obesity reviews
doi: 10.1111/obr.12179
Obesity Prevention
Establishing a method to estimate the cost-effectiveness of a kindergarten-based, family-involved intervention to prevent obesity in early childhood. The ToyBox-study L. Pil1, K. Putman2, G. Cardon3, I. De Bourdeaudhuij3, Y. Manios4, O. Androutsos4, M. Lateva5, V. Iotova5, K. Zych6, M. Góz´dz´6, E. M. González-Gil7, P. De Miguel-Etayo7, C. Geyer8, J. Birnbaum8 and L. Annemans1,2 on behalf of the ToyBox-study group 1
Department of Public Health, Ghent University,
Ghent, Belgium; 2Faculty of Medicine and Pharmacy, Vrije Universiteit Brussel, Brussels, Belgium; 3Department
of Movement and Sports Sciences,
Ghent University, Ghent, Belgium; 4Department of Nutrition and Dietetics, Harokopio University, Athens, Greece; 5Clinic of Paediatric Endocrinology, Medical University of Varna, Varna, Bulgaria; 6Public Health Division, The Children’s Memorial Health Institute, Warsaw, Poland; 7GENUD (Growth, Exercise, Nutrition and Development) Research Group, University of Zaragoza, Zaragoza, Spain; 8Division of Metabolic and Nutritional Medicine Dr. von Hauner Children’s Hospital Ludwig-Maximilians-University of Munich, München, Germany
Received 3 April 2014; accepted 8 April 2014
Address for correspondence: L Pil, Department of
Summary Overweight and obesity in children are recognized as a major health problem. The ToyBoxintervention was developed with the aim of preventing obesity in pre-schoolers. Because it is increasingly important to inform policy makers not only on the effects of prevention interventions, but also on their costs and cost-effectiveness, our purpose was to establish a method to estimate the cost-effectiveness of the ToyBox-intervention. In order to estimate the long-term impact of the ToyBox-intervention on health and societal costs, extrapolations of the intervention effect will be conducted to predict children’s weight status (based on the body mass index) at adult age. Effects of the adult weight status on the prevalence of obesity-related complications will be modelled through a Markov model, with a total time horizon of 70 years and a cycle length of 1 year. The model will be conducted in six European countries participating in the ToyBox-intervention, based on country-specific economic and epidemiological data. This study describes the methodological rationale and implementation of an analytic model to examine the cost-effectiveness of the ToyBoxintervention for six European countries, in order to inform decision-makers on the value for money of this intervention in the prevention of obesity in pre-schoolers.
Public Health, Ghent University, De Pintelaan 185, 9000 Gent, Belgium.
Keywords: Cost-effectiveness, Markov model, obesity prevention, ToyBox-intervention.
E-mail: [email protected]
obesity reviews (2014) 15 (Suppl. 3), 81–89
ToyBox-study group: Coordinator: Yannis Manios; Steering committee: Yannis Manios, Berthold Koletzko, Ilse De Bourdeaudhuij, Mai Chin A Paw, Luis Moreno, Carolyn Summerbell, Tim Lobstein, Lieven Annemans, Goof Buijs; External advisors: John Reilly, Boyd Swinburn, Dianne Ward; Harokopio University (Greece): Yannis Manios, Odysseas Androutsos, Eva Grammatikaki, Christina Katsarou, Eftychia Apostolidou, Eirini Efstathopoulou; Ludwig Maximilians Universitaet Muenchen (Germany): Berthold Koletzko, Kristin Duvinage, Sabine Ibrügger, Angelika Strauß, Birgit Herbert, Julia Birnbaum, Annette Payr, Christine Geyer; Ghent University (Belgium): Department of Movement and Sports Sciences: Ilse De Bourdeaudhuij, Greet Cardon, Marieke De Craemer, Ellen De Decker; Department of Public Health: Lieven Annemans, Stefaan De Henauw, Lea Maes, Carine Vereecken, Jo Van Assche, Lore Pil; VU University Medical Center EMGO Institute for Health and Care Research (the Netherlands): EMGO Institute for Health and Care Research: Mai Chin A Paw, Saskia te Velde; University of Zaragoza (Spain): Luis Moreno, Theodora Mouratidou, Juan Fernandez, Maribel Mesana, Pilar De Miguel-Etayo, Esther M. González-Gil, Luis Gracia-Marco, Beatriz Oves; Oslo and Akershus University College of Applied Sciences (Norway): Agneta Yngve, Susanna Kugelberg, Christel Lynch, Annhild Mosdøl, Bente B Nilsen; University of Durham (UK): Carolyn Summerbell, Helen Moore, Wayne Douthwaite, Catherine Nixon; State Institute of Early Childhood Research (Germany): Susanne Kreichauf, Andreas Wildgruber; Children’s Memorial Health Institute (Poland): Piotr Socha, Zbigniew Kulaga, Kamila Zych, Magdalena Góz´dz´, Beata Gurzkowska, Katarzyna Szott; Medical University of Varna (Bulgaria): Violeta Iotova, Mina Lateva, Natalya Usheva, Sonya Galcheva, Vanya Marinova, Zhaneta Radkova, Nevyana Feschieva; International Association for the Study of Obesity (UK): Tim Lobstein, Andrea Aikenhead; CBO B.V. (the Netherlands): Goof Buijs, Annemiek Dorgelo, Aviva Nethe, Jan Jansen; AOK-Verlag (Germany): Otto Gmeiner, Jutta Retterath, Julia Wildeis, Axel Günthersberger; Roehampton University (UK): Leigh Gibson; University of Luxembourg (Luxembourg): Claus Voegele.
© 2014 World Obesity
81 15 (Suppl. 3), 81–89, August 2014
obesity reviews
82 Cost-effectiveness of obesity prevention L. Pil et al.
Figure 1 Model structure: extrapolation of the two possible intervention effects to the weight status at adult age (25,27–30) and the long-term effect on the prevalence of the chronic diseases. Narrow arrows: intervention effect; broad arrows: modelling, the ToyBox-study. EBRB, energy balance-related behaviour.
Introduction The prevalence of overweight and obesity in pre-school children has substantially increased worldwide (1). Nonactive lifestyles and non-healthy eating patterns have an important impact on this trend (2,3). Obese children have a higher risk to be obese at adult age (4,5). Additionally, obesity in children is associated with a higher risk for later chronic diseases such as type 2 diabetes and heart disease (3,6). The trend of increasing prevalence of overweight and obese children and adults results in a rising societal impact due to higher healthcare costs and productivity loss (7,8). The ToyBox-study (short for ‘Multifactorial evidence based approach using behavioural models in understanding and promoting fun, healthy food, play and policy for the prevention of obesity in early childhood’) aimed to develop, implement and evaluate a kindergarten-based, familyinvolved intervention to prevent obesity in early childhood (9). Previous studies have demonstrated the importance of calculating the cost-effectiveness of obesity prevention programmes (10–13). Wang et al. (11) stated that more costeffectiveness studies of healthcare programmes are needed in order to help programme planners to better understand the cost-effectiveness of similar or different healthcare programmes and to assess whether they can afford a particular programme. Hence, reporting the results of a costeffectiveness analysis allows decision-makers to make direct comparisons among different healthcare interventions (13). The current paper describes the design and data inputs of the health economic model used to estimate the long-term costs and effects for society of implementing the ToyBoxintervention, in order to evaluate the potential impact in the six intervention countries, namely, Belgium, Bulgaria, Germany, Greece, Poland and Spain.
Methods The ToyBox-intervention was implemented throughout the academic year 2012–2013 (24 weeks) in Belgium, Bulgaria, Germany, Greece, Poland and Spain (9,14,15) (www.toybox-study.eu). Alongside the intervention, a 15 (Suppl. 3), 81–89, August 2014
health economic evaluation will be conducted, estimating the long-term cost benefits of the ToyBox-intervention. The health economic model developed for ToyBox is a combined model consisting of a decision analytic model to represent either the probability of improved energy balance-related behaviours (EBRBs) or improved weight status (based on body mass index [BMI]) and a Markov model simulating over a lifetime the occurrence of obesityrelated complications with and without the intervention. The target population of the model consists of European pre-schoolers, between 4 and 6 years old. The difference in costs over a period of 70 years will be divided by the net effects (in quality-adjusted life years, QALYs) to obtain the primary outcome measure, the incremental costeffectiveness ratio: (CostsIntervention group – CostsControl group)/ (QALYsIntervention group – QALYsControl group). QALYs are calculated by multiplying the utility level for a given condition (a health-related quality-of-life weight ranging between 0 and 1) with the numbers of years an individual lives with the particular condition. A utility of 1 is equal to perfect health, whereas 0 stands for death.
Model structure The ToyBox health economic model consists of different parts, whereby the structure of the first part accounts for the type of effect of the intervention. There are two effect scenarios based on the possible anticipated consequences of the intervention: an effect either on the key EBRBs of the pre-schoolers targeted in the ToyBox-intervention (snacking behaviour, drinking behaviour, physical activity, sedentary behaviour), on the weight status of the pre-schoolers (defined as weight categories normal weight, overweight and obesity1) or on both (Fig. 1). As chronic diseases start to develop at adult age, a long-term extrapolation of the effect of the ToyBox-intervention on the pre-schoolers’ EBRBs or on the pre-schoolers’ weight status to the adult age (30–34 year) is necessary. The effect scenario and the long-term extrapolation make up the first part of the 1
According to the cut-off values developed by Cole et al. (37) and
adopted by the International Obesity Task Force (IOTF).
© 2014 World Obesity
obesity reviews
model. The second part of the model is a Markov model projecting the incidences of five chronic diseases associated with obesity (type 2 diabetes, stroke, coronary heart disease [CHD], breast cancer, colorectal cancer) and mortality from the age category 30–34 years onwards. For the first part of the model, we use tracking studies that estimate the relationship between EBRB or weight status at pre-school age, on the one hand, and the weight status at adult age, on the other hand2. If weight status and/or healthy EBRBs in childhood improve as a consequence of the ToyBox-intervention, this will result in a shift in the prevalence of overweight or obesity at adult age. The relationship between weight status at adult age and obesityrelated complications is obtained from the International Association for the Study of Obesity (16). Hence, the impact on those obesity-related complications (and associated costs) at older age with the early childhood intervention can be calculated indirectly. The Markov model is designed as a stochastic probabilistic model in which all events occur with specific probabilities, accounting for the uncertainty surrounding those probabilities (17). Consequently, the stochastic characteristics of the data that are used to populate the model allow uncertainty around the outcomes and therefore make the results more informative for decision-makers.
Extrapolation to adult weight status First scenario: effect on weight status of the pre-schoolers A first approach is modelling the (possible) intervention effect on the weight status of the children (i.e. change in the proportions of pre-schoolers who are normal weight, overweight or obese), whereby we make use of the tracking study of Venn et al. (4) to estimate the proportions of pre-schoolers that will be normal weight, overweight or obese at 30–34 years based on their pre-school weight post-intervention. In this way the (possible) intervention effect on weight status can be extrapolated from preschooler age to adult age. Second scenario: effect on pre-schoolers’ energy balance-related behaviours A second approach is the modelling based on change of EBRBs. If the ToyBox-intervention would have an effect on the EBRBs of the children that are targeted in the ToyBoxintervention, two methods could be used to extrapolate the effect to the adult weight status (Fig. 1). The effect on childhood behaviours can be extrapolated directly to the weight status at adult age, based on published relative risks, or it can be extrapolated indirectly to the weight status at 2
According to the WHO norms: normal weight:
doi: 10.1111/obr.12179
Obesity Prevention
Establishing a method to estimate the cost-effectiveness of a kindergarten-based, family-involved intervention to prevent obesity in early childhood. The ToyBox-study L. Pil1, K. Putman2, G. Cardon3, I. De Bourdeaudhuij3, Y. Manios4, O. Androutsos4, M. Lateva5, V. Iotova5, K. Zych6, M. Góz´dz´6, E. M. González-Gil7, P. De Miguel-Etayo7, C. Geyer8, J. Birnbaum8 and L. Annemans1,2 on behalf of the ToyBox-study group 1
Department of Public Health, Ghent University,
Ghent, Belgium; 2Faculty of Medicine and Pharmacy, Vrije Universiteit Brussel, Brussels, Belgium; 3Department
of Movement and Sports Sciences,
Ghent University, Ghent, Belgium; 4Department of Nutrition and Dietetics, Harokopio University, Athens, Greece; 5Clinic of Paediatric Endocrinology, Medical University of Varna, Varna, Bulgaria; 6Public Health Division, The Children’s Memorial Health Institute, Warsaw, Poland; 7GENUD (Growth, Exercise, Nutrition and Development) Research Group, University of Zaragoza, Zaragoza, Spain; 8Division of Metabolic and Nutritional Medicine Dr. von Hauner Children’s Hospital Ludwig-Maximilians-University of Munich, München, Germany
Received 3 April 2014; accepted 8 April 2014
Address for correspondence: L Pil, Department of
Summary Overweight and obesity in children are recognized as a major health problem. The ToyBoxintervention was developed with the aim of preventing obesity in pre-schoolers. Because it is increasingly important to inform policy makers not only on the effects of prevention interventions, but also on their costs and cost-effectiveness, our purpose was to establish a method to estimate the cost-effectiveness of the ToyBox-intervention. In order to estimate the long-term impact of the ToyBox-intervention on health and societal costs, extrapolations of the intervention effect will be conducted to predict children’s weight status (based on the body mass index) at adult age. Effects of the adult weight status on the prevalence of obesity-related complications will be modelled through a Markov model, with a total time horizon of 70 years and a cycle length of 1 year. The model will be conducted in six European countries participating in the ToyBox-intervention, based on country-specific economic and epidemiological data. This study describes the methodological rationale and implementation of an analytic model to examine the cost-effectiveness of the ToyBoxintervention for six European countries, in order to inform decision-makers on the value for money of this intervention in the prevention of obesity in pre-schoolers.
Public Health, Ghent University, De Pintelaan 185, 9000 Gent, Belgium.
Keywords: Cost-effectiveness, Markov model, obesity prevention, ToyBox-intervention.
E-mail: [email protected]
obesity reviews (2014) 15 (Suppl. 3), 81–89
ToyBox-study group: Coordinator: Yannis Manios; Steering committee: Yannis Manios, Berthold Koletzko, Ilse De Bourdeaudhuij, Mai Chin A Paw, Luis Moreno, Carolyn Summerbell, Tim Lobstein, Lieven Annemans, Goof Buijs; External advisors: John Reilly, Boyd Swinburn, Dianne Ward; Harokopio University (Greece): Yannis Manios, Odysseas Androutsos, Eva Grammatikaki, Christina Katsarou, Eftychia Apostolidou, Eirini Efstathopoulou; Ludwig Maximilians Universitaet Muenchen (Germany): Berthold Koletzko, Kristin Duvinage, Sabine Ibrügger, Angelika Strauß, Birgit Herbert, Julia Birnbaum, Annette Payr, Christine Geyer; Ghent University (Belgium): Department of Movement and Sports Sciences: Ilse De Bourdeaudhuij, Greet Cardon, Marieke De Craemer, Ellen De Decker; Department of Public Health: Lieven Annemans, Stefaan De Henauw, Lea Maes, Carine Vereecken, Jo Van Assche, Lore Pil; VU University Medical Center EMGO Institute for Health and Care Research (the Netherlands): EMGO Institute for Health and Care Research: Mai Chin A Paw, Saskia te Velde; University of Zaragoza (Spain): Luis Moreno, Theodora Mouratidou, Juan Fernandez, Maribel Mesana, Pilar De Miguel-Etayo, Esther M. González-Gil, Luis Gracia-Marco, Beatriz Oves; Oslo and Akershus University College of Applied Sciences (Norway): Agneta Yngve, Susanna Kugelberg, Christel Lynch, Annhild Mosdøl, Bente B Nilsen; University of Durham (UK): Carolyn Summerbell, Helen Moore, Wayne Douthwaite, Catherine Nixon; State Institute of Early Childhood Research (Germany): Susanne Kreichauf, Andreas Wildgruber; Children’s Memorial Health Institute (Poland): Piotr Socha, Zbigniew Kulaga, Kamila Zych, Magdalena Góz´dz´, Beata Gurzkowska, Katarzyna Szott; Medical University of Varna (Bulgaria): Violeta Iotova, Mina Lateva, Natalya Usheva, Sonya Galcheva, Vanya Marinova, Zhaneta Radkova, Nevyana Feschieva; International Association for the Study of Obesity (UK): Tim Lobstein, Andrea Aikenhead; CBO B.V. (the Netherlands): Goof Buijs, Annemiek Dorgelo, Aviva Nethe, Jan Jansen; AOK-Verlag (Germany): Otto Gmeiner, Jutta Retterath, Julia Wildeis, Axel Günthersberger; Roehampton University (UK): Leigh Gibson; University of Luxembourg (Luxembourg): Claus Voegele.
© 2014 World Obesity
81 15 (Suppl. 3), 81–89, August 2014
obesity reviews
82 Cost-effectiveness of obesity prevention L. Pil et al.
Figure 1 Model structure: extrapolation of the two possible intervention effects to the weight status at adult age (25,27–30) and the long-term effect on the prevalence of the chronic diseases. Narrow arrows: intervention effect; broad arrows: modelling, the ToyBox-study. EBRB, energy balance-related behaviour.
Introduction The prevalence of overweight and obesity in pre-school children has substantially increased worldwide (1). Nonactive lifestyles and non-healthy eating patterns have an important impact on this trend (2,3). Obese children have a higher risk to be obese at adult age (4,5). Additionally, obesity in children is associated with a higher risk for later chronic diseases such as type 2 diabetes and heart disease (3,6). The trend of increasing prevalence of overweight and obese children and adults results in a rising societal impact due to higher healthcare costs and productivity loss (7,8). The ToyBox-study (short for ‘Multifactorial evidence based approach using behavioural models in understanding and promoting fun, healthy food, play and policy for the prevention of obesity in early childhood’) aimed to develop, implement and evaluate a kindergarten-based, familyinvolved intervention to prevent obesity in early childhood (9). Previous studies have demonstrated the importance of calculating the cost-effectiveness of obesity prevention programmes (10–13). Wang et al. (11) stated that more costeffectiveness studies of healthcare programmes are needed in order to help programme planners to better understand the cost-effectiveness of similar or different healthcare programmes and to assess whether they can afford a particular programme. Hence, reporting the results of a costeffectiveness analysis allows decision-makers to make direct comparisons among different healthcare interventions (13). The current paper describes the design and data inputs of the health economic model used to estimate the long-term costs and effects for society of implementing the ToyBoxintervention, in order to evaluate the potential impact in the six intervention countries, namely, Belgium, Bulgaria, Germany, Greece, Poland and Spain.
Methods The ToyBox-intervention was implemented throughout the academic year 2012–2013 (24 weeks) in Belgium, Bulgaria, Germany, Greece, Poland and Spain (9,14,15) (www.toybox-study.eu). Alongside the intervention, a 15 (Suppl. 3), 81–89, August 2014
health economic evaluation will be conducted, estimating the long-term cost benefits of the ToyBox-intervention. The health economic model developed for ToyBox is a combined model consisting of a decision analytic model to represent either the probability of improved energy balance-related behaviours (EBRBs) or improved weight status (based on body mass index [BMI]) and a Markov model simulating over a lifetime the occurrence of obesityrelated complications with and without the intervention. The target population of the model consists of European pre-schoolers, between 4 and 6 years old. The difference in costs over a period of 70 years will be divided by the net effects (in quality-adjusted life years, QALYs) to obtain the primary outcome measure, the incremental costeffectiveness ratio: (CostsIntervention group – CostsControl group)/ (QALYsIntervention group – QALYsControl group). QALYs are calculated by multiplying the utility level for a given condition (a health-related quality-of-life weight ranging between 0 and 1) with the numbers of years an individual lives with the particular condition. A utility of 1 is equal to perfect health, whereas 0 stands for death.
Model structure The ToyBox health economic model consists of different parts, whereby the structure of the first part accounts for the type of effect of the intervention. There are two effect scenarios based on the possible anticipated consequences of the intervention: an effect either on the key EBRBs of the pre-schoolers targeted in the ToyBox-intervention (snacking behaviour, drinking behaviour, physical activity, sedentary behaviour), on the weight status of the pre-schoolers (defined as weight categories normal weight, overweight and obesity1) or on both (Fig. 1). As chronic diseases start to develop at adult age, a long-term extrapolation of the effect of the ToyBox-intervention on the pre-schoolers’ EBRBs or on the pre-schoolers’ weight status to the adult age (30–34 year) is necessary. The effect scenario and the long-term extrapolation make up the first part of the 1
According to the cut-off values developed by Cole et al. (37) and
adopted by the International Obesity Task Force (IOTF).
© 2014 World Obesity
obesity reviews
model. The second part of the model is a Markov model projecting the incidences of five chronic diseases associated with obesity (type 2 diabetes, stroke, coronary heart disease [CHD], breast cancer, colorectal cancer) and mortality from the age category 30–34 years onwards. For the first part of the model, we use tracking studies that estimate the relationship between EBRB or weight status at pre-school age, on the one hand, and the weight status at adult age, on the other hand2. If weight status and/or healthy EBRBs in childhood improve as a consequence of the ToyBox-intervention, this will result in a shift in the prevalence of overweight or obesity at adult age. The relationship between weight status at adult age and obesityrelated complications is obtained from the International Association for the Study of Obesity (16). Hence, the impact on those obesity-related complications (and associated costs) at older age with the early childhood intervention can be calculated indirectly. The Markov model is designed as a stochastic probabilistic model in which all events occur with specific probabilities, accounting for the uncertainty surrounding those probabilities (17). Consequently, the stochastic characteristics of the data that are used to populate the model allow uncertainty around the outcomes and therefore make the results more informative for decision-makers.
Extrapolation to adult weight status First scenario: effect on weight status of the pre-schoolers A first approach is modelling the (possible) intervention effect on the weight status of the children (i.e. change in the proportions of pre-schoolers who are normal weight, overweight or obese), whereby we make use of the tracking study of Venn et al. (4) to estimate the proportions of pre-schoolers that will be normal weight, overweight or obese at 30–34 years based on their pre-school weight post-intervention. In this way the (possible) intervention effect on weight status can be extrapolated from preschooler age to adult age. Second scenario: effect on pre-schoolers’ energy balance-related behaviours A second approach is the modelling based on change of EBRBs. If the ToyBox-intervention would have an effect on the EBRBs of the children that are targeted in the ToyBoxintervention, two methods could be used to extrapolate the effect to the adult weight status (Fig. 1). The effect on childhood behaviours can be extrapolated directly to the weight status at adult age, based on published relative risks, or it can be extrapolated indirectly to the weight status at 2
According to the WHO norms: normal weight:
