Table 3: Unadjusted quality of life Generic Scaled Scores (out of 100), as assessed by the parent proxy-reports. Data are means \u00b1 SD. \r\n \r\n \r\n \r\n \r\n \r\n Type 1 diabetes\r\n \r\n Siblings\r\n \r\n p-value\r\n \r\n \r\n \r\n n\r\n \r\n 42\r\n \r\n 35\r\n \r\n \r\n \r\n \r\n \r\n Total Generic Scaled Score\r\n \r\n 75.9 \u00b1 13.4\r\n \r\n 80.8 \u00b1 14.0\r\n \r\n 0.14\r\n \r\n \r\n \r\n Physical Scaled Score\r\n \r\n 80.0 \u00b1 17.9\r\n \r\n 82.7 \u00b1 14.2\r\n \r\n 0.52\r\n \r\n \r\n \r\n Psychosocial Scaled Score\r\n \r\n 73.7 \u00b1 13.1\r\n \r\n 79.8 \u00b1 15.0\r\n \r\n 0.08\r\n \r\n \r\n \r\n Adjusted regression analysis on the total cohort indicated no difference in the Scaled Scores between the siblings and the diabetes groups (Table 4).Table 4: Regression analysis on quality of life Generic Scaled Scores comparing patients with type 1 diabetes and siblings (n=77). Statistical analyses have adjusted for age group, gender, ethnicity, and New Zealand Deprivation Index. Model-adjusted mean estimate of difference in Scaled Scores between two groups, standard error of mean (SEM), and 95% confidence interval (CI) are reported with p-value. \r\n \r\n \r\n \r\n \r\n \r\n Estimate\r\n \r\n SEM\r\n \r\n 95% C.I.\r\n \r\n p-value\r\n \r\n \r\n \r\n Generic Scaled Score\r\n \r\n -4.37\r\n \r\n 3.16\r\n \r\n -10.67, 1.92\r\n \r\n 0.17\r\n \r\n \r\n \r\n Physical Scaled Score\r\n \r\n -1.55\r\n \r\n 3.72\r\n \r\n -8.97, 5.87\r\n \r\n 0.68\r\n \r\n \r\n \r\n Psychosocial Scaled Score\r\n \r\n -5.92\r\n \r\n 3.32\r\n \r\n -12.54, 0.69\r\n \r\n 0.08\r\n \r\n \r\n \r\nFor pre-defined confounding variables, the Deprivation Index was a significant confounding factor in all regression models on the total cohort, indicating a strong negative association with the Scaled Scores. That is, children living in more socially deprived areas had poorer HRQOL, regardless of having diabetes or not. The mean Generic Scaled Score was 1.82 lower (95% CI [-3.26, -0.39]; p-value 0.01) with per unit increase in Deprivation Index. Compared with male patients, female patients also had a significantly lower Psychosocial Scaled Score (mean -8.11, 95% CI [-14.47, -1.76]; p-value 0.01) and Generic Scaled Score (mean -6.91, 95% CI [-12.96, -0.86]; p-value 0.03). There was no association on HRQOL found with age group and ethnicity.For those type 1 diabetes patients, three potential predictors of interest were fitted in the adjusted regression model to see whether they had any predictive effect on the scaled scores. The mean HbA1c had a significant effect on the Physical Scaled Score (mean -0.61, 95% CI [-1.01, -0.21]; p=0.004), and the Generic Scaled Score (mean -0.32, 95% CI [-0.63, -0.01]; p-value 0.04). Any hospital admission for diabetes was a significant predictor of higher HRQOL in all domains (Table 5). Using the HbA1c ranges of control did not show any significant association with the scaled scores. None of the confounding factors were statistically significant in these models.Table 5: Regression analysis on quality of life Generic Scaled Scores among patients with type 1 diabetes (n=42), with the predictors of interest. Statistical analyses have adjusted for age group, gender, ethnicity, and New Zealand Deprivation Index. Model-adjusted mean estimate of change in Scaled Scores associated with each predictor, standard error of mean (SEM), and 95% confidence interval (CI) are reported with p-value. \r\n \r\n \r\n \r\n Predictors of quality of life scores\r\n \r\n Estimate\r\n \r\n SEM\r\n \r\n 95% C.I.\r\n \r\n p-value\r\n \r\n \r\n \r\n Generic Scaled Score\r\n \r\n \r\n \r\n HbA1c (mean)\r\n \r\n -0.32\r\n \r\n 0.15\r\n \r\n -0.63, -0.01\r\n \r\n 0.04\r\n \r\n \r\n \r\n Diabetes duration\r\n \r\n 0.78\r\n \r\n 0.63\r\n \r\n -0.51, 2.06\r\n \r\n 0.23\r\n \r\n \r\n \r\n Admission for diabetes\r\n \r\n 14.99\r\n \r\n 5.61\r\n \r\n 3.51, 26.5\r\n \r\n 0.01\r\n \r\n \r\n \r\n Physical Scaled Score\r\n \r\n \r\n \r\n HbA1c (mean)\r\n \r\n -0.61\r\n \r\n 0.19\r\n \r\n -1.01, -0.21\r\n \r\n 0.004\r\n \r\n \r\n \r\n Diabetes duration\r\n \r\n 1.17\r\n \r\n 0.82\r\n \r\n -0.50, 2.84\r\n \r\n 0.16\r\n \r\n \r\n \r\n Admission for diabetes\r\n \r\n 18.62\r\n \r\n 7.28\r\n \r\n 3.74, 33.5\r\n \r\n 0.02\r\n \r\n \r\n \r\n Psychosocial Scaled Score\r\n \r\n \r\n \r\n HbA1c (mean)\r\n \r\n -0.15\r\n \r\n 0.16\r\n \r\n -0.47, 0.17\r\n \r\n 0.35\r\n \r\n \r\n \r\n Diabetes duration\r\n \r\n 0.53\r\n \r\n 0.66\r\n \r\n -0.83, 1.88\r\n \r\n 0.44\r\n \r\n \r\n \r\n Admission for diabetes\r\n \r\n 12.84\r\n \r\n 5.93\r\n \r\n 0.72, 25.0\r\n \r\n 0.04\r\n \r\n \r\n \r\n DiscussionThis study represents the perceived HRQOL of Taranaki children and young people with diabetes. Overall, there was no difference in HRQOL found between the diabetes group and their siblings. Females however, had lower overall HRQOL and psychosocial quality of life compared to males in both the sibling and diabetes groups. The psychosocial component makes up the majority of the PedsQLTM Generic 4.0 questionnaire, with questions relating to emotions, fears, school and social functioning. The observed poorer HRQOL in females is found in many other studies7,11,24 and may reflect eventual higher rates of psychological diagnoses, such as depression, somatic complaints and anxiety in women.Higher deprivation scores in the diabetes group compared to sibling group likely reflects the greater number of households sampled in the diabetes group, as those diabetes patients without siblings were still included in the study.Poorer diabetes control was associated with significantly lower physical quality of life and overall HRQOL with increasing HbA1c. This finding is replicated in other larger studies.11,25,26 Persistently raised blood glucose levels contribute to adverse effects on mood and coordination, and possibly neurocognitive function, but long- term studies are minimal in children and adolescents.23 A Swiss study did find that in boys with type 1 diabetes, there was a significant decline in verbal intelligence quotient between age 7 and 16 years if diagnosed before age 6, and this was correlated with high long-term HbA1c and degree of metabolic deterioration at diagnosis. These findings were not replicated in girls, or boys diagnosed after 6 years.27An unexpected finding was that a diabetes-related hospital admission was associated with higher HRQOL scores. This finding has not been described elsewhere, and may be due to the small numbers in this study. Information on whether the admission was due to diabetic ketoacidosis, poorly-controlled blood glucose levels, or an episode of hypoglycaemia was outside the scope of the study. It may be that not all admissions for diabetes have equal impact. Intensive education and support gained while on the ward perhaps has a beneficial effect on those patients.Parent-proxy and child self-report responses were strongly correlated for both the diabetes and sibling groups. Self-reports were not collected on the younger children (2-7 years), but given the correlation found with our older children, it was assumed that parent responses in these cases were a representative reflection of their childs. Other studies have found that parents tend to report their childs HRQOL as being poorer and more restricted by the burden of chronic disease than do the children themselves.8The main limitations of this study were the small sample size, and the lack of a population-based control group. The raw scores of our sibling group (Generic Scaled Score mean 80.8 \u00b1 13.97), however, are comparable to a large population study of children using the same PedsQLTM 4.0 instrument (10,241 children aged 2-16 assessed at enrolment in Californias Childrens Health Insurance Program 2001-2003; Generic Scaled Score mean 81.3 \u00b1 15.9).28 Siblings, with the same family background and living in identical environments, may actually be more closely correlated to our patient group than a group of non-related controls. It is also difficult to take into account the impact on a siblings quality of life by having a brother or sister with a chronic illness, such as diabetes, in the family. We were unable to conduct a paired one-to-one comparison with diabetes patients and their siblings, as there were many with no, unwilling or ineligible siblings. Other potential limitations include undertaking the questionnaire before or after the clinic appointment, and the use of parent reports for analysis. With a relatively small regional population, the decision to use parent reports was deemed justified by the authors. However, while there was good correlation between child-parent reports, it is acknowledged these will not be identical.Diabetes is a life-long illness with potentially major effects on a childs physical and mental health. Impacts on psychological health may be difficult to quantify and HRQOL assessment is one useful tool in the evaluation of a patients well-being.Factors such as family dynamics, parental separation, sibling relationships, behavioural problems, and school performance should be explored. It may be a diagnosis of diabetes un-masks or worsens underlying psychosocial stressors. Early identification of these is important for families to be given support and coping strategies.In summary, children and adolescents with type 1 diabetes reported a quality of life surprisingly similar to their siblings. While siblings might be adversely affected by having a family member with diabetes, these results were similar to the limited background international population data available. These results are encouraging as type 1 diabetes may not adversely affect quality of life to the degree expected, but need to be interpreted with caution, given the lack of a population-based sample. The current strategies used in paediatric diabetes care may be effective in at least addressing some of the psychological challenges children with diabetes face. It does not minimise, however, the burden of psychological stress experienced in this population, and the need for access to appropriate psychological services. Further study of HRQOL with parent and child reports in a larger cohort of New Zealand children and adolescents with diabetes is warranted.\r\n