Cost-effectiveness analysis of bariatric surgery for morbid obesity in Belgium

Abstract

Aims: This study presents the cost-effectiveness analysis of bariatric surgery in Belgium from a third-party payer perspective for a lifetime and 10-year horizon.

Materials and methods: A decision analytic model incorporating Markov process was developed to compare the cost-effectiveness of gastric bypass, sleeve gastrectomy, and adjustable gastric banding against conventional medical management (CMM). In the model, patients could undergo surgery, or experience post-surgery complications, type 2 diabetes, cardiovascular diseases, or die. Transition probabilities, costs, and utilities were derived from the literature. The impact of different surgical methods on body mass index (BMI) level in the base-case analysis was informed by the Scandinavian Obesity Surgery Registry and the Swedish Obese Subject (SOS) study. Healthcare resource use and costs were obtained from Belgian sources. A base-case analysis was performed for the population, the characteristics of which were obtained from surgery candidates in Belgium.

Results: In the base-case analysis over a 10-year time horizon, the increment in quality-adjusted life-years (QALYs) gained from bariatric surgery vs CMM was 1.4 per patient, whereas the incremental cost was €3,788, leading to an incremental cost-effectiveness ratio (ICER) of €2,809 per QALY. Over a lifetime, bariatric surgery produced savings of €9,332, an additional 1.1 life years and 5.0 QALYs. Bariatric surgery was cost-effective at 10 years post-surgery and dominant over conventional management over a lifetime horizon.

Limitations: The model did not include the whole scope of obesity-related complications, and also did not account for variation in surgery outcomes for different populations of diabetic patients. Also, the data about management of patients after surgery was based on assumptions and the opinion of a clinical expert.

Conclusions: It was demonstrated that a current mix of bariatric surgery methods was cost-effective at 10 years post-surgery and cost-saving over the lifetime of the Belgian patient cohort considered in this analysis.

Keywords:

• Bariatric surgery
• obesity
• cost-effectiveness analysis
• Belgium

Introduction

Obesity is a severe condition of global concern and a risk factor for diabetes, cancer, cardiovascular, musculoskeletal, and gynecological disorders^1–9. When conventional methods to manage obesity fail, bariatric surgery can become an effective method of weight-loss and may prevent development of common obesity-related comorbidities¹⁰.

In these conditions, evaluation of the economic and clinical impact of bariatric surgery is essential for appropriate resource allocation and informed decision-making. Critical areas for evaluation include optimal criteria for commissioning of bariatric surgery and the cost-effectiveness of a current mix of surgical techniques in Belgium for the entire population of surgical candidates and specific cohorts of patients, which varied in levels of obesity and diabetes status.

The objective of this study was to evaluate the cost-effectiveness of bariatric surgery in Belgium from a third-party payer perspective in the mid (10-years) and long-term (lifetime).

Materials and methods

Overview of the model

A state-transition Markov model was constructed to evaluate the costs and outcomes of bariatric surgery compared with conventional medical management (CMM)¹¹. Full details of the modeling approach, data input, and validation activities are reported elsewhere¹². In brief; morbidly obese patients could undergo surgery or continue with CMM, experience post-surgery complications or have no complications, develop T2D (type 2 diabetes) or cardiovascular obesity-related comorbidities (angina, myocardial infarction, stroke, heart failure, or peripheral artery disease), have T2D remission, or die (Figure 1). In each model cycle, patients could either move between health states or stay in the previous state. The cycle length in the model was set at 1 month. Monthly transition probabilities between the model states were derived from the literature or calculated using a log-normal transformation^13–25. The cost-effectiveness was evaluated over 10-year and lifetime horizons.

Figure 1. Structure of the model. Abbreviation. MI, myocardial infarction. Reprinted with permission of Borisenko Oleg from Borisenko et al. Bariatric surgery can lead to net cost savings to health care systems: results from a comprehensive European decision analytic model. Obes Surg 2015;25:1559–68¹².

It was assumed in the model that the risk of developing obesity-related comorbidities depends on the patient’s risk profile, which includes age, gender, smoking status, BMI, the level of systolic blood pressure (SBP), and the presence of T2D. Another assumption was made that bariatric surgery can reduce the risk of obesity-associated comorbidities and mortality by decreasing BMI and SBP and reducing the risk of developing T2D^26–28. By modeling the risk in the operated patient cohort and a hypothetical cohort of patients, who continue with a CMM and do not undergo surgical weight reduction, it is possible to quantify the effect of the surgery on the probability of long-term obesity-related complications.

The model was constructed using Microsoft Excel 2016 (Microsoft Corp., Redmond, WA).

Clinical input

Risk equation from the Framingham Heart Study was used to calculate the 10-year risks of cardiovascular events, which were re-stratified into monthly probabilities^13,29,30. Diabetes incidence was BMI-related and was determined by polynomial regression, as reported by Picot et al.³¹, and based on estimates from Colditz et al.³². Remission of diabetes was informed by 5-year data from a randomized control trial, in which outcomes in T2D patients undergoing medical weight loss therapy were compared to outcomes in patients receiving surgical treatment³³. The risk of short-term (30-day) mortality and severe post-surgery adverse events in the base-case analysis was extracted from a Belgian source³⁴. A Scandinavian Obesity Surgery Registry (SOReg) report provided information about the 2-year risk of post-surgery complications (cholecystectomy, gastric stricture, abdominal hernia repair, gastric ulcer, leakage, and abscess)³⁵. The model accounted for possible conversion and revisional procedures, which follow the index bariatric surgery in the case of its failure or the occurrence of serious complications. The rate of conversion/revisional procedures for sleeve gastrectomy was obtained from a randomized controlled study of bariatric surgery in T2D patients³³. The conversion/revisional probability for gastric bypass was conservatively assumed to be equal to the probability for sleeve gastrectomy. The latter assumption was verified by a clinical expert. Data on conversion/revisional rates for adjustable gastric banding was extracted from a Belgian study of hospital records with 9 years of follow-up³⁶.

Normal mortality was acquired from gender-specific Belgian life tables; mortality due to ischemic heart disease (ICD-10 codes I20-I25) was subtracted from all-cause mortality to calculate normal non-ischemic heart disease mortality^37,38. The presence of one of the cardiovascular events or diabetes increases the risk of developing associated disorders (e.g. patients with heart failure are more likely to have a stroke) and mortality, which was informed by a number of epidemiological studies^14–25.

Three surgical techniques commonly received by bariatric patients in Belgium were considered in the analysis: gastric bypass (GBP), sleeve gastrectomy (SG), and adjustable gastric banding (AGB). The SOReg report provided information about the impact of different surgical methods on the BMI level in the base-case analysis³⁵. The impact of surgery on BMI level was extrapolated using data from the SOS study from the time-point of the last follow-up (2 years)³⁹. After 15 years, the BMI level was assumed to be permanent until death. Any change in BMI in the conventional medical management arm was based on the change in BMI of the conservative management arm of the SOS study³⁹. The change in SBP was derived from the SOS study for non-diabetics⁴⁰ and from a study by Ikramuddin et al.⁴¹ for diabetic patients. A number of individual studies were used for analysis in different individual cohorts of patients^41–50. Details of the methodology and data sources have been reported elsewhere¹². Data on the clinical efficacy of bariatric surgery is presented in Table 1.

Table 1. Absolute BMI reduction.

Parameter	Value	Range for one-way sensitivity analysis	Distribution for PSA (SD)	Source
GBP, 1-year, males	12.7	8.7–37.7	Normal (2.2)	SOReg^35 2011
GBP, 2-year, males	12.6	8.6–37.4	Normal (2.2)
SG, 1-year, males	9.7	5.9–25.5	Normal (1.7)
SG, 2-year, males	9.4	5.7–24.7	Normal (1.6)
AGB, 1-year, males	5.6	3.9–16.9	Normal (1.0)
AGB, 2-year, males	6.9	4.8–20.9	Normal (1.2)
GBP, 1-year, females	13.5	9.5–41.1	Normal (2.4)
GBP, 2-year, females	13.5	9.5–41.2	Normal (2.4)
SG, 1-year, females	12.5	7.5–32.8	Normal (2.2)
SG, 2-year, females	14.7	8.9–38.0	Normal (2.6)
AGB, 1-year, females	5.5	3.9–17.0	Normal (0.9)
AGB, 2-year, females	5.1	3.6–15.7	Normal (0.9)

Abbreviations. PSA, probabilistic sensitivity analysis; GBP, gastric bypass; SG, sleeve gastrectomy; AGB, adjustable gastric banding.

Resource utilization and cost data

Contrary to conducting the analysis from the perspective of each surgery method, the modeled cost of bariatric surgery was driven by the proportional case-mix of all three surgical types, which allowed for capturing the actual cost of bariatric surgery in Belgium. The number of surgical procedures, as well as the distribution of different surgery types (GBP =75%; SG =20%; AGB =5%), was obtained from a clinical expert. The cost of the bariatric surgery procedure was informed by Belgian DRG data from 2010, updated for 2012 (APR-DRG 403, weighted average value for minor to moderate level of severity for GBP, SG, and AGB). Inpatient care cases with severe complications were assigned an uplifted DRG tariff (APR-DRG 403, weighted average value for high to the extreme level of severity for GBP, SG, and AGB). The costs of bariatric surgery procedures and post-surgery complications are presented in Table 2.

Table 2. Cost of bariatric surgery procedures and post-surgery complications (€2,012).

Parameter	Value	Range for one–way sensitivity analysis	Source
Cost of GB without complications	4,433	3,546–5,320	Weighted average value for minor to moderate level of severity APR-DRG 403
Cost of GB with complications	14,577	11,662–17,492	Weighted average value for high to extreme level of severity APR-DRG 403
Cost of SG without complications	4,433	3,546–5,320	Weighted average value for minor to moderate level of severity APR-DRG 403
Cost of SG with complications	14,577	11,662–17,492	Weighted average value for high to extreme level of severity APR-DRG 403
Cost of AGB without complications	4,433	3,546–5,320	Weighted average value for minor to moderate level of severity APR-DRG 403
Cost of AGB with complications	14,577	11,662–17,492	Weighted average value for high to extreme level of severity APR-DRG 403
Cost of removal/revision of AGB without complications	4,433	3,546–5,320	Weighted average value for minor to moderate level of severity APR-DRG 403
Cost of abdominal hernia procedure	2,818	2,254–3,382	APR-DRG 227 value for 2010, updated to 2012
Cost of cholecystectomy	3,103	2,482–3,724	APR-DRG 263 value for 2010, updated to 2012
Cost of leakage and abscess	9,183	7,346–11,020	Weighted average for APR-DRG 222 and APR-DRG 220 value for 2010, updated to 2012
Cost of obstruction	4,211	3,369–5,053	APR-DRG 247 value for 2010, updated to 2012
Cost of stricture	4,211	3,369–5,053	APR-DRG 247 value for 2010, updated to 2012
Cost of gastric ulcer	22	17.6–26.4	An 8-week course of 40 mg omeprazole. Cost was obtained from INAMI^55

Abbreviations. PSA, probabilistic sensitivity analysis; GBP, gastric bypass; SG, sleeve gastrectomy; AGB, adjustable gastric banding; APR-DRG, All Patient Refined Diagnosis-Related Groups; INAMI, Institut National d'Assurance Maladie-Invalidité (National Institute for Health and Disability Insurance).

The costs of the obesity-related comorbidities were derived from the Belgian literature^51–56.

Pre- and post-surgery care utilization was informed by expert opinion. It was assumed that pre-operatively a patient would have two consultations with a surgeon, and one visit to an endocrinologist. After surgery, two visits to a surgeon and four visits to an endocrinologist were considered during the first month. For the second and third year of follow-up, two and one annual consultations by a surgeon were considered, accordingly. It was assumed that surgery candidates who do not receive surgery require two annuals visit to a general practitioner (GP). Unit costs for outpatient visits were obtained from the National INAMI (Institut National d’Assurance Maladie-Invalidité) Nomenclature of procedures and Belgian Health Care Knowledge Centre (KCE)^55,56.

Table 3 summarizes costs of obesity-related comorbidities and unit costs for specialist visits.

Table 3. Cost of obesity-related comorbidities and unit costs (€2,012).

Parameter	Value	Range for one-way sensitivity analysis	Distribution for PSA	Source
Cost of T2D	4,418	3,534–5,302	Gamma (α = 100; λ = 44.18)	Jönsson et al.^51 (2002)
Cost of acute stroke	9,248	7,398–11,098	Gamma (α = 100; λ = 92.48)	Annemans et al.^52 (2003)
Cost of post-stroke 1 year	11,253	9,002–13,504	Gamma (α = 100; λ = 112.53)	Annemans et al.^52 (2003)
Cost of post-stroke 2 year	4,454	3,563–5,345	Gamma (α = 100; λ = 44.54)	Annemans et al.^52 (2003), extrapolation based on first 6 months costs
Cost of acute TIA	4,359	3,487–5,231	Gamma (α = 100; λ = 43.59)	APR-DRG 047 value for 2010, updated to 2012
Cost of acute AMI	7,756	6,205–9,307	Gamma (α = 100; λ = 77.56)	Annemans et al.^52 (2003)
Cost of post-MI state	4,842	3,874–5,810	Gamma (α = 100; λ = 48.42)	Annemans et al.^52 (2003)
Cost of HF	2,194	1,755–2,633	Gamma (α = 100; λ = 21.94)	Annemans et al.^53 (1998)
Cost of angina	5,518	4,414–6,622	Gamma (α = 100; λ = 55.18)	Annemans et al.^53 (1998)
Cost of PAD	4,363	3,490–5,236	Gamma (α = 100; λ = 43.63)	Wilson et al.^54 (2012)
Cost of GP visit	15	12–18	—	Nomenclature, INAMI^55
Cost of nurse visit	18	14.4–21.6	—	KCE report, updated to 2012^56
Cost of surgeon visit	31	24.8–37.2	—	Nomenclature, INAMI^55
Cost of endocrinologist visit	56	44.8–67.2	—	Nomenclature, INAMI^55

Abbreviations. PSA, probabilistic sensitivity analysis; T2D, type 2 diabetes; TIA, transient ischemic attack; AMI, acute myocardial infarction; MI, myocardial infarction; HF, heart failure; PAD, peripheral arterial disease; GP, general practitioner; INAMI, Institut National d’Assurance Maladie-Invalidité (National Institute for Health and Disability Insurance); KCE, Federaal Kenniscentrum voor de Gezondheidszorg (The Belgian Health Care Knowledge Centre).

Only direct medical costs were included in the analysis. All costs from prior years were adjusted to 2012-year Euros using the Belgian Consumer Price Index⁵⁷.

Utility data

Health-related quality-of-life was expressed by the generic HRQoL instrument; EuroQol-5D (EQ-5D). Health-related quality-of-life was dependent on BMI and the presence of diabetes⁵⁸. Also, quality-of-life effects of obesity-related comorbidities were incorporated in the model. They were informed by data from UK adaptations of the US Medical Expenditure Panel Survey EQ-5D mapping study⁵⁹.

Analysis

The analysis was performed from a third-party payer perspective over a 10-year and a lifetime horizon. All costs beyond the first year were discounted by 3.0% annually, all health outcomes – by 1.5%, according to the Belgian recommendations⁶⁰. Surgery was considered cost-effective if the incremental cost-effectiveness ratio (ICER), calculated by dividing the difference in cumulative costs between the two arms by the difference in QALYs gained, was below the willingness-to-pay threshold of €30;000/QALY⁶¹.

Base-case analysis

First, the base-case analysis considered the hypothetical cohort of candidates for bariatric surgery in Belgium, the characteristics of which were extracted from the large single-center Belgian study, the SOS study, and Factbook 2011 of the Organization for Economic Co-operation and Development (OECD)^34,40,62. This analysis was performed for a cohort of patients whose average age was 39 years; 21% of whom were males; with a mean BMI of 41 kg/m², a mean SBP of 140 mm Hg, the prevalence of T2D of 9%, and smoking prevalence of 21%. Patient baseline characteristics are presented in Table 4.

Table 4. Patient baseline characteristics.

Parameter	Value	Range for one-way sensitivity analysis	Distribution for PSA	Source
Age, years	39.2	21–56	Normal (SD =4.5)	Dillemans et al.^34 (2009)
Gender, males (%)	20.69	NA	Beta (α = 1,035; β = 3,965)	Dillemans et al.^34 (2009)
Body mass index, kg/m^2	41.44	23–58	Normal (SD =4.5)	Dillemans et al.^34 (2009)
Diabetes mellitus (%)	9.17	NA	Beta (α = 459; β = 4,542)	Dillemans et al.^34 (2009)
Systolic blood pressure, mmHg	140.1	125–200	Gamma (α = 55.53; λ = 2.52)	Sjöström^40 (2004)
Smoking (%)	20.5	NA	Beta (α = 1,025; β = 3,975)	OECD Factbook^62 (2011)

Abbreviations. PSA, probabilistic sensitivity analysis; OECD, Organization for Economic Co-operation and Development.

Second, the cost-effectiveness of bariatric surgery was calculated for 16 cohorts of 39.2-year old non-smoking males and females with moderate (start BMI – 33 kg/m²); severe (start BMI – 37 kg/m²); morbid (start BMI – 42 kg/m²) and super obesity (start BMI – 52 kg/m²) with and without the presence of T2D.

One-way sensitivity analysis

To determine key model drivers and examine how the output changed when input parameters were altered; a one-way sensitivity analysis was performed over a lifetime horizon. Parameters were varied one at a time, within a pre-determined minimum and maximum range, while the remaining parameters were unaltered. The analysis was performed for a single cohort of 39.2-year-old males with a BMI of 41 kg/m², with non-smoking and diabetes-free status.

Specific conditions were applied to the binary input parameters (gender, smoking, and diabetes status). For the “gender” parameter, “male gender” was considered as maximum input (value of 1), “female gender” as minimum input (value of 0). For diabetes and smoking, their presence was considered as maximum input (value of 1), their absence as minimum input (value of 0).

Probabilistic sensitivity analysis

A probabilistic sensitivity analysis (PSA) was undertaken by varying all input parameters simultaneously across distributions to evaluate multivariate and stochastic uncertainties in the model. Key input parameters that were considered to be random variables were baseline patient characteristics, costs, utility decrements, probabilities, and relative risks. The choice of distributions was guided by the nature of the underlying parameter¹³. Gamma distribution was used for cost variables available only as single-point estimates. A beta distribution was assigned to the probabilities, utility decrements, while relative risks were assumed to follow a log-normal distribution. Distribution parameters were dependent on the nature of inputs, with beta distribution being used for the probabilities, utilities, and decrements, and a log-normal distribution was used for relative risks. Costs were assigned with a gamma distribution, assuming 10% standard deviation around the mean values. A normal distribution was assigned to the patient age and BMI, while SBP was assumed to follow a gamma distribution. For parameters based on expert or analyst assumption, uniform distributions were used. Reimbursement tariffs were not tested in probabilistic analysis.

During the PSA, 5,000 iterations of the model were run over a lifetime horizon. The result was presented in the form of a scatterplot of the incremental cost-effectiveness of surgery mix compared with usual care.

Model validation

The model underwent a previously reported extensive three-step validation process¹².

Results

Base-case analysis

In the base-case analysis at a 10-year horizon, bariatric surgery generated an additional 1.3 QALYs and the extra cost of €3,788 compared to conventional medical management, resulting in an incremental cost-effectiveness ratio of €2,809/QALY. In the base-case analysis over a lifetime horizon, bariatric surgery produced cost savings of €9,332, and generated an additional 1.1 years of life and 5.0 QALY. Owing to better clinical outcomes, as well as to lower treatment costs, bariatric surgery was found to be a dominant strategy (more effective and less costly) over conventional medical management over a lifetime horizon (Table 5).

Table 5. Base-case results of cost-effectiveness analysis.

	Cost, €	Δ cost, €	LYG, years	Δ LYG	QALY	Δ QALY	ICER, €/QALY
Ten years
CMM arm	6,599	3,788	9.2	0.0	4.5	1.3	2,809
Surgical arm	10,387		9.2		5.8
Lifetime
CMM arm	36,284	–9,332	27.9	1.1	12.9	5.0	Dominates
Surgical arm	26,953		29.0		17.9

Abbreviations. ICER, incremental cost-effectiveness ratio; LYG, life years gained; CMM, conventional medical management; QALY, quality-adjusted life years.

Also, it was observed that surgery had the potential to reduce the risk of obesity-associated cardiovascular comorbidities, for both 10-year and lifetime horizons (Table 6).

Table 6. Absolute and relative risks of obesity-related comorbidities in the model.

	Angina	MI total non-fatal	MI fatal	Stroke total non-fatal	Stroke fatal	TIA	HF	PAD	Diabetes
Ten years
Absolute risk in the surgical arm	0.02	0.02	0.0004	0.02	0.003	0.002	0.02	0.01	0.04
Absolute risk in the CMM arm	0.02	0.04	0.0006	0.03	0.004	0.003	0.02	0.02	0.15
Relative risk	0.71	0.71	0.66	0.71	0.71	0.71	0.71	0.71	0.27
Lifetime
Absolute risk in the surgical arm	0.11	0.19	0.02	0.17	0.03	0.01	0.13	0.09	0.13
Absolute risk in the CMM arm	0.13	0.24	0.02	0.21	0.04	0.02	0.17	0.11	0.39
Relative risk	0.81	0.78	0.70	0.78	0.78	0.82	0.79	0.82	0.34

Abbreviations. HF, heart failure; MI, myocardial infarction; PAD, peripheral artery disease; TIA, transient ischemic attack; CMM, conventional medical management.

Analysis of patient cohorts

The analysis shows that, over a 10-year horizon, bariatric surgery was cost saving in all the eight diabetic cohorts considered (moderately, severely, morbidly, and super obese males and females). In the non-diabetic cohorts, surgery was cost-effective in all groups; moderately obese males (ICER €6,399/QALY) and females (ICER €6,802/QALY), severely obese males (ICER €5,062/QALY) and females (ICER €5,356/QALY), morbidly obese males (ICER €2,114) and females (ICER €2,288); and super obese males (ICER €636/QALY) and females (ICER €652/QALY). The cost-effectiveness of surgery increased with the increase of baseline BMI of the cohort.

Over a lifetime horizon, it was demonstrated that bariatric surgery was dominating over conventional medical management in all the eight diabetic and non-diabetic cohorts considered.

Sensitivity analysis

Deterministic one-way sensitivity analysis showed that the model was most sensitive to age (surgery is less cost-effective in older patients), baseline BMI (surgery becomes more cost-effective with increase in BMI), effectiveness of GBP at 2 years in males (surgery is less cost-effective with lower efficacy), presence of T2D (surgery is more cost-effective in patients with diabetes), and gender (surgery is more cost-effective in women) (Figure 2).

Figure 2. Tornado diagram. One-way sensitivity analysis was produced for a single cohort of 39.2-year non-smoking males with BMI of 41 kg/m². Only variables affecting the results for more than €100 are shown.

Under the probabilistic sensitivity analysis conditions, the model demonstrated that bariatric surgery produced clinical benefits (defined as additional QALYs) in 100% of the patients, and had a cost-saving effect in 99% of cases, while in the remaining 1% it was cost-effective (Figure 3).

Figure 3. Cost-effectiveness acceptability plane. A scatter plot depicting the ICER derived from each iteration of the model from the probabilistic sensitivity analysis.

Discussion

This analysis of currently used surgical techniques in Belgium indicates that bariatric surgery is cost-effective in the mid-term and can save the healthcare system money over a longer time horizon. Results are consistent with the findings of recent Belgian analysis⁶³, in which a one unit of BMI decrease over a 20-year horizon was associated with improved health outcomes (up to 0.14 QALY) and cost savings (up to €1,039 and €864 per female and male patient, correspondingly). However, global evidence suggests different estimates of the cost-effectiveness of bariatric surgery, depending on the country-specific context, severity of obesity status, and type of surgical approach. The results indicate that bariatric surgery can be either a cost-saving or cost-effective alternative to conventional methods for managing obesity^{31,58,64–74}. In Germany, bariatric surgery (GBP and SG) was found to produce extra costs of €921 and yield less clinical benefits (2.55 QALY) within a 20-years horizon⁶⁶. On the contrary, in the French setting bariatric surgery was associated with a cost-saving effect over a 10-year horizon and incremental QALY gains of 1.34 and 1.03 for GBP and AGB, correspondingly⁵⁸.

The demonstrated lifetime economic benefits could translate into potential cost reduction of ∼ €98M over a lifetime period when 10,500 patients who annually undergo bariatric surgery in Belgium are considered. Likewise, clinical benefits could add up to ∼11,293 person-years and 51,482 quality-adjusted person-years for the lifetime of the operated cohort.

Nevertheless, these results have originated from a decision analytic modeling, which has inherent limitations of being a simplification of the current clinical practice. Several empirical studies, performed mainly in Sweden and the US, failed to demonstrate a cost-saving effect with the use of bariatric surgery. However, they shared similar limitations, including a short time horizon, reporting data for discontinued surgical methods (e.g. vertical banded gastroplasty in the SOS study), and extensive use of the open surgical approach^75,76. Due to delays in publication and the natural development of surgical methods, empirical economic evidence may not reflect current safety and efficacy of the bariatric procedures. Decision analytic modeling can complement empirical data and bridge knowledge gaps promptly.

The results of this study revealed that bariatric surgery is cost-effective over mid-term, cost-saving over a lifetime, and is also associated with substantial clinical benefits. A recent study on the utilization of bariatric surgery in seven European countries showed that Belgium has a favorable clinical and political opinion on bariatric surgery with a high utilization level (928 procedures per 1 million (M) population), compared with Europe’s average (401 procedures per 1 M population)⁷⁷.

The favorable cost-benefit effect of bariatric surgery implies that the policy-makers might have additional evidence for further prioritizing bariatric surgery in Belgium and diverting the appropriate resources on the provision of bariatric surgery to those who are most likely to benefit. The ability of bariatric surgery to recoup the upfront costs by future savings might be especially pronounced in high-risk patients, e.g. those with extreme BMI or in diabetic patients, in which savings from prevented CVD comorbidities accumulate with the reduced costs of diabetic complications (such as “diabetic foot”).

Like most modeling analyses, this study has a number of limitations, as previously discussed¹². In brief, the model did not include the whole scope of obesity-related complications, which could lead to under-estimating cost benefits from the surgery. The model used here did not account for variation in surgery outcomes for different populations of diabetic patients, which could have an impact on the overall clinical effectiveness of the therapy. The rates of post-surgery BMI reduction were extracted from the SOReg report, which was the most comprehensive source at the time of analysis; however, they might not be representative of the most recent developments in clinical practices and patient outcomes. Finally, the data about management of patients after surgery or surgical candidates who do not receive surgery was based on assumptions and opinion of a clinical expert.

Conclusion

In a comprehensive decision analytic model, the current mix of surgical methods for bariatric surgery available in Belgium was cost-effective at 10 years and cost-saving over the lifetime of the operated cohort.

Transparency

Declaration of funding

This study was funded by Covidien, Inc. (now Medtronic, Inc.).

Declaration of financial/other interests

OB and VL are employees of Synergus AB, which received consultancy fees to conduct the research from Covidien, Inc. (now Medtronic, Inc.). Covidien, Inc. did not have any influence on the interpretation of data as well as the final conclusions drawn. ID and BD report no other conflicts of interest in this work. JME peer reviewers on this manuscript have no relevant financial or other relationships to disclose.

Acknowledgments

No assistance in the preparation of this article is to be declared.