SEARCH

SEARCH BY CITATION

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods and Procedures
  5. Results
  6. Discussion
  7. Acknowledgments
  8. References

Objective

To examine the cost-effectiveness of Be Active Eat Well (BAEW), a large, multifaceted, community-based capacity-building demonstration program that promoted healthy eating and physical activity for Australian children aged 4-12 years between 2003 and 2006.

Design and Methods

A quasi-experimental, longitudinal design was used with anthropometric data collected at baseline (1001 children—intervention; 1183—comparator) and follow-up. A societal perspective was employed, with intervention resource use measured retrospectively based on process evaluation reports, school newsletters, reports, and key stakeholder interviews, and valued in 2006 Australian dollars (AUD). Outcomes were measured as Body Mass Index (BMI) units saved and Disability Adjusted Life Years (DALYs) averted over the predicted cohort lifetime, and reported as incremental cost-effectiveness ratios (with 95% uncertainty intervals).

Results

The intervention cost AUD0.34M ($0.31M; $0.38M) annually, and resulted in savings of 547 (−104; 1209) BMI units and 10.2 (−0.19; 21.6) DALYs. This translated to modest cost offsets of AUD27 311 (−$1803; $58 242) and a net cost per DALY saved of AUD29 798 (dominated; $0.26M).

Conclusions

BAEW was affordable and cost-effective, and generated substantial spin-offs in terms of activity beyond funding levels. Elements fundamental to its success and any potential cost efficiencies associated with scaling-up now require identification.


Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods and Procedures
  5. Results
  6. Discussion
  7. Acknowledgments
  8. References

The prevalence of childhood obesity worldwide has increased markedly in recent years [1]. In Australian children, overweight and obesity levels have doubled between 1985 and 1997 [2, 3]. In a recent review of 41 Australian studies of childhood weight status, current prevalence rates of Australian children (aged 2-18 years) are reported as 21-25% for overweight and obese combined and 5-6% for obese alone [4]. Whilst the study concluded that prevalence rates have started to plateau in the last 10 years, they remain high and warrant urgent action.

On a global level, concern about the high prevalence of overweight and obesity amongst children has given rise to a considerable number of multifaceted childhood obesity prevention programs [5-14]. In particular, the reviews by Foltz et al. [13] and Waters et al. [14] highlighted a range of population level strategies, which are often multifaceted, multilevel, and sometimes multisetting. The Cochrane Review [14] found strong evidence to support beneficial effects of child obesity prevention programs on BMI, particularly amongst primary school aged children, and that such programs included a broad range of components. In a review of 51 school-based interventions, Shaya et al. found 40 studies achieved positive statistically significant results between baseline and follow-up, although there was generally no persistence of positive results [15]. However, few childhood obesity programs have included an economic evaluation [16]. The key question answered by an economic evaluation is whether the benefit gained from an intervention is worth the cost involved, or, in other words, does it represent “value-for-money”. Policy-makers are increasingly requiring evidence to support the economic credentials of proposed interventions in order to inform decisions around resource allocation. However, the availability of this data has been very limited for several reasons. First, there have been relatively few primary prevention interventions that have effectively reduced the prevalence of obesity or anthropometric outcomes in children or adolescents [17], and secondly, the methods to undertake such economic analyses of large multifaceted, community-based interventions are not well developed. The data that are available are patchy, but The APPLE project in New Zealand, a multifaceted school-based program for 5-12 year old children reported an intervention cost per child of NZD1 281, and an incremental cost-effectiveness ratio (ICER) of NZD664 - 1708 per kg of weight-gain prevented over four years (NZD1.00 = USD0.67) (18-19). A school-based obesity prevention program in the United States was reported as cost-saving [20].

In addition to the above studies, which included empirical data collection, the Assessing Cost-Effectiveness in Obesity (ACE-Obesity) study conducted in Australia in 2004-2005 modeled, using the best available evidence, several multifaceted, school-based programs among the 13 evaluated interventions [21]. Multifaceted school-based programs (with and without physical activity components) were found to be cost-effective [21], whilst an after-school hours activity program [22], a walking school bus program [23] and a TravelSMART Schools program [24] were found not to be cost-effective under the existing delivery arrangements.

The Be Active Eat Well (BAEW) program was a successful intervention, which focused on the provision of opportunities, resources and support to achieve positive changes in communities, children, and their families [25]. The activities focused primarily on promoting healthy eating, physical activity, and healthy weight, with community-driven and context specific decision-making. The intervention was implemented by the local health services, Colac Area Health (CAH), in conjunction with key stakeholder organizations, such as local government. Evaluation and support was provided by the academic partners, Deakin University. The program was shown to be more effective at slowing the rate of weight gain (by ∼1 kg) and waist gain (by ∼3 cm) of primary school-aged children in the intervention area (Colac) over the three-year intervention period than in children in the comparison group [25].

This paper aims to measure the cost-effectiveness and affordability of the BAEW program during the period of direct project funding. It aims to determine if this injection of funds into the community (a “capacity boosting” approach) represented a good investment for government and the community, and is appropriate for other communities endeavoring to prevent or reduce childhood obesity.

Methods and Procedures

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods and Procedures
  5. Results
  6. Discussion
  7. Acknowledgments
  8. References

Overview of economic evaluation

A retrospective cost-effectiveness evaluation was undertaken of the program's high investment phase, which equates to the period between mid 2003 to mid 2006, where there was an injection of external funding into the intervention community (Colac). Whilst the evaluation addresses issues of both allocative efficiency (“what to do”) and technical efficiency (“how to do it”), the focus is on the former—that is, whether a community capacity building approach as employed in BAEW was cost effective as a childhood obesity prevention measure. Economic evaluation involves the incremental analysis of both the costs and outcomes of the intervention against those of the specified comparator (usually “current practice”), to determine if the intervention offers a good return on investment.

The cost-effectiveness evaluation was undertaken from a societal perspective as the intervention was community-based and most of the intervention activities were in settings outside of the health sector (e.g., schools, preschools, take-away food outlets, and sporting clubs). The intervention timeframe was three years, whilst the timeframe for modeling costs and benefits was rest-of-life, until the cohort of children had either died or reached age 100 years [21]. All costs and benefits were discounted at 3% in accordance with the recommendations of the US Consensus Panel on Cost-Effectiveness [26]. The reference year for costing was 2006. ICER were calculated as the incremental costs divided by the incremental benefits, and expressed in terms of Australian dollars (AUD) per unit of body mass index (BMI) saved and AUD/disability-adjusted life year (DALY) averted.

The intervention

The BAEW program was conducted in the rural town of Colac (population 11,000) in the state of Victoria, Australia [27]. Colac Area Health was awarded the contract (AUD100,000 per year over four years) to facilitate BAEW. Other organizations, such as Colac Otway Shire and Colac Neighborhood Renewal, were also involved in the design, planning and implementation through the provision of in-kind support [28]. Deakin University was awarded separate funding for program support, training, and evaluation of the project.

The intervention was complex, multifaceted and whole-of-community, with a particular focus on the primary school setting. Six primary schools and four preschools in Colac participated in the intervention. The intervention strategies and activities were detailed in the BAEW Action Plan (June 2006) and process evaluation reports [28]. The interventions targeted evidence-based behavior change —reduction of television viewing; reduced consumption of sugar sweetened drinks, and increased water consumption; reduced consumption of energy dense snacks and increased consumption of fruit and vegetables; increased active play after school and at weekends; and increased active transport to schools [25].

Current practice

The costs and benefits of the BAEW intervention were incrementally assessed against current practice in 12 primary schools across Victoria's Barwon South Western Region, in which no specific intervention was offered. Current practice covered any initiatives (which may or may not have been school-specific) introduced into the school environment to address concerns about healthy eating, physical activity, or childhood obesity, over and above normal school curriculum activities (such as physical education classes) which are common to all schools (both intervention and control).

Assessment of benefit

Health benefits arising from BAEW were, in the first instance, measured as change in BMI, and secondly, as, Disability-adjusted Life Years (DALYs) saved over the lifetime of the cohort. The advantage of using DALYs (as distinct from Quality-adjusted Life Years [QALYs]) are that they combine both mortality and morbidity. Children's measured anthropometric and demographic data were collected at baseline and intervention completion, using a quasi-experimental, longitudinal design. Given the absence of long-term follow-up results, 100% of the effect size was assumed to be maintained, however, alternative decay of effect scenarios were tested under sensitivity analyses.

The reduction in BMI for the intervention was converted to a saving in terms of DALYs using a model developed as a component of the ACE-Obesity study undertaken by Deakin Health Economics in 2004-2005 in conjunction with the Victorian Department of Human Services. The details of this model are reported separately [21]. In brief, DALYs averted as a result of participation in the intervention were calculated as the difference in future mortality and morbidity outcomes between current practice and the intervention scenario. The impact of the change on BMI distribution, as a result of the intervention, was determined using potential impact fractions (PIFs), defined as the proportional change in expected disease or death, that is attributable to a change in exposure to the risk factor in the population. The diseases for which PIFs were calculated were ischaemic heart disease, ischaemic stroke, hypertensive heart disease, type 2 diabetes, osteoarthritis, endometrial cancer, colon cancer, postmenopausal breast cancer, and kidney cancer. The model takes the current prevalent cohort of children and adolescents (age 5–19 years) and follows them in five-year age and gender groups for their remaining life span until death (or age 100 years). The model [21] is based on 2001 population, epidemiology, and disease cost data, which means that there is a mismatch between the reference year used for costing (2006) and for DALY and cost-offset calculations, however, the task of updating the model to the BAEW reference year was beyond the scope of this study.

Assessment of costs

Pathway analysis (who does what, to whom, when, where, and how often) was used to identify the component activities of the intervention in order to ascertain the associated resource utilization. The use of a societal perspective meant that all costs, unless otherwise specified, were included regardless of where the cost was borne. The intervention was measured in steady state, and costs were reported both with and without the evaluation component.

The opportunity cost principle was employed in determining whether costs should be included. Costs were assigned to the time expended by staff, parents, and volunteers in delivering or attending intervention activities, where their attendance diverted them away from their usual work or leisure activities. Key costs excluded were: costs of student time expended in participating in intervention activities; costs associated with any spin-off activities; and the costs of changes in the physical activity or eating patterns of participating families as a result of BAEW participation.

Resource use associated with the intervention was derived using pathway analysis, built-up from the ten objectives, 33 sub-objectives, and 182 activities recorded in the BAEW Action Plan [25]; the detailed process evaluation data collected prospectively during the program; and the detailed process evaluation reports prepared by CAH around each of the strategic objectives (http://www.goforyourlife.vic.gov.au/hav/articles.nsf/pracpages/Be_Active_Eat_Well). The comprehensive process evaluation data and reports enabled the extraction of detailed data on costs (materials, direct costs, travel, etc.) incurred and personnel time invested, for each of the program objectives. The resource use data collection was further supplemented with detail from financial acquittals, interviews with key stakeholders involved in the program, plus interviews with personnel from each of the six participating primary schools. Within the comparator area, information on current practice was obtained from schools through interviews with principals and key staff (who were at the school during the time of the BAEW program), newsletters and reports.

All costs were measured in Australian dollars (AUD1.00 = USD0.748447 as at 1 July 2006) and adjusted to real prices for the 2006 reference year using the relevant consumer price index [29]. For resources not traded in the market place (such as donated goods or volunteer time), a market value was imputed. Unit prices were obtained from the most up-to-date and recognized sources. Table 1 provides a summary of unit cost information, data sources, and assumptions. Costs are reported by item type, payer sector, and strategic objective.

Table 1. Summary of unit cost information, data sources and assumptions
Element costedUnit costs (per hour) 2006 AUDaSourceAssumption
  1. a

    AUD1.00 = USD0.748447 (as at 1 July 2006)

  2. b

    All personnel rates are based on rates per hour, unless otherwise specified.

  3. c

    Hourly rates include a 30% loading to cover salary oncosts (including superannuation, recreational leave, etc), unless otherwise specified.

  4. d

    No additional allowance made for salary oncosts.

Personnel Timeb,c   
Average wage rate26.30Australian Bureau of Statistics [30]Average hourly earnings for all full-time non-managerial adult employees = AUD26.30. Assumed to be cost for persons outside of the public sector or other organisations where hourly rate is not available.
Volunteer/Parent time6.58Jacobs and Fassbender [31]Leisure time is valued at 25% of the average wage rate.
School principal71.86Victorian Department of Education and Early Childhood Development salary rates, 2008Midpoint of salary scale used.
Assistant school principal67.33Victorian Department of Education and Early Childhood Development salary rates, 2008Midpoint of salary scale used.
Teacher36.21Victorian Department of Education and Early Childhood Development salary rates, 2008Midpoint of salary scale used.
Casual teachers (per day)255.47Victorian Department of Education and Early Childhood Development salary rates, 2008As above. Costs to replace teacher if out of school
School nurse38.75Victorian Department of Education and Early Childhood Development salary rates, 2008Midpoint of salary scale used.
Canteen Attendant17.71Victorian Department of Education and Early Childhood Development salary rates, 2008Minimum rate
Social marketing consultant54.00Deakin University, Enterprise bargaining agreement 2005 -2008Based on Associate Professor rates
Casual research assistant30.42Deakin University, Enterprise bargaining agreement 2005 -2008 
Chief Executive Officer98.32State Services Authority, Victorian public service executive employment handbook, State Govt of Victoria 2007.Mid point - Band 3
Top Level Public Servants58.62Australian Public Service Commission Collective Agreement 2006-09Middle band of EL2 range. Also assumed to be salary for senior personnel in local and state government and community health
Mid Level Public Servants41.45Australian Public Service Commission Collective Agreement 2006-09Assumed to be salary for all personnel with a professional involvement in the program, plus for all managerial and coordinator positions
Community Development Officer33.11State Services Authority, Victorian public service executive employment handbook, State Govt of Victoria 2007.Midpoint salary, Class III 3rd year assumed
Food Service Chef23.91Work Place Authority, 2006 Pay Scale Summaries, Australian Government, 2007Chef Grade A - Food Service 8 (Highest bracket for Health and Allied Services)
Manager Nutrition & Dietetics47.46Work Place Authority, 2006 Pay Scale Summaries, Australian Government, 2007Manager Nutrition & Dietetics - Grade 4, Classification: JC1 (Gr 4 Yr 1) up to JC41 (Gr 4 Yr 5)
Dietitian32.17Colac Area HealthAn average of hourly rates paid to dietitians on staff
Paediatriciand285.68Medicare Benefits Schedule, 2009, Department of Health and Ageing45 minutes consultation
GPd94.35Medicare Benefits Schedule, 2009, Department of Health and AgeingItem 23 Level B, Category 1 – Professional attendances
Practice Nurse37.43Work Place Authority, 2006 Pay Scale Summaries, Australian Government, 2007Nurse Practitioner - Year 2
Retailers20.19Australian Bureau of Statistics [36]Average wage rate
Administrative and Clerical Employees25.36Work Place Authority, 2006 Pay Scale Summaries, Australian Government, 2007Mid band of administration support staff
Other costs   
Catering (per person per day)24.56www.deakin.edu.auUsed where cost and invoices unavailable. Includes lunch and snacks
Travel ($/km)0.67Royal Automobile Club of Victoria www.racv.com.auAssumes use of medium 2-3 litre vehicles. Excludes travel to meetings within 10km
Bus hire (daily rate)191.55Hertz 200925 seat bus
Training resources (per person)1.50EstimatePaper/printing costs for training handouts

Uncertainty analysis

Given the need to make assumptions because of the uncertainty or lack of evidence around some parameters, extensive use was made of probabilistic uncertainty analysis (Table 2). Simulation-modeling techniques (using the @RISK software and Monte Carlo simulations) were used to facilitate the presentation of a 95% uncertainty range around the health benefits, costs, and ICERs. In uncertainty analysis, all relevant variables are allowed to vary simultaneously, with parameters selected randomly from the specified range. Given the fact that resource use and associated costs were based on detailed process evaluation data, uncertainty distributions were not attached to intervention costs.

Table 2. Uncertainty distributions used in modelling
ParametersValuesUncertainty distributionSources and assumptions
  1. a

    Values are minimum, most likely and maximum.

  2. b

    In a triangular distribution, the greatest probability of being chosen is the value representing the top of the triangle (the most likely value), whereas the probability of other values being chosen tapers off towards the extremes of the base of the triangle (minimum and maximum values).

  3. c

    As the ACE-Obesity BMI to DALY model,only commences at age 5, the DALYs averted for children < 5 years were calculated by running the relevant BMI savings through the age 5-9 years component of the model.

  4. d

    AUD = Australian dollars (AUD1.00=USD0.748447).

  5. e

    Values are minimum and maximum.

  6. f

    In a uniform distribution, every value in the specified range has an equal probability of being chosen in each iteration of the simulation.

BMI saved per child
Boys <5 years (n=65)−0.3449 (−1.2348; 0.5448)aTriangularbEffectiveness study [25]
Boys 5-9 years (n=592)0.0453 (−0.7318; 0.6412)aTriangularbEffectiveness study [25]
Boys10-12 years (n=222)1.1551 (0.2985; 2.0116)aTriangularbEffectiveness study [25]
Girls <5 years (n=88)−0.8148(−1.7963; 0.1668)aTriangularbEffectiveness study [25]
Girls 5-9 years (n=623)0.5348 (−0.1411; 1.2170)aTriangularbEffectiveness study [25]
Girls 10-12 years (n=213)1.2758 (−0.9015; 3.4532)aTriangularbEffectiveness study [25]
DALYs averted per child
Boys <5 yearsc−0.0051 (−0.01594; 0.0089)aTriangularbCalculated using ACE-Obesity BMI to DALY model [21]
Boys 5-9 years−0.0007 (−0.0099; 0.0105)aTriangularbAs above
Boys10-12 years0.01958 (0.0044; 0.0377)aTriangularbAs above
Girls <5 yearsc−0.0102(−0.0205; 0.0023)aTriangularbAs above
Girls 5-9 years0.0077 (−0.0020; 0.0187aTriangularbAs above
Girls 10-12 years0.0191(−0.0114; 0.0658)aTriangularbAs above
Cost-offsets per child (AUDd)
Boys <5 yearsc−AUD15.05 (−AUD47.43; AUD25.97)aTriangularbAs above
Boys 5-9 years−AUD2.07 (−AUD29.12; AUD30.67)aTriangularbAs above
Boys10-12 yearsAUD57.60 (AUD13.23; AUD109.75)aTriangularbAs above
Girls <5 yearsc−AUD30.52 (−AUD62.94; AUD67.85)aTriangularbAs above
Girls 5-9 yearsAUD22.10 (−AUD6.01; AUD52.69)aTriangularbAs above
Girls 10-12 yearsAUD53.82 (−AUD35.05; AUD163.20)aTriangularbAs above
Gross costs of current practice (AUDd)AUD75 461; AUD150 922eUniformfMinimum estimated from collected data; maximum is an estimate (double minimum value)

Sensitivity analysis

Given uncertainty around maintenance of the intervention effect, different scenarios were tested to determine what component of the effect would need to be lost before the intervention was rendered cost-ineffective. A scenario also tested the impact if only 50% of children received the benefit.

Modeling to a national level

To facilitate comparison with interventions evaluated as part of the ACE-Obesity study [21], the intervention was also modeled at a national level for one year. Given the number of programs competing for time in the school setting, it was conservatively assumed that the intervention would be taken up by10% of Australian primary schools. In the primary scenario, it was assumed that the benefit would be received by 100% of children in those schools, but the impact if a lower proportion received the benefit was tested under sensitivity analysis.

Results

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods and Procedures
  5. Results
  6. Discussion
  7. Acknowledgments
  8. References

Effectiveness

As reported in an earlier paper, children in the BAEW intervention showed lower increases in BMI scores −0.28(−0.7; 0.15) than those in the comparator group [25]. Whilst this difference in BMI between the intervention and comparator groups at follow-up was not statistically significant, it was trending in the right direction. Furthermore, children in the intervention group gained less weight −0.92 (−1.74; −0.11), and showed significantly lower increases in waist circumference −3.14(−5.07; −1.22), BMI-Z score −0.11(−0.21 to−0.01) and waist/height ratio −0.02(−0.03; −0.004) compared with the comparison population.

Intervention costs

The average annual cost of the intervention was AUD61 200 per school community. Personnel time accounted for the majority of costs (64.1%), and the key sectors that incurred costs were health (as the intervention was delivered by Colac Area Health) and local government (Table 3). In terms of strategic objectives, the largest component of costs (42.2%) was spent on the active play objective, given the substantial expenditure incurred on personnel and venues associated with after-school programs.

Table 3. Annual average incremental costs per school community of the intervention by item type, payer sector, and strategic objective ($AUD), 2006
 Incremental one year costsa,b
Cost parameterNo.% of total
  1. a

    Incremental costs = costs of intervention minus costs in comparator area.

  2. b

    Costs are exclusive of evaluation, which is specified separately in Objective 3.

Cost by item type  
Personnel$39 20364.1
Professional development-$220-0.4
Resources and equipment$4 2647.0
Travel$3060.5
Floor space$7281.2
Administration$1 6122.6
Other$15 30625.0
Total$61 199100.0
Cost by payer sector  
Health$36 76660.0
Local government$15 31725.0
Education$9051.5
Recreation and sport$7 28211.9
Commercial$4100.7
Other$5540.9
Total$61 234100.0
Costs by strategic objective  
Objective 1: Capacity building$12 34920.2
Objective 2: Awareness raising$6 45810.6
Objective 4: Decrease television viewing$2750.1
Objective 5: Decrease soft drink, increase water consumption$6431.1
Objective 6: Increase fruit and vegetable consumption$2 3953.9
Objective 7: Increase active transport to and from school$3 6696.0
Objective 8: Increase active play after school and at weekend$25 80142.2
Objective 9: Improve quality of deep-fried takeaway chips$2 5234.1
Objective 10: Pilot a Healthy Happy Families program$7 09111.6
Total (excluding evaluation)$61 204100.0
Objective 3: Evaluation$19 789 
Total (including evaluation)$89 091 
Table 4. Cost Effectiveness results ($AUD), 2006
 Colac interventionModelled for 10% Australian population
  1. BMI Body Mass Index; DALY Disability-Adjusted Life Year; M million; AUD1.00 = USD0.748447; dominated = lower health gains for greater costs.

Total BMI saved547 (−104;1 209)82,899 (−24 428; 192 013)
Total DALYs saved10.2 (−0.19;21.58)1 521 (−142; 3 356)
Incremental cost$0.34M ($0.31M;$0.38M)$37.65M ($34.35M; $41.04M)
Gross cost per BMI saved$576 (dominated;$6,125)$399 (dominated; $3 983)
Gross cost per DALY saved$32 429 (dominated;$0.26M)$22 978 (dominated; $0.21M)
Cost offsets$27 311 (−$1 803; $58 242)$4.10M (−$0.56M; $9.17M
Net cost per DALY saved (with cost-offsets)$29 798 (dominated;$0.26M)$20 227(dominated; $0.20M)

Every AUD1.00 of project funding invested generated an additional AUD2.80 worth of activity at the community level (excluding the evaluation). When the evaluation was included, this equated to AUD3.35 generated by the community for every AUD1.00 invested.

The total cost to evaluate BAEW (including both the intervention and comparator areas) was AUD1.16M. This meant that the cost of evaluating this large demonstration project was similar to the cost of planning and delivery of the intervention itself (AUD1.15M). Demonstration projects are typically more highly evaluated than regular programs, the purpose being to demonstrate their effectiveness and transferability.

Cost-effectiveness

For Colac, the intervention resulted in savings of 547 BMI units (−104; 1209), which translated to 10.2 (−0.19; 21.58) DALYs averted. This resulted in modest cost-offsets of AUD27 311 (-AUD1803; AUD58 242). Therefore, under current modeling assumptions, the BAEW intervention was cost-effective as an obesity prevention intervention with net costs per DALY saved of AUD29 798 (dominated; AUD0.26M) (“dominated” means lower health gains for greater costs). The majority of the 6000 iterations modeled entailed increased benefits coupled with greater costs. On current assumptions, there was a 73.2% chance that the intervention would cost less than the commonly used Australian benchmark for cost-effectiveness of AUD50 000 per DALY saved [32, 33] (Figure 1). There were only 169 or 2.7% of iterations, which involved a health loss coupled with additional costs. The wide uncertainty intervals were because of the negative change in BMI values for some age and gender groups (drawn from the effectiveness study and shown in Table 2) modeled as part of the uncertainty analysis.

image

Figure 1. Cost-effectiveness plane.

Download figure to PowerPoint

When modeled for 10% of Australian primary school children, the intervention would reach 181,212 children across 656 schools, and result in savings of 82,899 BMI units and 1521 DALYs. The incremental cost was AUD37.65M, which translates to a gross cost per BMI unit saved of AUD399. It resulted in modest cost-offsets of AUD4.1M, resulting in a net cost per DALY saved of AUD20 227. These results are essentially a straight linear extrapolation of the Colac results. The difference in the cost-effectiveness ratios is merely a product of the variation in the age and gender distribution of children in the intervention arm in Colac compared to that of the national primary school population.

In accordance with the assumptions employed in the ACE-Obesity study [21], the results assumed full maintenance of the BMI benefit into adulthood. If the benefit was only received by 50% of the children in the participating schools, the ICER would be considerably higher, but still just below the cost-effectiveness threshold. If, on the other hand, the population receiving the benefit remain unchanged, but only 50% of the effect was maintained over time, the total BMI units saved would be halved and the cost-effectiveness ratio doubled, meaning that the intervention was no longer cost-effective. At least 70% of the intervention effect would need to be retained to ensure the intervention remained cost-effective.

Discussion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods and Procedures
  5. Results
  6. Discussion
  7. Acknowledgments
  8. References

The BAEW intervention was cost-effective (measured against the benchmark of AUD50,000 per DALY averted) in terms of its effect on BMI in children under the current assumptions. Despite the wide confidence intervals around the ICER, the modeling indicated the likelihood that the intervention would fall below the cost-effectiveness threshold in the case of the majority of modeling iterations. The evaluation shows that such a multi-faceted community capacity-building approach to childhood obesity prevention has a high chance (73.2%) of offering “value-for-money” when the benefits are measured against the costs involved.

This paper reports costs in a number of different ways, each of which reflects different project objectives. The total intervention cost provides an indicator of the affordability of the intervention. Based on a cost of AUD37.7M to deliver the intervention to 10% of Australia's primary school children, the cost of extending this to a larger proportion of children may or may not be prohibitive, depending on the size of budget allocations to the prevention of childhood obesity in Australia. The total intervention costs may be somewhat inflated by the inclusion of some cost items incurred by other government programs, such as the Walking School Bus and the Active After School Communities Program, which were running concurrently in some intervention and comparator schools. Other factors contributing to the high costs are the fact that context-specific programs generally are more expensive than standardized programs. Evaluation costs would also presumably be lower. On the other hand, translation to scale will involve additional costs related to national coordination and different contexts.

A further factor that may lead to over-stating of the gross intervention costs is the potential under-reporting of the costs of current practice in the comparator schools. This economic evaluation was undertaken retrospectively following the end of the BAEW project, rather than alongside the effectiveness study. Whilst we are confident that the intervention costs are accurately reflected, given that they were based on detailed process data, there is less certainty surrounding comparator costs, which were based on stakeholder recall. The recording of current practice was hampered by a relatively high level of staff turnover in comparator schools in the period since completion of the intervention. To allow for this potential under-estimation, a multiplier of 2 was placed around the comparator costs as part of the uncertainty analysis. However, an audit of nutrition-related initiatives in the region conducted in 2002 suggested that the level of activity actually occurring was low [34].

The net costs take into account the relatively modest cost-offsets achievable by the intervention. The achievability of these long-term savings in costs will be dependent on the maintenance of the intervention effect over time. Given that the maintenance of effect is unknown, it is standard practice in economic analyses to begin with an assumption of 100% maintenance of effect to allow for a subsequent estimation of how the effect could be reduced before the intervention would cease to be cost-effective. This provides valuable information to decision makers about the level of maintenance that needs to be achieved and they can then assess the likelihood of whether that is achievable.

We acknowledge that an assumption of 100% maintenance of effect is a limitation of the study and is likely to overstate the cost-effectiveness results. There is a dearth of published evidence relating to long-term follow-up of weight reduction programs in children [15], which would underpin an alternative assumption on which to base the quantitative modeling. As a consequence, we have qualified this assumption by conducting a threshold analysis to determine the extent to which it could be relaxed before the intervention would cease to be cost-effective [21]. The BAEW intervention would remain cost-effective if effectiveness remained at 70% or higher. Having said this, it is likely that our original calculation understated the intervention's cost-effectiveness. The calculated benefit in terms of DALYs averted is likely to be conservative given that the DALYs arising from obesity-related diseases would be greater had we modeled them for the 2006 reference year in line with costs rather than for 2001.

Finally, the intervention compares favorably in terms of cost per child with interventions assessed as part of the ACE-Obesity project [35]. When modeled, the BAEW intervention cost AUD344 per child, which falls within the range of costs (AUD0.54 to AUD2908) of the eight nontargeted interventions evaluated as part of the ACE-Obesity study.

The costs reported here are exclusive of the evaluation component delivered by Deakin University. The actual economic cost of the evaluation of this large demonstration project across both the intervention and comparator areas was AUD1.14M over the three years project period. This is equivalent to the actual cost of planning and delivering the intervention (AUD1.12M), which demonstrates the very high cost associated with the provision of evaluation to such large community-based obesity demonstration projects. However, given that the purpose of a demonstration project is to inform future activities, they warrant substantial (if not equivalent) resources being allocated for rigorous and comprehensive evaluation. The need for high quality evaluations is well recognized and is critical to inform governments and communities' investments [36, 37].

This evaluation has been concerned with the measurement of all economic costs associated with delivering the BAEW intervention by taking into account all resources used, irrespective of whether they were paid for (such as donated goods and services). As a consequence, the gross intervention costs far exceed the funding received by Colac Area Health to underpin implementation of the intervention. Every one dollar of project funding invested generated an additional AUD2.80 worth of activity at the community level (excluding Deakin University evaluation and support). This is illustrative of the enormous capacity of projects like BAEW to generate substantial spin-off investment and activity over and above the original project funding levels.

This is one of the first community-based intervention studies targeting childhood obesity that has been subjected to an economic evaluation [18, 19]. The application of economic evaluation to complex, multifaceted, nonstandardized interventions in multisettings is innovative in itself. However, the establishment of the cost-effectiveness and affordability of this large-scale demonstration approach is only the first stage in a long process. The next task is to determine what this means if the approach was translated to scale—what are the fundamental elements, and which elements could be omitted without impacting on the intervention effect. The other question is what cost efficiencies would be gained through scaling up of the project. Beyond that, the final stage will entail the integration of these findings into the existing system. This current stage of economic evaluation is a precursor to building that systems-based model for obesity prevention.

Acknowledgments

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods and Procedures
  5. Results
  6. Discussion
  7. Acknowledgments
  8. References

We acknowledge all primary schools, principals and school staff, parent and students for their contributions during this phase of the study. We also thank Ruth Cutler and Mark Brennan who provided essential support to carry out this study. The authors also thank Dr Michelle Haby and the Department of Health for permission to use the BMI to DALY model developed as part of the Assessing Cost-Effectiveness In Obesity (ACE Obesity) project.

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods and Procedures
  5. Results
  6. Discussion
  7. Acknowledgments
  8. References
  • 1
    Lobstein T, Baur L, Uauy R. Obesity in children and young people: A crisis in public health. Obes Rev 2004;5:4-85.
  • 2
    Magarey AM, Daniels LA, Boulton TJC. Prevalence of overweight and obesity in Australian children and adolescents: Reassessment of 1985 and 1995 data against new standard international definitions. Med J Aust 2001;174:561-564.
  • 3
    Booth ML, Chey T, Wake M, et al. Change in the prevalence of overweight and obesity among young Australians, 1969-1997. Am J Clin Nutr 2003;77:29-36.
  • 4
    Olds TS, Tomkinson GR, Ferrar KE, Maher CA. Trends in the prevalence of childhood overweight and obesity in Australia between 1985 and 2008. Int J Obes 2009;34:57-66.
  • 5
    Smith A, Coveney J, Carter P, Jolley G, Laris P. The Eat Well SA project: an evaluation-based case study in building capacity for promoting healthy eating. Health Promot Int 2004;19:327-334.
  • 6
    de Silva-Sanigorski AM, Bell AC, Kremer P, et al. Reducing obesity in early childhood: results from Romp & Chomp, an Australian community-wide intervention program. Am J Clin Nutr 2010;91:831-840.
  • 7
    de Silva-Sanigorski AM, Bolton K, Haby M, et al. Scaling up community-based obesity prevention in Australia: background and evaluation design of the Health Promoting Communities: Being active eating well initiative. BMC Public Health 2010;10:65-65.
  • 8
    De Henauw S, Verbestel V, Marild S, et al. The IDEFICS community-oriented intervention programme: a new model for childhood obesity prevention in Europe? Int J Obes 2011;35 (Suppl 1):S16-S23.
  • 9
    Borys JM, Le Bodo Y, Jebb SA, et al. EPODE approach for childhood obesity prevention: Methods, progress and international development. Obes Rev.2012;13:299-315.
  • 10
    Gortmaker SL, Peterson K, Wiecha J, et al. Reducing obesity via a school-based interdisciplinary intervention among youth: Planet Health. Arch Pediatr Adolesc Med 1999;153:409-418.
  • 11
    Luepker R, Perry CL, McKinlay SM, et al. Outcomes of a field trial to improve children's dietary patterns and physical activity. The Child and Adolescent Trial for Cardiovascular Health. JAMA 1996;275:768-76.
  • 12
    Sallis JF, McKenzie TL, Alcaraz JE, et al. The effects of a 2-year physical education program (SPARK) on physical activity and fitness in elementary school students. Sports, Play, and Active Recreation for Kids. Am J Public Health. 1997;87:1328-1334.
  • 13
    Foltz JL, May AL, Belay B, et al. Population-level intervention strategies and examples for obesity prevention in children. Annu Rev Nutr 2012;32:391-415.
  • 14
    Waters E, de Silva-Sanigorski A, Hall BJ, et al. Interventions for preventing obesity in children. Cochrane Database Syst Rev 2011;12:CD001871.
  • 15
    Shaya FT, Flores D, Gbarayor CM, Wang J. School-based obesity interventions: A literature review. J School Health 2008;78:189-196.
  • 16
    Dalziel K, Segal L, Mortimer D. Review of Australian health economic evaluation—245 interventions: What can we say about cost effectiveness? Cost Eff Resour Alloc 2008;6:1-12.
  • 17
    Millar L, Kremer P, de Silva-Sanigorski A, et al. Reduction in overweight and obesity from a 3-year community-based intervention in Australia: The ‘It's Your Move!’ project. Obes Rev 2011;12(Suppl 2):20-28.
  • 18
    McAuley KA, Taylor RW, Farmer VL, et al. Economic evaluation of a community-based obesity prevention program in children: The APPLE Project. Obesity 2010;18:131-136.
  • 19
    Taylor RW, McAuley KA, Barabezat W, Farmer VL, Williams SM, Mann JI. Two-year follow-up of an obesity prevention initiative in children: The APPLE project. Am J Clin Nutr 2008;88:1371-1377.
  • 20
    Wang LY, Yang Q, Lowry R, Wechsler H. Economic analysis of a school-based obesity prevention program. Obesity 2003;11:1313-1324.
  • 21
    Haby MM, Vos T, Carter R, et al. A new approach to assessing the health benefit from obesity interventions in children and adolescents: The assessing cost-effectiveness in obesity project. Int J Obesity 2006;30:1463-1475.
  • 22
    Moodie ML, Carter RC, Swinburn BA, Haby MM. The cost-effectiveness of Australia's Active After-school Communities Program. Obesity 2010;18:1585-1592.
  • 23
    Moodie M, Haby M, Galvin L, Swinburn B, Carter R. Cost-effectiveness of active transport for primary school children—Walking School Bus program. Int J Behav Nutr Phys Act 2009;6:6-63.
  • 24
    Moodie M, Haby M, Swinburn B, Carter R. Assessing cost-effectiveness in obesity: Active transport programs for primary school children—TravelSMART Schools Curriculum Program’. J Phys Act Health 2011,8:503-515.
  • 25
    Sanigorski AM, Bell AC, Kremer PJ, Cuttler R, Swinburn BA. Reducing unhealthy weight gain in children through community capacity-building: Results of a quasi-experimental intervention program, Be Active Eat Well. Int J Obesity 2008;32:1060-1067.
  • 26
    Gold MR SJ, Russell LB, Weinstein MC (eds). Cost-effectiveness in Health and Medicine. New York: Oxford University Press; 1996.
  • 27
    Australian Bureau of Statistics. 3235.0—Population by Age and Sex, Australia, 2006. ABS, 2007.
  • 28
  • 29
    ABS. Consumer Price Index cat. no. 6401.0. Canberra, 2008.
  • 30
    ABS. Employee Earnings and Hours, Cat. 6306.0. Australian Bureau of Statistics: Canberra, 2006.
  • 31
    Jacobs P, Fassbender K. The measurement of indirect costs in the health economics evaluation literature. A review. Int J Technol Assess Health Care 1998;14:799-808.
  • 32
    Haby MM CR, Mihalopoulos C, Magnus A, Sanderson K, Andrews G, Vos T. Assessing cost-effectiveness -mental health: Introduction to the study and methods. A N Z J Psychiatry 2004;38:569-578.
  • 33
    George B HA, Mitchell A. Cost-effectiveness analysis and the consistency of decision making: evidence from pharmaceutical reimbursement in Australia (1991 to 1996). Pharmacoeconomics 2001;19:1103-1109.
  • 34
    Bell AC. Nutrition programs and research in the Barwon-South Western Region. Deakin University and Department of Human Services (Barwon-South Western Region), 2002.
  • 35
    Carter R, Moodie M, Marwick A, et al. Assessing cost-effectiveness in obesity (ACE-Obesity): An overview of the ACE approach, economic methods and cost results. BMC Public Health 2009;9:419.
  • 36
    Swinburn B BC, King L, Margarey A, O'Brien K, Waters E,. Obesity prevention programs demand high-quality evaluations. Aust N Z J Public Health 2007;31:305-307.
  • 37
    Swinburn B GT, Kumanyika S. Obesity prevention: A proposed framework for translating evidence into action. Obes Rev 2005;6:23-33.