The Need for Early Referral: Characteristics of Children and Adolescents Who Are Overweight and Obese Attending a Multidisciplinary Weight Management Service

There is a need to examine the issue of childhood obesity from a systems perspective. This study aimed to describe the baseline characteristics of children attending pediatric multidisciplinary weight management services and understand how this information will inform future service delivery. A total of 51 children and adolescents who were overweight and obese (27 male) and aged between two and 16 years participated. Body size measures such as body mass index (BMI) and body fat percentage were collected. Participants and their parents/guardians also completed questionnaires on dietary intake, behaviors and habits, physical activity and health-related quality of life. A total of 72% of participants were classified as morbidly obese. Adolescents had significantly lower scores for overall diet, physical activity and particular health-related quality of life scores. No significant correlations were found between BMI z-scores and diet, physical activity and health-related quality of life. In adolescents, correlations were detected between dietary scores and health-related quality of life. Results confirm the need to critically examine the current context to adapt and tailor interventions to individual circumstances, and when combined with focused referral, triaging and screening processes, should assist in delivering the right care at the right time.


Introduction
There is a clear need to understand the issue of childhood obesity from a systems perspective. Overweight and obesity in childhood and adolescence remains a critically important health care issue, both in Australia and internationally [1,2]. Obesity is a complex health issue, with many social, environmental, biological, and individual factors contributing to weight gain across the life span [3]. Approximately one in four Australian children aged five to 17 years (27.4%) are classified as overweight or obese [1], which is similar to the average rate for the countries that form the Organization for Economic Co-Operation and Development (OECD), at approximately 23% [4]. While there appears to be no change in the proportion of children who were overweight or obese since 2011-2012 [1], the associated health risks remain considerable. Pediatric overweight and obesity can significantly affect processes are used within the system, and how this can help us to provide clearer direction for services, modify clinical practices, up-skill staff, assess effectiveness and design more appropriate interventions. In turn, it is hoped that this will reduce attrition rates and contribute to a sustainable service. The aim of this study, therefore, was to describe the baseline characteristics of children attending a pediatric multidisciplinary weight management clinic over a one-year time frame at a tertiary hospital and satellite site in Queensland, Australia, in the context of health services implementation research.

Materials and Methods
This prospective cross-sectional study was part of a larger project investigating the expansion, evaluation and sustainability of a public multidisciplinary pediatric weight management service in Queensland, Australia. Ethics approval was obtained from the Human Research Ethics Committees at both Children's Health Queensland Hospital and Health Service (CHQ HHS) and Bellberry Limited respectively. The study is registered under the Australian New Zealand Clinical Trials Registry (ID: ACTRN12616000101482) and is a pragmatic research trial. Written informed consent was provided by all parents/guardians and obtained from participants where appropriate. Data collection occurred between June 2016 and June 2017 at two clinic locations: Lady Cilento Children's Hospital (LCCH), a large central service and at a satellite site, The University of Queensland (UQ) Ipswich Nutrition and Dietetics practice (IPSWICH), both situated in Queensland, Australia. All children and their parents/guardians attending the clinic were invited to participate in the study. Inclusion criteria for acceptance in the clinics were the following: (i) referred by a pediatrician, primary care provider (PCP), member of the education department or allied health and aged between 0 and 18 years; and (ii) participants less than two years of age were required to be above the 97th percentile on the WHO weight-for-age growth charts [16] and gaining weight rapidly, or been assessed by a pediatrician or PCP as having weight concerns and above the healthy weight range; participants older than two years of age were required to have a body mass index (BMI) >85th percentile on the United States Centers for Disease Control and Prevention (US-CDC) charts [17] or the WHO charts [18]. Referral guidelines to the clinic were based on National Health and Medical Research Council (NHMRC) Clinical Practice Guidelines for the Management of Overweight and Obesity in Adults, Adolescents and Children in Australia [3]. Children with serious co-morbidities requiring urgent weight management or a suspected underlying metabolic or endocrine cause were eligible to participate in the study. Families who could not understand English well, and where an interpreter was not available were excluded from the study. The research data, including demographic information, was collected either immediately after the first clinic attendance, or at a separate appointment that was scheduled within two weeks of the first attendance.
The multidisciplinary weight management clinic offered the same service to children, adolescents and families across the two sites, and involved a medical practitioner (pediatrician or endocrinologist), dietitian and psychologist. It was an intensive, clinical management program in a real-life setting, and all advice and guidance provided was individualized and evidence-based, following recommendations from the Clinical Practice Guidelines for the Management of Overweight and Obesity in Adults, Adolescents and Children in Australia [3]. Advice and guidance included areas such as diet modification, nutrition counselling, promotion of an active lifestyle, behavior modification strategies, supportive parental involvement, and a comprehensive medical evaluation.

Anthropometric Measures
Height was measured to the nearest 0.1 cm using a stadiometer (Seca Model 213 or 264, Seca, Hamburg, Germany) and weight to the nearest 0.1 kg was collected using digital scales (Seca Medical Scales Model 876, Seca, Hamburg, Germany) with participants wearing light clothing and no shoes. BMI (kg/m 2 ) was calculated using the above information and BMI z-score for age and gender was determined via US-CDC or WHO z-score data files [16][17][18]. The International Obesity Taskforce (IOTF) classifications were used to categories BMI [19]. A measure of percentage body fat was collected using foot-to-foot bioelectrical impedance analysis (BIA) scales.

Dietary Intake, Behaviours and Habits
All parents/guardians or participants, depending on age and capacity, completed a paper-based version of the E-KINDEX (Electronic Kids Dietary Index) questionnaire [20]. This tool was selected to assess a child's obesogenic habits, and contains three sections; dietary intake, dietary behaviors and eating beliefs, and dietary practices or habits. The overall maximum score is 87 with higher scores indicating a healthier food intake as well as behaviors and habits associated with a lower BMI [20].

Physical Activity
Physical activity (PA) was assessed via paper-based questionnaires and completed by participants or parent/guardian depending on age and capacity. The validated pre-school age physical activity questionnaire (Pre-PAQ) was used to assess young children who were not yet of school age [21]. For this study, only question 34 and 37 were completed, as these are the only scored questions, which have been previously validated against accelerometer data at the group level [21]. The other questions provide demographic, contextual and parental PA information which had either been collected, or was not of interest to our study. Scoring of questions 34 and 37 were reported as total minutes of Level 3 (light activity) to Level 5 (moderate to vigorous) per day [21]. The physical activity questionnaire for older children (PAQ-C) and physical activity questionnaire for adolescents (PAQ-A) were used for participants in primary school (approximate age 6-13 years) and high school (approximate age 13-17 years) respectively [22,23]. These valid and reliable questionnaires provide a general measure of physical activity and generate an overall score from one to five, with one indicating low PA, and five indicating high PA [22,23].

Health-Related Quality of Life
Health-related Quality of Life (HRQOL) was measured using the paper-based PedsQL™ Measurement Model for the Pediatric Quality of Life Inventory™ questionnaire, validated for ages 2-18 years [24]. PedsQL™ has separate questionnaires for different age groups (2-4 years, 5-7 years, 8-12 years and 13-18 years), with a parent proxy and child version for each age group with the exception of children aged 2-4 years. Each parent/guardian and the participant (depending on age and capacity) completed the questionnaire separately. The questionnaire is separated into four sections: physical, emotional, social and school functioning. From this, three scores are generated; physical, social (emotional, social and school sections) and an overall score. The maximum for each score is 100, with a higher score representing a higher HRQOL [24].

Statistical Analyses
Statistical analyses were performed using Statistical Package for the Social Sciences version 23 (IBM SPSS Statistics 23.0). Standard diagnostics tests such as normality and homogeneity of variance were performed. All tests were two tailed and statistical significance was set at an α value of 0.05 apart from Mann-Whitney U tests to identify differences between the three age groups and here it was set at 0.017 to avoid Type 1 error. Participants were grouped according to age: 0-4.99 years (young children), 5-11.99 years (children) and 12-17.99 years (adolescents) and descriptive statistics determined. Differences in variables between groups were assessed using either one-way ANOVA with the post-hoc Tukey tests, Kruskal Wallis tests with post-hoc Mann-Whitney U tests, t-tests and Mann-Whitney U tests (for variables compared across two age groups only) or via ANCOVA with post-hoc multiple comparisons, adjusted for the Index of Relative Socio-Economic Disadvantage Score (IRSD). Differences between child-reported and parent proxy-reported quality of life scores were calculated using paired samples t-tests and Wilcoxon signed rank test. Relationships between key variables were assessed using partial correlation, controlling for the effect of IRSD scores.

Results
Fifty-one children and adolescents (53% male) ranging in age from 1.6 years to 16.3 years participated in the study over a one-year period (July 2016 to June 2017). At the time of assessment, 53% of the participants attended the LCCH clinic, and 47% attended the IPSWICH clinic. Based on the IOTF classification for BMI for age and gender, 14% (n = 7) were classified as overweight, 14% (n = 7) as obese and 72% (n = 37) as morbidly obese [19]. A total of 28 participants (55%) lived in areas that indicate relatively greater disadvantage according to the IRSD score [25]. Medical specialists within the LCCH were the most frequent referral source (n = 18 participants, 35%), followed by paediatricians within the LCCH (n = 9, 17.5%), external general practitioners (n = 7, 14%), external dietitians (n = 7, 14%), obesity specialists within the LCCH (n = 6, 11.5%) and external paediatricians (n = 4, 8%). All participants were either Anglo-Celtic or East Asian in ethnicity.
When grouping participants according to age, 12% (n = 6) were aged between 0 and 4.99 years (young children), 45% (n = 23) were aged between 5.00 and 11.99 years (children), and 43% (n = 22) were aged between 12.00 and 17.99 years (adolescents). Comparisons were made between the three age groups for anthropometric variables (Table 1), and dietary, PA and HRQOL results ( Table 2). Young children had the highest BMI z-scores of all participants (p < 0.006), and there were no significant differences between BMI z-scores for children and adolescents. No differences in body fat percentage were found between children and adolescents, and median body fat percentage results for both groups were considerably higher than the 98th centile for all ages and genders on current reference curves [26]. Body fat percentage was not able to be measured for all young children due to difficulties in standing still for assessment (Table 1). When considering the overall dietary score, adolescents scored significantly lower compared to children (mean difference (MD) = 6.6, 95% confidence of interval (CI) = 0.24-13.0, p = 0.039). Young children had the highest scores for dietary behaviors and eating beliefs, significantly greater than scores for adolescents (MD = 5.7, 95% CI = 0.94-10.47, p = 0.014). Dietary intake scores were similar across all age groups ( Table 2).
Comparisons for PA could only be made between the children and adolescents, due to the different questionnaire used for the young children. Adolescents had significantly lower PA levels when compared to children (MD = 0.48, 95% CI = 0.1-0.86, p = 0.014) ( Table 2).
Results regarding HRQOL showed that parent proxy-reported social HRQOL was significantly lower for adolescents when compared to young children (MD = 23.70, 95% CI = 2.30-45.08). No other significant differences were found between the age groups for child or parent proxy-reported HRQOL. When comparing the child and parent proxy-reported HRQOL scores within groups, parent scores were generally lower overall and for each section. For the children, parent proxy-reported HRQOL was lower for overall (f (20) = 2.75, p < 0.05) and social scores (z = −2.016, p < 0.05). Within the adolescents, parent proxy-reported HRQOL was lower overall (z = −2.25, p < 0.05) and for both physical (z = −2.23, p < 0.05) and social scores (f (18) = 2.55, p < 0.05) ( Table 2). BMI z-score was compared to HRQOL, dietary overall and PA scores across age groups, but no obvious significant relationships were found. The relationship between HRQOL and dietary measures was investigated, with significant findings in the adolescent age group (Table 3). A strong, positive correlation was evident between adolescent-reported overall HRQOL and overall diet score (r = 0.709, p < 0.01), dietary intake score (r = 0.604, p < 0.01) and dietary behaviors and eating beliefs score (r = 0.496, p < 0.05). A medium to strong correlation was found between parent proxy-reported HRQOL and dietary practices or habits score (r = 0.541, p < 0.05) in the adolescent aged group. No relationships between HRQOL and dietary measures were evident in the children and young children age groups.

Discussion
The aim of this study was to describe the baseline characteristics of children attending a pediatric multidisciplinary weight management clinic in Queensland, Australia, in the context of health services implementation research. The authors wanted to understand, at the systems level, what strategies are working to support the prevention and management of pediatric overweight and obesity in Queensland, and what needs to change. Results showed that children entering the clinics were not assessed on parameters of diet, physical activity or HRQOL prior to referral, and were considerably overweight as per international definitions, with 72% of participants classified as morbidly obese. This may have been due, in part, to triaging practices within the system, where children classified as overweight were generally placed into a dietitian-only clinic, rather than a multidisciplinary clinic. These results may also reflect referring practitioner decisions, who may only be referring children and adolescents when their BMI is within the obese category and other concerns are apparently obvious. This is supported by the data which show that only 14% (n = 7) of children or adolescents were referred by an external general practitioner, compared to 65% (n = 33) referred by a specialist or pediatrician within the LCCH. Of these 33 children and adolescents, data regarding their contact with the public or private health care system was not collected, so it is unclear why concerns regarding weight were not flagged earlier. Ideally, this should be happening at the primary care level. Additionally, the sample size did not include children or adolescents of ethnicities considered to be at higher risk of overweight and obesity in Australia, such as Indigenous Australians or those from Maori and Pacific Islander cultures [27,28]. Ensuring timely identification, referral and management for these children and adolescents is critical, and needs to be investigated further from a systems-level.
Within this cross-sectional study, measures of diet, PA and aspects of HRQOL generally declined in the older age groups. Scores for overall diet, behavior and eating beliefs and practices or habits sections of the dietary questionnaire were significantly lower in adolescents, and we speculate that this could be related to increased exposure to discussions of weight, previous dieting attempts or engaging in more obesogenic habits such as skipping breakfast [29]. Dietary intake scores were fairly similar across the three age groups. This stability in scores may reflect the measurements within the tool, rather than the absence of any differences, as the tool measures types of food and the number of times the foods are consumed across a week, rather than quantity which is typically highest in adolescents [29].
PA was also significantly lower within the adolescent age group which matches trends found in national health surveys in Australia [30]. While scores were significantly different, both children and adolescents were still doing only low amounts of activity when compared to recommendations [3].
Outcomes concerning HRQOL showed that parents proxy-reported lower HRQOL scores when compared to reports by their children. Within the adolescent age group, dietary scores (including overall diet, dietary intake, and dietary behaviors and eating beliefs) increased as self-reported HRQOL scores increased, highlighting an important relationship that was not evident in the children in the study. This has been described previously in the literature in a large prospective study in Australia by Jacka et al. [31] who found a relationship between unhealthy dietary patterns and poorer mental health in adolescents-higher unhealthy diet scores at baseline predicted lower HRQOL scores (using the same tool as the current study) at follow up one year later, and improvements in diet quality were reflected by improvements in mental health [31]. These outcomes support the need for new triaging and screening options to identify those adolescents who would benefit from targeted, intensive psychological input.
The absence of any significant relationships between BMI z-score and diet, PA and HRQOL measures was unexpected. It was thought that these results may have provided an interesting insight into what to target in multicomponent interventions for specific age groups. These findings may be due to the fact, that 72% of the cohort were classified as morbidly obese, and the variation in BMI z-scores across all age groups was small.
Clearly, more research is required to understand why children and adolescents are not being identified and referred to services in a timely manner, and what can be done from a systems perspective to change this practice. This is important, given that recent evidence from similar multidisciplinary clinics in real-life settings on an international level suggests that treatment outcomes are better when children were younger [32] and that early intervention for weight management is preferred and considered more cost-effective [15]. Consideration also needs to be given to ensure that the most appropriate triaging guidelines are utilized once children and adolescents have been referred to services. Currently, triaging guidelines for the LCCH and IPSWICH clinics are generally based on weight and associated co-morbidities only. The results of this study, however, show that other factors such as age, diet, PA and/or HRQOL should also be considered. Screening for these factors during the triaging process can help determine the level of support and intensity required for patients, which will in turn contribute to an efficient use of services, which helps contain costs. Results also confirm that clinics cannot use a one-size-fits-all approach to treatment and management given the differences evident at commencement of the clinic between the groups of children. Services may need to consider if the formation of an adolescent-specific clinic, where psychological input occurs more frequently that for the younger children, would be a beneficial use of resources given their clinical and service context.
The main limitation of this study is the small sample size potentially causing insufficient power to detect correlations. The population of interest is difficult to recruit, and attrition is high, resulting in small sample sizes which may not be representative of the larger population of children and adolescents who are overweight or obese. Missing data was also a common theme as a result of unavoidable reasons such as child age or behavior. Another limitation was that outcome measures for dietary intake, behaviors and habits and physical activity were self-reported rather than objectively measured. Given the pragmatic nature of the study, and the limitations when embedding research into clinical practice, this was a necessary compromise. Furthermore, the E-KINDEX tool has, at this stage, only been validated for children aged 9 to 13 years [20]. Lastly, due to the choice of tools for physical activity, comparisons were unable to be made between young children and the other two age groups.

Conclusions
There is no single solution to the problem of pediatric overweight or obesity, however it remains a significant health problem and a focus on timely identification of concerns, appropriate referral systems, and dedicated services is critical. Results of this study found that adolescents are already experiencing detrimental consequences related to being overweight or obese, such as lower scores related to diet and dietary practices or habits, and lower levels of PA. Results identified that system referral processes are not effective, and confirm the need to critically examine the current context to further investigate these processes and adapt interventions to individual circumstances, supported by focused triaging and screening processes to enhance the opportunity for individual success. There is a clear need to implement these research findings for system-wide change to enable efficient, effective service delivery options at all levels of health care that can be up-scaled throughout the public health system.