Skip to main content
Advertisement
Browse Subject Areas
?

Click through the PLOS taxonomy to find articles in your field.

For more information about PLOS Subject Areas, click here.

  • Loading metrics

Cost-Effectiveness of Pharmacotherapy to Reduce Obesity

Abstract

Aims

Obesity causes a high disease burden in Australia and across the world. We aimed to analyse the cost-effectiveness of weight reduction with pharmacotherapy in Australia, and to assess its potential to reduce the disease burden due to excess body weight.

Methods

We constructed a multi-state life-table based Markov model in Excel in which body weight influences the incidence of stroke, ischemic heart disease, hypertensive heart disease, diabetes mellitus, osteoarthritis, post-menopausal breast cancer, colon cancer, endometrial cancer and kidney cancer. We use data on effectiveness identified from PubMed searches, on mortality from Australian Bureau of Statistics, on disease costs from the Australian Institute of Health and Welfare, and on drug costs from the Department of Health and Ageing. We evaluate 1-year pharmacological interventions with sibutramine and orlistat targeting obese Australian adults free of obesity-related disease. We use a lifetime horizon for costs and health outcomes and a health sector perspective for costs. Incremental Cost-Effectiveness Ratios (ICERs) below A$50 000 per Disability Adjusted Life Year (DALY) averted are considered good value for money.

Results

The ICERs are A$130 000/DALY (95% uncertainty interval [UI] 93 000–180 000) for sibutramine and A$230 000/DALY (170 000–340 000) for orlistat. The interventions reduce the body weight-related disease burden at the population level by 0.2% and 0.1%, respectively. Modest weight loss during the interventions, rapid post-intervention weight regain and low adherence limit the health benefits.

Conclusions

Treatment with sibutramine or orlistat is not cost-effective from an Australian health sector perspective and has a negligible impact on the total body weight-related disease burden.

Introduction

Obesity in adults is a global public health concern because excess weight increases the relative risk of disease and mortality [1], [2], [3]. A range of diseases, notably cardiovascular disease, diabetes and a number of cancers, are related to excess weight [4]. High body mass was responsible for 7.5% of the total burden of disease and injury in Australia in 2003 [5].

Several treatment options for obesity exist. Obese persons may lose up to 5 to 8.5 kilograms with diet [6], 1 to 2 kg with exercise [7], 2.5 to 5.5 kg with drugs [8], or 20 to 30 kg with bariatric surgery [9].

Most weight loss drugs are either lipase inhibitors or appetite suppressants. Sibutramine (also known under trade names Meridia or Reductil) is a norepinephrine and serotonin reuptake inhibitor. Sibutramine putatively aids weight loss by suppressing appetite and increasing thermogenesis. Orlistat (also known as Xenical or Alli) is a lipase inhibitor which is assumed to aid weight loss by preventing the digestion and absorption of dietary fat. Following a standard dose of 120 mg three times per day before meals, orlistat inhibits approximately 30% of fat absorption [10].

In Australia, sibutramine has been listed for reimbursement since 2001 for obese patients with an initial body mass index (BMI) greater than or equal to 30 kg per square metre or greater than or equal to 27 kg per square metre in the presence of other obesity-related risk factors (e.g. diabetes, dyslipidemia, hypertension). Up to 2010, sibutramine could be prescribed to patients who have not adequately responded to an appropriate weight-reducing regimen alone (hypo caloric diet and/or exercise) [11], [12]. Orlistat is reimbursed for patients with a BMI greater than or equal to 35 with no known co-morbidities, or a BMI greater than or equal to 30 with co-morbidities (diabetes, ischemic heart disease, psychiatric conditions, and/or hypertension). Patients are eligible for 1 continuous 12-month treatment in a lifetime, and should receive dietetic and weight management advice. Orlistat is discontinued after 3 months if the patient's initial body weight has not been reduced by 2.5 kg or 2.5% (whichever is the lesser), and after 6 months if the loss was less than 5 kg or 5% [13].

Cost-effectiveness is one of the criteria used to decide whether to fund health care interventions. Cost-effectiveness analysis is a form of economic analysis that compares the relative costs and outcomes (effects) of two or more courses of action, one of which is often ‘do nothing’ or ‘business as usual’. In the health field, the outcomes are often expressed in quality-adjusted life years (QALYs) or disability-adjusted life years (DALYs) [14]. Both of these measures combine effects on duration and quality of life, but while the QALY is a measure of health gain, the DALY is a measure of health loss. The DALY concept was developed for the Global Burden of Disease study (GBD) in the early 1990s [15]. To enable comparison of the severity of different diseases, a coherent set of ‘disability weights’ was based on expert opinion. The DALY used in economic evaluations varies from the DALY in burden of disease studies: a) fatal health loss is based on mortality rates in the population of interest rather than being mirrored against a standard life expectancy; and b) the GBD age weights are not used. In fact, the only difference between estimating DALYs and QALYs in our cost-effectiveness studies is the use of the GBD disability weights rather than QALY utility weights.

A recent systematic review of cost-effectiveness analyses of pharmacological anti-obesity treatments found that most published studies reported findings that were within the cost-effective range [16]. However, 11 of the 14 studies were sponsored by the companies that market the drugs, and the 3 independent studies found much higher incremental costs per quality-adjusted life year (an average of €62,000/QALY for independent studies versus €15,000/QALY for industry-sponsored studies). Although attrition rates in anti-obesity pharmacotherapy are high [17], this was not clearly accounted for in any of the studies. The authors of the review concluded that there was a need for independently conducted studies that explicitly address the high discontinuation rates and compare different drugs head-to-head. The present research was carried out independently from industry and evaluates the cost-effectiveness of treatment with sibutramine or orlistat to reduce obesity in Australia, 2003. In addition, it estimates by how much these drugs could potentially reduce the burden of disease due to excess body mass.

The study forms part of the ‘Assessing the Cost-Effectiveness (ACE) Prevention’ project, which was established to provide a comprehensive analysis of the cost-effectiveness of interventions that address the non-communicable disease burden in Australia. The project takes a health sector perspective and applies a lifetime time horizon. The rationale and methodology have been explained elsewhere [18]. Earlier studies reported on weight reducing diets [19], nutrition labelling and a ‘junk food’ tax [20], physical activity interventions [21] and fruit and vegetable interventions [22].

Methods

Modelling health outcomes and costs

We constructed a proportional multi-state life table Markov model [23], [24] to calculate health outcomes resulting from a reduction in average body weight in the population. The model simulates and compares two populations in separate life tables: a reference population based on existing levels of morbidity and mortality for 2003 and an intervention population which is identical except that it receives the intervention. Table 1 lists all effectiveness and costing parameters in the model, along with their distributions and sources.

thumbnail
Table 1. Input parameters relating to effectiveness and intervention costs.

https://doi.org/10.1371/journal.pone.0026051.t001

Interventions.

In the model, pharmaceutical intervention with either sibutramine 120 mg or orlistat 3×15 mg daily is offered to all obese Australian adults (21% of men and 23% of women) [25]. Recruitment and prescribing are assumed to be via opportunistic screening by GPs. Since 84% of Australians are seen by a GP in any one year and we (perhaps optimistically) assume 50% of the eligible population would agree to participate, we estimated recruitment at 42% of the target population. We assume a maximum treatment duration of one year and compared against a no-intervention scenario.

Weight Change.

A recent meta-analysis estimated that sibutramine reduced body mass by 4.45 kg (standard error [SE] 0.426) at the end of 12 months, and orlistat by 2.89 kg (SE 0.316) [10]. A slightly more recent study found very similar results but did not give results specifically at 12 months, which we preferred because it accords well with Australian reimbursement practice and with the structure of our model (which has one-year cycles) [8]. We implicitly assume that adherence in practice was no different from adherence in the trials that were included in the meta-analysis. Consistent with the intent-to-treat analysis used in the primary studies, we apply the weight loss to all persons who started the drug and do not reduce the effect for discontinuation of treatment or non-adherence. Weight which is lost with drugs tends to be regained. For both drugs, we apply the weight regain of 0.385 kg/month observed in the STORM trial with sibutramine [26] to this average weight loss at the end of the intervention year, and assume that no permanent weight loss is achieved.

Health effects.

Effectiveness is modelled by dividing the populations into 5-year cohorts and simulating each cohort in the life table until all persons have died or reached 100 years of age. Years of life lived are adjusted at each age for time spent in poor health due to disease or injury. The model was implemented in Microsoft Excel 2003 (Microsoft Corp., Redmond, Washington).

The model explicitly simulates nine obesity-related diseases: stroke, ischemic heart disease, hypertensive heart disease, diabetes mellitus, osteoarthritis, post-menopausal breast cancer, colon cancer, endometrial cancer, kidney cancer [4].

For each disease, a simple disease model calculates the effect on prevalence and disease-specific mortality in subsequent years as a consequence of changes in disease incidence [27]. Each disease state is assigned a disability weight (Table 2) based on the Australian Burden of Disease Study [5]. Disease specific changes in prevalence and mortality are linked to the life table and influence total mortality rates and the average health-related quality of life at each age and sex, and therefore the total number of disability-adjusted life years lived by the cohort.

thumbnail
Table 2. Disability weights for prevalent diseases, by sex, at baseline [5].

https://doi.org/10.1371/journal.pone.0026051.t002

The reference population is simulated based on the measured body weight distribution in year 1999/2000 [25] and expected future trends in obesity in Australia [28]. The characteristics of the intervention population are the same except that the obese lose a number of kilograms, which lowers the risk of each of the diseases related to excess weight. The change in disease incidence is estimated through Potential Impact Fraction (PIF) calculations [29]. The relative risks used [4], [30] are detailed in Table 3.

thumbnail
Table 3. Relative risks of disease per 1 unit increase of BMI [4], [30].

https://doi.org/10.1371/journal.pone.0026051.t003

Estimates of disease incidence and mortality are based on the Australian Burden of Disease 2003 study [5]. In the absence of coherent data on the prevalence of these diseases among the obese, which are needed to define the baseline situation, we generate estimates using the model itself. This requires assuming the population is in a steady state (i.e., ignoring trends) and defining the intervention scenario as all Australians having a BMI over 30 (by using the part of the BMI distribution that exceeds 30 kg.m2 and inflating it to the size of the total Australian population). This allows calculating PIF values, which are applied to the Australian incidence rates to estimate disease incidence among obese Australians. Running the model over the lifetime of the youngest cohort then gives the estimates of the baseline prevalence that we need as input for our modelling. The prevalence estimates obtained this way for IHD, stroke and diabetes were comparable to those obtained by analysing measured data from the AusDiab study [25] without the random variation by age (results not shown).

Costs.

Intervention costs are assessed from a health sector perspective. We include the costs of pharmaceuticals, GP visits, and lifetime health care costs. We separately present results that additionally include participants' travel costs and the cost of time spent visiting the GP. In the base case scenario, participants' time and travel costs are included. Participants' time is valued at 25% of the wage rate, i.e. A$17.44 per hour [31]. Table 1 summarizes the assumptions and costing figures.

Intervention costs include 10 minutes GP time during the initial GP appointment, the cost of medication for up to one year, and an average of 1.3 (sibutramine) or 1.6 (orlistat) medication-related follow-up visits per person starting on medication. This low number of doctor visits is partly due to non-adherence of 48% for sibutramine and 33% for orlistat [17]. Half of the attrition is estimated to occur immediately after prescription (incurring 1 month of drug costs), the rest on average 6 months after starting therapy (incurring 6 months of both drug costs). Based on information of the Pharmaceutical Benefit Schedule, a full year of medication is estimated to cost A$1,467 for sibutramine and A$1,486 for orlistat.

The intervention costs are partially offset by reduced health care expenditure for diseases related to obesity in later years. For cancers, these costs are evaluated for each incident case averted and for the remaining diseases for each prevalent case averted. The literature is divided as to whether health care costs in the years of life that were added by interventions should or should not be included [32], [33], [34]. We present results both ways. Disease treatment costs are drawn from the Australian Institute of Health and Welfare [35] and are detailed in Table 4.

thumbnail
Table 4. Average health care costs per prevalent or incident case of disease.

https://doi.org/10.1371/journal.pone.0026051.t004

All costs are adjusted to real prices in the 2003 reference year using the relevant Health Price Index from the Australian Institute of Health and Welfare [35], or the Consumer Price Index from the Australian Bureau of Statistics [36] if the costs occurred outside the health sector.

Calculation of DALYs and ICERs.

DALYs averted by the intervention are calculated as the increase in the discounted number of life years lived in the intervention population compared to the reference population, with an adjustment for disability. To show the impact of interventions on the total burden of high body mass in the Australian population, we also express the health gains at the population level as a percentage of the gains that could be achieved under an ideal scenario in which the population has a BMI of 21 (SD 1), which is the theoretical minimum risk scenario used in the World Health Organization's Comparative Risk Assessment study [37], over the rest of their lifetime. We calculate incremental cost-effectiveness ratios (ICERs) as the discounted net cost of the intervention (total cost less cost offsets) divided by the DALYs gained by the intervention, compared to no intervention. Discounting of costs and benefits is at 3% per year [38]. In the ACE Prevention project, interventions with ICERs under A$50,000/DALY (in 2003, A$1 was approximately US$0.67 or €0.58) are considered cost-effective in the Australian context, although in the light of revealed preferences some may consider this a rather low threshold [39].

Uncertainty.

Uncertainty intervals for DALYs, net costs and the ICER are estimated by Monte Carlo simulation (2000 iterations) using the Excel add-in Ersatz (www.epigear.com). Normal distributions are assumed around the change in average weight that results from each intervention and around the natural log of the relative risk of obesity related disease. A triangular distribution was assumed around a few of the costing components. Importantly, no uncertainty distribution was assumed around the cost of medication, which makes up over 90% of the total intervention costs. This is because both drugs are (or, in the case of sibutramine, were) subsidized to this centrally determined price level. Table 1 lists the parameters with their distributions and values.

Sensitivity Analyses.

We explore the range of cost-effectiveness outcomes by including patients' time and travel costs, and by including health care costs for unrelated diseases into the cost-offset calculations. To enable comparison with previous studies, we also assess the effect of halving the rate at which weight is regained; of linear weight regain over 3 years following the cessation of therapy; of assuming 23% of the weight loss is permanent; and of including a utility-weight of 0.017 per BMI-unit lost for one year. Also presented are a scenario without discounting, one in which all obese are assumed free of obesity related disease at baseline, and results by age.

Results

We simulated a total loss of DALYs due to excess body mass over the lifetime of Australians aged 20 years and above in 2003 of 5.8 million (95% confidence range 4.9–6.6 million). Neither of the evaluated interventions averts more than 0.2% of this disease burden (Table 5).

thumbnail
Table 5. Cost-effectiveness of sibutramine and orlistat when compared with current practice.

https://doi.org/10.1371/journal.pone.0026051.t005

With ICERs between A$93 000 and A$350 000 per DALY, treatment of obesity with sibutramine or orlistat is not cost-effective in any of the costing scenarios (Table 5).

The sensitivity analysis (Table 6) shows that halving the rate at which weight is regained does not materially change the results. Assuming weight is regained linearly over 3 years doubles the health impact but does not result in ICERs in the cost-effective range. In contrast, the assumption that 23% of the weight lost is not regained dramatically improves the cost-effectiveness ratios. In this scenario, treatment with either drug would likely be considered cost-effective. Attributing a utility value to BMI-loss per se also greatly improves the cost-effectiveness ratios. This adds so many QALYs to the DALYs we used in our standard methods that it pushes sibutramine into the cost-effective range, with orlistat coming close. Discounting has significant impact on the cost-effectiveness ratios but discounting at 0% does not change the conclusions. Neither does age. In our model, the benefits in the 20–29 age category are zero because the risk of disease below age 35 was assumed unrelated to BMI (Table 3). Above that age the cost-effectiveness improves up to age 60–69, but only to $88 000 per DALY for sibutramine and $160 000 for orlistat. Above that age it deteriorates again due to lower relative risks and the lower life expectancy of the additional survivors. The starting prevalence of obesity-related diseases had no material influence on the results, probably because the model does not vary the risk of incident disease based on medical history.

Discussion

By the criteria used in this study, treating obese persons with sibutramine or orlistat is not cost-effective in the Australian setting and has a negligible impact on the burden of disease due to excess body mass.

Comparison of results with other studies

In contrast to our findings, the majority of previous studies have concluded that sibutramine and orlistat are cost-effective. Our cost-effectiveness ratios for sibutramine and orlistat are over twice as high as those in earlier studies [16]. Two factors explain part of this difference.

Firstly, in many previous studies the rate at which lost weight is regained is more optimistic than in ours. Many studies assume an arbitrary linear regain over 3 or 5 years, while this study uses an empirical rate of regain [26] which results in regain of all lost weight within 2 years, on average. One study assumed that 23% of the weight loss lasts a lifetime [40], based on an older review of dietary interventions (rather than pharmaceutical interventions) [41]. Our sensitivity analysis shows that the rate of weight regain has a modest impact on the results, but the results are very sensitive to assumptions around the permanency of weight loss.

Secondly, many studies attribute a (substantial) ‘utility’ gain to weight loss per se, while this study only includes reductions in disease-related quality and length of life. We thus ignore any improvements in mental health, fitness and wellbeing that may be the result of weight loss, even in the absence of changes in disease status. Adding such effects would improve the cost-effectiveness estimates. A number of previous studies [42], [43], [44], [45], [46] use a utility value of 0.017 per BMI unit lost per year from a pharma-sponsored study that used data from an RCT [47]. This value is high; it would imply that a person with a BMI of 38 and a quality of life of 80% of perfect health could achieve a quality of life better than perfect when reaching normal weight. Studies that use data from trials of weight reducing interventions to derive quality of life values for body mass changes may measure not only health-related quality of life, but also the satisfaction of achieving weight loss or the disappointment of having failed to lose weight, which in our study is not taken into account. In the study from which the 0.017 utility value was derived, the measurement of persons health status with a visual analogue scale was preceded by questions that framed ‘health’ in terms of physical attractiveness, social functioning, health distress and emotions [47]. This may have influenced participants' responses.

Limitations

Consistent with our health sector perspective, our study does not take into account any increases in productivity that may result from reduced BMI. Including these benefits would improve the cost-effectiveness of both interventions for the working-age population. However, this would be partly or completely offset by older persons, who are generally less productive, living longer.

Persons taking weight reducing drugs are also advised to follow a diet. Our analysis did not include a diet, but since the diet recommendation would apply equally to the intervention and the reference group, this makes no difference to the results.

We did not model any persons discontinuing therapy because of insufficient weight loss, but it seems very unlikely that such stopping rules would improve the cost-effectiveness enough to make treatment with either sibutramine or orlistat cost-effective. A quick back-of-the-envelope calculation shows that even if we cost only half of the persons after 3 months but retain the health effect of the full cohort, the ICERs are around A$85 000 for sibutramine and A$160 000 for orlistat.

On the effect side, we assumed the attrition for drug intervention would be the same as in trials, which is probably too optimistic. No additional costs, health risks or loss of quality of life were included for side-effects of the drugs (mainly increased blood pressure and heart rate for sibutramine and gastrointestinal complaints for orlistat). Quite aside from considerations of efficiency, sibutramine was recently shown to increase the risk of cardiovascular disease among subjects with pre-existing cardiovascular conditions [48]. On that basis, its licence has been withdrawn in Australia [12] and in the European Union [49].

Conclusion

Given the expected rates of weight regain and the high costs of medication, the implementation of the pharmaceutical interventions for primary prevention of obesity-related disease in Australia is unlikely to offer value for money. Based on current evidence, the overall effect of these pharmaceuticals on the obesity-related burden of disease is negligible.

Acknowledgments

We would like to thank Health Department partners in the research as well as the input of the ACE Prevention Steering Committee and Technical Groups, and two anonymous reviewers for their helpful comments.

Author Contributions

Conceived and designed the experiments: TV JJB MF. Performed the experiments: MF JLV. Analyzed the data: MF JLV. Contributed reagents/materials/analysis tools: JJB. Wrote the paper: JLV. Contributed to the interpretation of results and writing of the paper: TV JJB.

References

  1. 1. Flegal K, Graubard BI, Williamson DF, Gail MH (2005) Excess Deaths Associated with Underweight, Overweight and Obesity. JAMA 293: 1864–1867.
  2. 2. Gronniger JT (2006) A Semiparametric Analysis of the Relationship of Body Mass Index to Mortality. American Journal of Public Health 96: 173–178.
  3. 3. Peeters AB, Barendregt JJ, Willekens F, Mackenbach JP, Al Mamun A, Bonneux L (2003) “Obesity in Adulthood and its Consequences for Life Expectancy: A Life Table Analysis.”. Ann Intern Medicine 38: 24–32.
  4. 4. James WP, Jackson-Leach R, Ni Mhurchu C, Kalamara E, Shayeghi M, et al. (2004) Overweight and obesity (high body mass index). In: Ezzati M, Lopez AD, Rodgers A, Murray CJL, editors. Comparative quantification of health risks. Geneva: World Health Organization. pp. 497–596.
  5. 5. Begg S, Vos T, Barker B, Stevenson C, Stanley L, et al. (2007) The Burden of Disease and Injury in Australia. Canberra: AIHW.
  6. 6. Franz MJ, VanWormer JJ, Crain AL, Boucher JL, Histon T, et al. (2007) Weight-loss outcomes: a systematic review and meta-analysis of weight-loss clinical trials with a minimum 1-year follow-up. J Am Diet Assoc 107: 1755–1767.
  7. 7. Wing RR (1999) Physical activity in the treatment of the adulthood overweight and obesity: current evidence and research issues. Med Sci Sports Exerc 31: S547–552.
  8. 8. Rucker D, Padwal R, Li SK, Curioni C, Lau DC (2007) Long term pharmacotherapy for obesity and overweight: updated meta-analysis. Bmj 335: 1194–1199.
  9. 9. Maggard MA, Shugarman LR, Suttorp M, Maglione M, Sugarman HJ, et al. (2005) Meta-Analysis: Surgical Treatment of Obesity. Annals of Internal Medicine 142: 547-W-118.
  10. 10. Li Z, Maglione M, Tu W, Mojica W, Arterburn D, et al. (2005) Meta-analysis: pharmacologic treatment of obesity. Ann Intern Med 142: 532–546.
  11. 11. Pharmaceutical Benefits Advisory Committee (2008) Public Summary Document Sibutramine hydrochloride. Canberra: Department of Health and Aging.
  12. 12. Therapeutic Goods Administration DoHaA, Government of Australia (2010) Sibutramine (Reductil) - withdrawal in Australia. Canberra: TGA.
  13. 13. Department of Health and Ageing (2009) Pharmaceutical Benefits Scheme. DHA. Available: www.pbs.gov.au Accessed 2009 Dec 23.
  14. 14. Gold MR, Stevenson D, Fryback DG (2002) HALYS and QALYS and DALYS, Oh My: similarities and differences in summary measures of population Health. Annu Rev Public Health 23: 115–134.
  15. 15. Murray CJ, Lopez AD (1994) Quantifying disability: data, methods and results. Bull World Health Organ 72: 481–494.
  16. 16. Neovius M, Narbro K (2008) Cost-effectiveness of pharmacological anti-obesity treatments: a systematic review. Int J Obes (Lond) 32: 1752–1763.
  17. 17. Padwal R, Li SK, Lau DC (2003) Long-term pharmacotherapy for overweight and obesity: a systematic review and meta-analysis of randomized controlled trials. Int J Obes Relat Metab Disord 27: 1437–1446.
  18. 18. Carter R, Vos T, Moody M, Haby MM, Magnus A, et al. (2008) Priority setting in health: origins, description and application of the Australian Assessing the Cost-Effectiveness of Prevention initiative. Expert Reviews Pharmacoeconomic Outcomes Research 8: 593–617.
  19. 19. Forster M, Veerman JL, Barendregt JJ, Vos T (2011) Cost-effectiveness of diet and exercise interventions to reduce overweight and obesity. Int J Obes (Lond).
  20. 20. Sacks G, Veerman JL, Moodie M, Swinburn B (2010) ‘Traffic-light’ nutrition labelling and ‘junk-food’ tax: a modelled comparison of cost-effectiveness for obesity prevention. Int J Obes (Lond): . Epub ahead of print 2010/2011/2017.
  21. 21. Cobiac LJ, Vos T, Barendregt JJ (2009) Cost-effectiveness of interventions to promote physical activity: a modelling study. PLoS Med 6: e1000110.
  22. 22. Cobiac LJ, Vos T, Veerman JL (2010) Cost-Effectiveness of Interventions to Promote Fruit and Vegetable Consumption. PLoS ONE 5: e14148.
  23. 23. Barendregt JJ, Van Oortmarssen GJ, Van Hout BA, Van Den Bosch JM, Bonneux L (1998) Coping with multiple morbidity in a life table. Math Popul Stud 7: 29–49, 109.
  24. 24. Forster M, Veerman JL, Barendregt JJ, Vos ET (2011) Cost-Effectiveness of Diet and Exercise Interventions to Reduce Overweight and Obesity. Int J Obes. In press.
  25. 25. Cameron AJ, Welborn TA, Zimmet PZ, Dunstan DW, Owen N, et al. (2003) Overweight and obesity in Australia: the 1999–2000 Australian Diabetes, Obesity and Lifestyle Study (AusDiab). Med J Aust 178: 427–432.
  26. 26. James WP, Astrup A, Finer N, Hilsted J, Kopelman P, et al. (2000) Effect of sibutramine on weight maintenance after weight loss: a randomised trial. STORM Study Group. Sibutramine Trial of Obesity Reduction and Maintenance. Lancet 356: 2119–2125.
  27. 27. Barendregt JJ, Van Oortmarssen GJ, Vos T, Murray CJ (2003) A generic model for the assessment of disease epidemiology: the computational basis of DisMod II. Popul Health Metr 1: 4.
  28. 28. Haby MM, Markwick A, Peeters A, Shaw J, Vos T (2011) Future predictions of body mass index and overweight prevalence in Australia, 2005–2025. Health Promot Int.
  29. 29. Barendregt JJ, Veerman JL (2009) Categorical versus continuous risk factors, and the calculation of potential impact fractions. J Epidemiol Community Health.
  30. 30. Ni Mhurchu C, Parag V, Nakamura M, Patel A, Rodgers A, et al. (2006) Body mass index and risk of diabetes mellitus in the Asia-Pacific region. Asia Pac J Clin Nutr 15: 127–133.
  31. 31. Jacobs P, Fassbender K (1998) The measurement of indirect costs in the health economics evaluation literature. A review. Int J Technol Assess Health Care 14: 799–808.
  32. 32. Garber AM, Phelps CE (2008) Future costs and the future of cost-effectiveness analysis. J Health Econ 27: 819–821.
  33. 33. Meltzer D (2008) Response to “Future costs and the future of cost-effectiveness analysis”. J Health Econ 27: 822–825.
  34. 34. van Baal PHM, Polder JJ, de Wit GA, Hoogenveen RT, Feenstra TL, et al. (2008) Lifetime Medical Costs of Obesity: Prevention No Cure for Increasing Health Expenditure. PLoS Med 5: e29.
  35. 35. AIHW (2008) Health Expenditure Australia. Canberra: Australian Institute of Health and Welfare.
  36. 36. ABS (2003) ABS Consumer Price Index. Canberra: Australian Bureau of Statistics.
  37. 37. Ezzati M, Lopez AD, Rogers A, Murray CJL (2003) Comparative Quantification of Health Risks. Geneva: World Health Organisation.
  38. 38. Lipscomb J (1989) Time preference for health in cost-effectiveness analysis. Medical Care 27: S233–252.
  39. 39. George B, Harris A, Mitchell A (2001) Cost-effectiveness analysis and the consistency of decision making: evidence from pharmaceutical reimbursement in australia (1991 to 1996). Pharmacoeconomics 19: 1103–1109.
  40. 40. van Baal PH, van den Berg M, Hoogenveen RT, Vijgen SM, Engelfriet PM (2008) Cost-effectiveness of a low-calorie diet and orlistat for obese persons: modeling long-term health gains through prevention of obesity-related chronic diseases. Value Health 11: 1033–1040.
  41. 41. Anderson JW, Konz EC, Frederich RC, Wood CL (2001) Long-term weight-loss maintenance: a meta-analysis of US studies. Am J Clin Nutr 74: 579–584.
  42. 42. Foxcroft DR (2005) Orlistat for the treatment of obesity: cost utility model. Obes Rev 6: 323–328.
  43. 43. Hertzman P (2005) The cost effectiveness of orlistat in a 1-year weight-management programme for treating overweight and obese patients in Sweden: a treatment responder approach. Pharmacoeconomics 23: 1007–1020.
  44. 44. Lacey LA, Wolf A, O'Shea D, Erny S, Ruof J (2005) Cost-effectiveness of orlistat for the treatment of overweight and obese patients in Ireland. Int J Obes (Lond) 29: 975–982.
  45. 45. Ruof J, Golay A, Berne C, Collin C, Lentz J, et al. (2005) Orlistat in responding obese type 2 diabetic patients: meta-analysis findings and cost-effectiveness as rationales for reimbursement in Sweden and Switzerland. Int J Obes (Lond) 29: 517–523.
  46. 46. Hampp C, Hartzema AG, Kauf TL (2008) Cost-utility analysis of rimonabant in the treatment of obesity. Value Health 11: 389–399.
  47. 47. Hakim Z, Wolf A, Garrison LP (2002) Estimating the Effect of Changes in Body Mass Index on Health State Preferences. Pharmacoeconomics 20: 393–404.
  48. 48. James WP, Caterson ID, Coutinho W, Finer N, Van Gaal LF, et al. (2010) Effect of sibutramine on cardiovascular outcomes in overweight and obese subjects. N Engl J Med 363: 905–917.
  49. 49. Curfman GD, Morrissey S, Drazen JM (2010) Sibutramine–another flawed diet pill. N Engl J Med 363: 972–974.
  50. 50. Dansinger ML, Tatsioni A, Wong JB, Mei C, Balk EM (2007) Meta-analysis: The Effect of Dietary Counseling for Weight Loss. Annals of Internal Medicine 147: 41–W47.
  51. 51. Beilby J, Furler J (2005) General practitioner services in Australia. general practice in australia: 2004. Canberra: Commonwealth of Australia.
  52. 52. Padwal R, Kezouh A, Levine M, Etminan M (2007) Long-term persistence with orlistat and sibutramine in a population-based cohort. Int J Obes (Lond) 31: 1567–1570.
  53. 53. Carter R, Rankin B (2006) Economic assessment of the NHFA proposal for a national CHD initiative in general practice. Report submitted to the National Heart Foundation of Australia and its CHD-GP Working Group. Program Evaluation Unit, School of Population Health, The University of Melbourne.
  54. 54. ABS (2003) Labour Force Australia 2003. Canberra: Australian Bureau of Statistics.
  55. 55. ABS (2003) Average Weekly Earnings. Canberra: Australian Bureau of Statistics.
  56. 56. Rankin S, Hughes-Anderson W, House J, Aitken J, Heath D, et al. (2002) Rural residents' utilisation of health and visiting specialist health services. Rural and Remote Health 2: 119.
  57. 57. Bamford E, Dunne L, Taylor D, Symon B, Hugo G, et al. (1999) Accessibility to general practitioners in rural South Australia. A case study using geographic information system technology. Medical Journal of Australia 171: 614–616.
  58. 58. Organisation for Economic Cooperation and Development (2006) Purchasing Power Parities (PPPs) for OECD Countries 1980–2006.
  59. 59. Stinnett AA, Paltiel AD (1997) Estimating CE ratios under second-order uncertainty: the mean ratio versus the ratio of means. Med Decis Making 17: 483–489.