Medical care of overweight children under real-life conditions: the German BZgA observation study

Abstract

Objective:

Current care for overweight children is controversial, and only few data are available concerning the process of care, as well as the outcome under real-life conditions.

Methods:

A nationwide survey of treatment programs for overweight children and adolescents in Germany identified 480 treatment centers. From 135 institutions that had agreed to participate in this study of process of care and outcome, 48 randomly chosen institutions were included in the study. All 1916 overweight children (mean age 12.6 years, 57% female, mean body mass index 30.0 kg/m2), who presented at these institutions for lifestyle interventions, were included in this study. Diagnostic procedures according to guidelines and effect of lifestyle interventions on weight status at end of treatment were analyzed.

Results:

Children treated <3 months were older and more obese, whereas children with >3 months treatment duration demonstrated more cardiovascular risk factors at baseline. On the basis of an intention-to-treat analysis, 75% of the children reduced their overweight. The reduction of overweight varied widely between the treatment institutions (intracluster correlation coefficient 0.15 in the multiple regression model reflecting the intracenter correlation). Screening for hypertension, disturbed glucose metabolism and dyslipidemia was performed in 52% of the children at baseline and in 10% at the end of intervention.

Conclusion:

Overweight reduction is achievable with lifestyle intervention in clinical practice. However, because the clientele, treatment approach and outcome varied widely between different institutions, and screening for comorbidities was seldomly performed as recommended, quality criteria for institutions have to be implemented to improve medical care of overweight children under real-life conditions.

Introduction

Current care for overweight and obese children is controversial, and only few data are available comparing the process of care, as well as the outcome, among different institutions with different treatment approaches. National and international guidelines concerning overweight children and adolescents recommend multidisciplinary behavior-oriented long-term lifestyle interventions, as well as measurements of cardiovascular risk factors such as blood pressure, blood glucose and lipids.1, 2, 3 However, only half of the German treatment institutions for children and adolescents measured these parameters in their overweight patients as demonstrated in a previous study in the year 2003.4

Most importantly, the effect of lifestyle intervention in overweight children is discussed controversially.5 Successful intervention studies in overweight children were conducted in single specialized treatment centers under ideal circumstances of a study trial questioning the generalizability of the findings, whereas studies performed in clinical practice under real-life conditions are very scarce. Treatment approaches are very heterogeneous, as in addition to the recommended longitudinal multidisciplinary behavior-oriented outpatient care, single diet or exercise therapy as well as inpatient rehabilitation for a period of 4–6 weeks is offered.6 Nearly 10 000 overweight children are treated in Germany yearly, most of them (>70%) in inpatient rehabilitation centers.6 The treatment experience is limited in most institutions, especially in outpatient treatment centers, and only six institutions in Germany have examined their treatment concept so far.6 Therefore, objective data reflecting current care in clinical practice under real-life conditions are urgently needed.

In 2005, a large observational study was initiated and financed by the BZgA (Federal Centre for Health Education) comparing patient selection, process of care (evaluation of comorbidities, type and intensity of education) as well as the outcome between different categories of lifestyle intervention programs separated by structural and quality criteria. The main aims of this study of medical care in real-life were to (1) determine the frequency of diagnostic procedures according to guidelines in general practice, (2) to study the clinical characteristics of overweight children treated in Germany, (3) to analyze the effects of different treatment approaches on weight status and comorbidities and (4) to identify predictive factors for treatment success. In this analysis, we focus on the quality of diagnostic procedures and the outcome at the end of interventions.

Methods

Initiated by the Federal Center for Health Education (BZgA) in 2004, a nationwide survey of treatment programs for overweight or obese children and adolescents in Germany identified 480 treatment centers.7 All institutions were contacted to participate in the following study. From the 135 institutions that agreed to participate, 52 were randomly chosen using a random number procedure, aiming at homogeneous distribution of patients among treatment strategies (for example, <3 months treatment duration and 3 months treatment) and reflecting currently available therapeutic programs. Four institutions discontinued the study because the treatment center was closed in the observation period. The patients of these institutions were excluded from further analyses. All overweight children aged 8 to <17 years, who presented to the remaining 48 treatment institutions between August 2005 and August 2006 for participating in a lifestyle intervention, were included in this study. The treatment institutions were divided into two groups according to treatment duration: group A, <3 months treatment duration and group B, 3 months treatment duration.

Anthropometrical data were measured at each institution at baseline and at the end of intervention. These data were transferred anonymously to the study center by a software for standardized documentation.4, 5, 6, 7, 8 Furthermore, the available data concerning the cardiovascular risk factor profile (blood pressure, fasting glucose and lipids) were transferred anonymously to the study center by the software. This software, the APV software, was started in 2001. A German demonstration version of the software can be downloaded from the APV website http://www.a-p-v.de. The software can be installed on single computers, on institutional networks (intranet) as well as on an internet server with secure remote access. The latter option is suitable when patient education, training and medical care are provided simultaneously by multiple, geographically separated institutions.

In addition, quality of life, the psychological situation, eating and exercise behavior, the social status determined by questionnaires for children and parents, as well as intensity of intervention in several categories (medical care, dietary counseling, psychological interventions, sports therapy, all for patients and/or parents) were measured. These data are not in the focus of this analysis and will be published separately.

Written informed consent was obtained from all children and their parents. The study was approved by the local ethics committee of the University of Ulm.

Overweight and obesity were defined according to the definition of the International Task Force of Obesity using population specific data.9, 10 Overweight was defined as a body mass index (BMI) above the 90th percentile, and obesity as a BMI of the 97th percentile. Because BMI is not normally distributed in childhood, we used the LMS method to calculate s.d. score of BMI (SDS-BMI) as a measurement for the degree of overweight. The LMS method summarizes the data in terms of three smooth age-specific parameters for a family of curves called L (lambda), M (mu), and S (sigma) based on German population-specific data.10, 11 The M and S parameters correspond to the median and coefficient of variation of BMI for German children at each age and gender, whereas the L parameter allows for the substantial age- and gender-dependent skewness in the distribution of BMI.10, 11 The assumption underlying the LMS method is that after Box–Cox power transformation the data at each age are approximately normally distributed.11

Blood pressure adjusted to height was elevated, if above the 95th percentile of European reference ranges,12 according to the guidelines of the German Hypertension League (Deutsche Hochdruckliga). Abnormal lipid levels (dyslipidemia) were defined according to the American Heart Association13 as follows: total cholesterol >5.18 mmol l−1, low-density lipoprotein (LDL) cholesterol >3.4 mmol l−1, high-density lipoprotein (HDL) cholesterol <0.9 mmol l−1 and triglycerides >1.71 mmol l−1. Impaired fasting glucose was defined as fasting glucose >6.1 mmol l−1.

Statistical analysis was performed using the SAS 9.1.3 software package (SAS Institute, Carey, NJ, USA). As not all variables were normally distributed, nonparametric statistical tests (Wilcoxon rank-sum test, Kruskal–Wallis test) were used to compare patient groups according to treatment strategies. To ensure strict control of overall type I error rate, P-values were adjusted for multiple comparisons as proposed by Holm.14 Treatment success was evaluated both on an intention-to-treat approach, defining success as any reduction of SDS-BMI, and in a per protocol analysis, based on the individual reduction of SDS-BMI. Success rates between the short- and long-treatment groups were compared by χ2-analysis. A multiple regression mixed model with SDS-BMI reduction as dependent variable and age, SDS-BMI at onset, gender, parental BMI, migration status and treatment strategy (the latter two as dichotomous variables) as fixed effects was used to elucidate patient covariates significantly related to treatment success. Variance due to center heterogeneity was accounted for by including treatment center (nested within treatment category) as a random categorical variable into the model, as only a subset of all treatment centers in Germany participated in the study. Intracluster correlation coefficients (ICCs) were calculated as the ratio of center variance (the random variable) as the numerator and total variance as the denominator.15 The model was implemented with SAS proc glimmix (version June 2006), with an identity-link function and Gaussian distribution, using residual subject-specific pseudolikelihood as estimation method and adjusting denominator degrees of freedom according to Kenward and Roger.16 Mean SDS-BMI reduction according to treatment strategy was calculated as least-square means based on observed margins, which provide model-based estimates. Center heterogeneity is visualized based on best linear unbiased predictors (BLUPs) for each individual center, incorporating both fixed and random effects from the mixed model. Significance was assigned to a P-value of <0.05.

Results

A total of 1916 children and adolescents participated in the study (Table 1). Of the 6, 5 treatment centers with treatment period <3 months (group A) offered an inpatient treatment program, whereas all the 42 treatment centers with treatment period 3 months (group B) offered an outpatient lifestyle intervention.

Table 1 Baseline characteristics (age, gender and degree of overweight as BMI and SDS-BMI) and frequency of diagnostic procedures in the study population and separated to treatment duration (group A: <3 months treatment duration; group B: 3 months treatment duration)

At baseline, the children did not differ significantly in respect of their age, gender, degree of overweight in the different treatment groups, despite the fact that the children in group A were older and more obese than the children in group B (see Table 1). Furthermore, the children in group B demonstrated significantly more cardiovascular risk factors as compared to children in group A (see Table 1). The frequency of diagnostic procedures according to the guidelines did not differ significantly between groups A and B at baseline. Of the screened overweight children, 51% demonstrated at least one cardiovascular risk factor (32% hypertension, 5% impaired fasting glucose, 1% diabetes mellitus, 31% dyslipidemia).

At the end of intervention, the majority of the children reduced their degree of overweight in both treatment groups (Table 2). On the basis of an intention-to-treat analysis, 75% of the children reduced their overweight (group A: 89% success, group B: 64% success, success defined as SDS-BMI reduction). The dropout rate was significantly (P<0.001) lower and the success rate was significantly (P<0.001) higher in treatment group A.

Table 2 Outcome of the patients at the end of intervention in the study population (n=1916) and separated to treatment duration (group A, n=875: treatment duration <3 months), group B, n=1041: treatment duration 3 months)

A mixed multiple linear model was used to elucidate significant confounders on SDS-BMI reduction and adjust for center heterogeneity in a hierarchical analysis simultaneously. Adjusted means for SDS-BMI reduction were −0.36 for group A and −0.17 for group B (P<0.0001). As shown in Table 3, younger age and lower SDS-BMI at onset were significantly associated with better SDS-BMI reduction, whereas gender, migration status and BMI of father did not contribute significantly. Furthermore, BMI of mother was significantly associated to SDS-BMI reduction of their children, but only in the model including only this one fixed effect in addition to treatment group.

Table 3 Factors related to the degree of SDS-BMI reduction in 1916 children participating in a lifestyle intervention Coefficients are based on a hierarchical mixed linear model with treatment center as random effect (level 2). Coefficients for models including only one fixed effect in addition to treatment group are compared to coefficients from the multiple models including all fixed effects simultaneously.

At the level of individual treatment centers within the two treatment categories, considerable heterogeneity was observed, which cannot be explained by differences in baseline characteristics of the patients. In the multiple regression mixed model with SDS-BMI reduction as dependent variable and age, SDS-BMI at onset, gender, parental BMI, migration status and treatment strategy as fixed effects, ICC was 0.15, indicative of a predominant patient effect, however also a relevant center effect. In Figure 1, adjusted mean SDS-BMI reductions, based on BLUPs, are shown. Center heterogeneity is clearly larger in group B (range −0.01 to −0.36) compared to group A (−0.27 to −0.42).

Figure 1
figure1

Adjusted mean s.d. score of body mass index (SDS-BMI) reductions during intervention in the individual 48 treatment centers based on best linear unbiased predictors (BLUPs) incorporating both fixed and random effects from the mixed model (line: mean of groups (short therapy=treatment duration <3 months, group A; long therapy=treatment duration3 months, group B)).

At the end of interventions, screening of cardiovascular risk factors was performed only in 194 (10%) children (group A: 2%, group B: 17%, P< 0.0001 group A compared to group B). In these screened children, most cardiovascular risk factors improved significantly (Table 4).

Table 4 Cardiovascular risk factors at baseline and the end of intervention in the 194 children with measurement of cardiovascular risk factors at baseline and the end of intervention

Discussion

This observational study demonstrated the high frequency of cardiovascular risk factors in overweight children and the heterogeneity of clinical care for overweight and obese children and adolescents in Germany in real-life condition. Our study demonstrated that even in real-life conditions overweight reduction in children is achievable with lifestyle interventions. Because cardiovascular risk factors improved in the children who were screened at baseline and follow-up, the effect of lifestyle intervention seems clinically relevant. Because younger age and lower degree of overweight were associated with better outcome, these findings underline the necessity of an early intervention in overweight children.

Many overweight children in Germany were treated in inpatient rehabilitation centers for weeks (group A) in contrast to the recommended long-term intervention. As the inpatient treatment centers with short-term intervention rendered the best outcome in this study, one could question the recommendation of an intensive multidisciplinary long-term intervention over months. The better outcome in group A might be explained by the fact that outpatient long-term interventions probably have more difficulties to achieve overweight reduction in real-life conditions than short-term interventions in inpatient treatment centers under fixed conditions with a defined amount of exercise treatment and a defined caloric intake. Most importantly, weight loss may not be sustained and this ongoing BZgA study will demonstrate which treatment approach has the best outcome 1 and 2 years after the end of intervention. So far, long-term success in overweight reduction has been demonstrated only after long-term outpatient lifestyle intervention programs but not after short-term inpatient treatment approaches.5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17

One of the most striking findings in our observational study was the great heterogeneity in the clientele, treatment approach and outcome. Especially older and more obese children were treated <3 months, whereas children with 3 months treatment duration demonstrated more cardiovascular risk factors at baseline. After adjusting for baseline age, gender, degree of overweight, migration status and parental BMI, especially the treatment centers in group B (3 months treatment duration) differed widely in the achieved weight reduction of their patients (see Figure 1). Furthermore, treatment centers in group A (<3 months treatment duration) took care of much more patients per institution as compared to treatment centers of group B. These discrepancies may be explained by the shorter treatment duration in group A but also by structural and financial differences in inpatient and outpatient treatment centers in Germany.18 For example, inpatient treatment centers are financed by annuity insurances in contrast to outpatient treatment centers financed by health insurances.

To improve the outcome after lifestyle interventions under real-life conditions, it is necessary to analyze why treatment institutions, especially in group B, differed so widely in their outcome. Probably benchmarking programs like APV can help to harmonize and optimize treatment modalities. Ideally, all treatment centers in Germany should participate in this quality process program. Accreditation of institutions as started by the German Working Group of Obesity in Childhood and Adolescence (AGA) may help achieve this.19

Diagnostic procedures as recommended in guidelines were not performed in nearly half of the children in specialized treatment centers. Especially, fasting glucose measurements were lacking. However, impaired fasting glucose and diabetes mellitus have a strong impact on the long-term morbidity and mortality of obesity.3 The finding of lacking diagnostic procedures concerning the cardiovascular risk factor profile is in concordance with a previous study in Germany in 2001.4 Most importantly, cardiovascular risk factors were only seldomly determined in follow-up even in the children in which cardiovascular risk factors have been detected at baseline. Because the improvement of cardiovascular risk factors cannot be predicted by weight loss,20 diagnostic procedures have to be performed in follow-up, as other treatment modalities (for example, drug therapy in diabetes, hypertension and dyslipidemia) may be necessary if comorbidities persist or worsen.

Because the national and international guidelines request measurement of blood pressure, lipids and glucose in all overweight and obese children,1, 2, 3 the treatment center should achieve a high frequency of complete diagnostic procedures at baseline and follow-up. Thresholds for diagnostic procedures are planned for certification of obesity treatment institutions by the German Working Group of Obesity in Childhood and Adolescence according to other guidelines.19 For example, the Center for Disease Control and Prevention has defined 80% as threshold for diagnostic procedures defining good quality in management of patients with diabetes mellitus.21

The strengths of this study are the recruitment of multiple treatment centers in real-life conditions and the large study cohort. Furthermore, studying social status, psychosocial factors and change of quality of life will help to identify predictive factors and to prove the relevance of overweight reduction in further analyses. However, this study has some potential important limitations. First, the children were not randomized to the different treatment approaches. Therefore, we have to be very cautious when comparing the different treatment groups because the outcomes are likely influenced by further factors not balanced in this study, for example, motivation. Moreover, the age and degree of overweight were different in the treatment groups, probably also influencing the findings. In addition, this study has no control group. Conversely, randomization and implementation of a control group generate a study population that is probably quite different from the population presenting for obesity treatment in real-life. Finally, the long-term effect after the end of intervention has to be studied to compare the different treatment approaches. Currently, the outcome of the treatment centers in our study is being studied 1 and 2 years after the end of intervention.

In summary, overweight reduction in children and improvement of their cardiovascular risk factors are achievable with lifestyle interventions in clinical practice under everyday circumstances. However, the clientele, practices of the institutions and outcomes varied widely between different institutions. Nearly half of the overweight children demonstrated cardiovascular risk factors but only a minority of the children was screened for them in follow-up even in specialized treatment centers. In conclusion, implementation of quality criteria for institutions seems necessary to improve medical care of overweight children in real life.

References

  1. 1

    Kunze D, Wabitsch M . Guidelines of the German Obesity Working Group. available at www.a-g-a.de.

  2. 2

    Wabitsch M . Overweight and obesity in European children: definition and diagnostic procedures, risk factors and consequences for later health outcome. Eur J Pediatr 2000; 159 (Suppl 1): S8–13.

    Article  Google Scholar 

  3. 3

    Barlow SE . Expert Committee. Expert committee recommendations regarding the prevention, assessment, and treatment of child and adolescent overweight and obesity: summary report. Pediatrics 2007; 120 (Suppl 1): S164–S192.

    Article  Google Scholar 

  4. 4

    Reinehr T, Wabitsch M, Andler W, Beyer P, Böttner A, Chen-Stute A et al. Medical care of obese children and adolescents—APV: a standardised multicentre documentation derived to study initial presentation and cardiovascular risk factors in patients transferred to specialised treatment institutions. Eur J Ped 2004; 163: 308–312.

    Article  Google Scholar 

  5. 5

    Campbell K, Waters E, ÓMeara S, Summerbell C . Interventions for preventing obesity in childhood. A systematic review. Obes Rev 2001; 2: 149–157.

    CAS  Article  Google Scholar 

  6. 6

    Reinehr T, Wabitsch M . Strukturierte Erfassung der Therapieangebote für adipöse Kinder und Jugendliche—ein Projekt der Arbeitsgemeinschaft Adipositas im Kindes- und Jugendalter (AGA). Monatsschr Kinderheilkd 2003; 151: 757–761.

    Article  Google Scholar 

  7. 7

    BzgA . Die Versorgung übergewichtiger und adipöser Kinder und Jugendlicher in Deutschland. Quantität und Qualität von Angeboten im Zeitraum 2004–2005. Gesundheitsförderung konkret 2007. Band 8.

  8. 8

    L'Allemand D, Wiegand S, Müller J, Reinehr T, Wabitsch M, Widhalm K et al. Cardiovascular risk in 26008 European overweight children as established by a multicenter database. Obesity 2008; 16: 1672–1679.

    Article  Google Scholar 

  9. 9

    Cole TJ, Bellizzi MC, Flegal KM, Dietz WH . Establishing a standard definition for child overweight and obesity worldwide: international survey. BMJ 2000; 320: 1–6.

    Article  Google Scholar 

  10. 10

    Kromeyer-Hauschild K, Wabitsch M, Geller F, Ziegler A, Geiβ HC, Hesse V et al. Percentiles of body mass index in children and adolescents evaluated from different regional German studies. Monatsschr Kinderheilkd 2001; 149: 807–818.

    Article  Google Scholar 

  11. 11

    Cole TJ . The LMS method for constructing normalized growth standards. Eur J Clin Nutr 1990; 44: 45–60.

    CAS  PubMed  Google Scholar 

  12. 12

    de Man SA, Andre JL, Bachmann H, Grobbee DE, Ibsen KK, Laaser U et al. Blood pressure in childhood: pooled findings of six European studies. J Hypertens 1991; 9: 109–114.

    CAS  Article  Google Scholar 

  13. 13

    Kavey REW, Daniels SR, Lauer RM, Atins DL, Hayman LL, Taubert K . American Heart Association Guidelines for Primary Prevention of Atherosclerotic Cardiovascular Disease Beginning in Childhood. Circulation 2003; 107: 1562–1566.

    Article  Google Scholar 

  14. 14

    Dmitrienko A, Molenberghs G, Chuiang-Stein C, Offen W . Analysis of Clinical Trials using SAS Multiple Comparisons and Multiple Endpoints. SAS Publishing: Carey, NJ, USA, 2005. pp 67–128.

  15. 15

    Shoukri MM, Chaudhary MA . Analysis of correlated data with SAS and R. 3rd edn 2007. Chapman and Hall/CRC, Boca Raton: London, New York.

  16. 16

    Littell RC, Milliken GW, Stroup WW, Wolfinger RD, Schabenberger O . SAS for Mixed Models. SAS Press: Carey, NJ, USA, 2006.

    Google Scholar 

  17. 17

    Reinehr T, Temmesfeld M, Kersting M, de Sousa G, Toschke AM . Four-year follow-up of children and adolescents participating in an obesity intervention program. Int J Obes 2007; 31: 1074–1077.

    CAS  Article  Google Scholar 

  18. 18

    Sawicki PT, Bastian H . German health care: a bit of Bismarck plus more science. BMJ 2008; 337: a1997.

    Article  Google Scholar 

  19. 19

    Reinehr T, Holl R, Wabitsch M . The German Working Group of Childhood and Adolescent Obesity (AGA): improving the quality of care for overweight and obese children in Germany. Obes Facts 2008; 1: 26–32.

    Article  Google Scholar 

  20. 20

    Toschke AM Reinehr T . Different anthropometric index changes in relation to cardiovascular risk profile change. Clin Nutr 2008; 27: 457–463.

    Article  Google Scholar 

  21. 21

    Centers for Disease Control and Prevention. National diabetes fact sheet: general information and national estimates on diabetes in the United States 2002. Department of Health and Human Services, Centers for Disease Control and Diabetes: Atlanta, GA, US, 2003.

Download references

Acknowledgements

We thank all children and parents participating at this study as well as all treatment centers for this kind support. We are indebted to Gideon de Sousa, Vestische Youth Hospital Datteln, for his support in editing the paper.

This study was funded by the Federal Centre for Health Education (Cologne, Germany), an agency of the German Federal Ministry for Health.

Author information

Affiliations

Authors

Corresponding author

Correspondence to T Reinehr.

Rights and permissions

Reprints and Permissions

About this article

Cite this article

Reinehr, T., Hoffmeister, U., Mann, R. et al. Medical care of overweight children under real-life conditions: the German BZgA observation study. Int J Obes 33, 418–423 (2009). https://doi.org/10.1038/ijo.2009.50

Download citation

Keywords

  • quality of care
  • children
  • overweight
  • intervention
  • diagnostic procedures

Further reading

Search