Comparison of Comorbidity Treatment and Costs Associated With Bariatric Surgery Among Adults With Obesity in Canada

Jason A Davis; Rhodri Saunders

doi:10.1001/jamanetworkopen.2019.19545

. 2020 Jan 17;3(1):e1919545. doi: 10.1001/jamanetworkopen.2019.19545

Comparison of Comorbidity Treatment and Costs Associated With Bariatric Surgery Among Adults With Obesity in Canada

Jason A Davis ^1,^✉, Rhodri Saunders ¹

¹Coreva Scientific GmbH & Co KG, Koenigswinter, Germany

Accepted for Publication: November 22, 2019.

Published: January 17, 2020. doi:10.1001/jamanetworkopen.2019.19545

^✉

Corresponding Author: Jason A. Davis, DPhil, Coreva Scientific GmbH & Co KG, Im Muehlenbruch 1, Third Floor, 53639 Koenigswinter, Germany (jd@coreva-scientific.com).

Author Contributions: Drs Davis and Saunders had full access to all of the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.

Concept and design: Both authors.

Acquisition, analysis, or interpretation of data: Both authors.

Drafting of the manuscript: Davis.

Critical revision of the manuscript for important intellectual content: Both authors.

Statistical analysis: Davis.

Obtained funding: Saunders.

Administrative, technical, or material support: Davis.

Supervision: Saunders.

Conflict of Interest Disclosures: Dr Davis reported serving as a consultant for Medtronic Inc during the conduct of the study and outside the submitted work. Dr Saunders reported serving as a consultant for Medtronic during the conduct of the study and outside the submitted work; and performing health-economic model development and burden of disease analysis for Cardinal Health and receiving consulting fees from Medasense Biometrics and Da Vinci Health group outside the submitted work.

Funding/Support: The study was funded by Medtronic Inc.

Role of the Funder/Sponsor: Medtronic Inc had no role in the design and conduct of the study; or collection, management, analysis, and interpretation of the data. Medtronic was permitted to review the manuscript before submission but did not have editorial control over the final manuscript, including the preparation or approval of the manuscript and decision to submit the manuscript for publication.

^✉

Corresponding author.

PMCID: PMC6991282 PMID: 31951277

Key Points

Question

How are surgical delays to sleeve gastrectomy and postsurgical weight trajectories associated with health costs?

Findings

In this economic evaluation of a hypothetical 100-patient cohort, a model with shorter time to surgery and less postsurgical weight gain was associated with reduced patient-years of treatment for diabetes, hypertension, and dyslipidemia and reduced 10-year costs.

Meaning

These findings suggest that improving presurgical and postsurgical care could realize considerable savings in patient-years of obesity-related comorbidities and costs that may provide impetus to invest in care pathway improvement.

This economic evaluation estimates the difference in patient-years of treatment for diabetes, hypertension, and dyslipidemia and public-payer cost between the Canadian standard of care and an improved bariatric surgery care pathway.

Abstract

Importance

Information on the associations between barriers to delivery of bariatric surgery and poor weight trajectory afterward is lacking. Estimates are needed to inform decisions by administrators and clinicians to improve care.

Objective

To estimate the difference in patient-years of treatment for diabetes, hypertension, and dyslipidemia and public-payer cost between the Canadian standard and an improved bariatric surgery care pathway.

Design, Setting, and Participants

Economic evaluation of a decision analytic model comparing the outcomes of the standard care in Canada with an improved bariatric care pathway with earlier sleeve gastrectomy delivery and better postsurgical weight trajectory. The model was informed by published clinical data (101 studies) and meta-analyses (11 studies) between January and May 2019. Participants were a hypothetical 100-patient cohort with demographic characteristics derived from a Canadian study.

Interventions

Reduction of Canadian mean bariatric surgery wait time by 2.5 years following referral and improvement of patient postsurgery weight trajectory to levels observed in other countries.

Main Outcomes and Measures

Modeling weight trajectory after sleeve gastrectomy and resolution rates for comorbidities in Canada in comparison with an improved care pathway to estimate differences in patient-years of comorbidity treatment over 10 years following referral and the associated costs.

Results

For the 100-patient cohort (mean [SD] 88.2% [1.4%] female; mean [SD] age, 43.6 [9.2] years; mean [SD] body mass index, 49.4 [8.2]; and mean [SD] comorbidity prevalence of 50.0% [4.1%], 66.0% [3.9%], and 59.3% [4.0%] for diabetes, hypertension, and dyslipidemia, respectively) over 10 years following referral, the improved vs standard care pathway was associated with median reduction in patient-years of treatment of 324 (95% credibility interval [CrI], 249-396) for diabetes, 245 (95% CrI, 163-356) for hypertension, and 255 (95% CrI, 169-352) for dyslipidemia, corresponding to total savings of $900 000 (95% CrI, $630 000 to $1.2 million) for public payers in the base case. Relative to standard of care, the associated reduction in costs was approximately 29% (95% CrI, 20%-42%) in the improved pathway. Sensitivity analyses demonstrated independent associations of earlier surgical delivery and various levels of postsurgical weight trajectory improvements with overall savings.

Conclusions and Relevance

This study suggests that health care burden may be decreased through improvements to delivery and management of patients undergoing sleeve gastrectomy. More data are needed on long-term patient experience with bariatric surgery in Canada to inform better estimates.

Introduction

Surgical intervention has been shown to be an invaluable tool for long-term health improvement in many patients with obesity and related comorbidities. A growing body of long-term evidence after various bariatric surgical procedures suggests that for suitable patients, surgery is associated with higher rates of persistent weight loss and improvement or resolution of comorbidities such as type 2 diabetes, hypertension, and dyslipidemia compared with nonsurgical treatments.^1,2,3 There remains, however, inconsistency in outcomes among patients undertaking a surgical path in their bariatric care. Some differences are expected, as different types of surgery offer varying degrees of weight loss and morbidity resolution.^2,4,5 Different weight trajectories have been observed after the same type of surgery,^6,7,8 with corresponding differences in comorbidity resolution.^6,7 Barriers to reaching surgery may also exist, including patient perceptions and health care system impediments such as capacity and meeting surgical criteria.^9,10,11,12 Resolving these issues to improve care and long-term patient outcomes requires allocation of potentially limited resources and information regarding the burden of the status quo vs researching improvements to care.

Canada is one jurisdiction facing these challenges in obesity and comorbidity management. Because the country has a publicly funded health care system, bariatric care, including surgery, is available to patients without personal expense, but inefficiencies in care and barriers to access may contribute to underuse of services, particularly surgery.^9,13 Patient perceptions and lengthy wait times across the country are likely contributors to this effect.^12,14,15 There is likely scope to improve provision of care in this setting, as an example of a locale dealing with barriers to achieving what has been shown to be an effective treatment strategy.

Among the types of surgery available in Canada, sleeve gastrectomy (SG) is gaining in prominence.^9,13,16 Some of the earliest Canadian reports of its use were of patients treated just before 2010,^17,18,19 meaning that there is a smaller body of evidence available regarding long-term outcomes compared with procedures with a longer history, such as Roux-en-Y gastric bypass (RYGB). Globally, the evidence for long-term outcomes after SG as a primary procedure is building, with multiple studies presenting positive outcomes of comorbidity improvement and persistent weight loss beyond 5 years following surgery.^20,21 There is a need to examine the application of SG in Canada to determine whether the current standard of care can be improved, drawing on outcomes that have been achieved in other locations for a sense of what may be feasible.

The present study is an economic evaluation of the delivery and management of SG surgery in the Canadian setting via a decision analytic model. It sought first to examine SG outcomes in Canada in the context of global outcomes to contextualize the current standard of care. The care pathway from surgical referral to the postsurgical period was considered, and burden in terms of comorbidity prevalence (type 2 diabetes, hypertension, and dyslipidemia) was quantified in patient-years of treatment and costs. The potential for savings from the perspective of the public payer was assessed by comparing the Canadian standard of care with an improved system, with realistic achievements in earlier time of delivery of surgery following referral, and in better weight trajectory after surgery. Estimates provide a measure of the scope of the current burden and may guide discussions around investments to improve care for patients who undergo bariatric surgery in Canada.

Methods

As an economic evaluation based on aggregate published clinical data, this study did not require ethical review. Prior to study commencement, the data protection officer of Coreva Scientific performed an assessment for risk of personal and/or identifiable data in compliance with the European General Data Protection Regulation. The study was deemed low risk as it relies solely on data in the public domain (published results). It adheres to the Consolidated Health Economic Evaluation Reporting Standards (CHEERS) reporting guideline. Data analysis took place between January and May 2019.

Postreferral Analysis

The primary outcome of the analysis was to determine the difference in comorbidity treatment costs from the perspective of a public payer resulting from introduction of improvements to bariatric surgical care management, specific to SG in the Canadian setting. As such, the study required estimation of comorbidity evolution over the period following referral, before and after surgery. Canadian data sources were used preferentially, supplemented with American data where Canadian data were unavailable. A time horizon of 10 years was chosen given mean waits of 3.5 years across provinces in Canada; this time horizon also served to capture sufficient postsurgical data. The analysis determines overall differences in comorbidity treatment in patient-years and associated costs, with savings defined as the total in the standard care pathway minus the total in the improved pathway.

Presurgical Period

Limited published data report on the postreferral, presurgical period. In the present study, patients were taken to experience mild weight gain of 0.75% total weight per year, determined as a weighted mean among patients from 2 identified studies presenting such data.^22,23 Baseline prevalence of comorbidities was taken from a Canadian surgical cohort,²³ with the remainder considered to be at risk of comorbidity onset. Incidence of comorbidities (type 2 diabetes, hypertension, and dyslipidemia) were estimated from power law modeling of wait-listed patients in the US setting.²⁴ During the wait for surgery, some patients are expected to drop out of the program; data from a Canadian study on wait-listed patients were used for survival analysis with an exponential model to estimate dropout rate as a function of time following referral. Patients who dropped out were assumed to have no comorbidity resolution.

The standard of care time for surgical delivery is taken to be 3.5 years, while the comparator improved pathway of surgery at 1.0 year is assumed a reasonable expectation given past achievement of median 1.2 years for delivery in a program in Ontario, Canada,²⁵ and the upper quartile of longest wait times in a US state of approximately 204 days.²⁶

Postsurgical Period

Previous studies^6,7,8 have noted considerable differences in postsurgical weight trajectory experienced by patients receiving the same surgery. These differences have further been attributed to impact on outcomes.^6,7 The present study characterizes SG trajectory outcomes using cohorts from published studies to better generalize outcome possibilities. Studies were identified from meta-analyses, de novo searches for SG reports in which weight loss was a primary outcome, and any reports of SG outcomes in the Canadian setting (eAppendix 1 in the Supplement). A total of 101 studies were identified with medium- to long-term data. Data from these studies were subjected to group-based trajectory analysis with mixed models to account for correlation of outcomes among study cohorts (see eAppendix 2 in the Supplement for details). Subsequent trajectory comparisons were made between fitted models.

Comorbidity Resolution After SG

In the absence of SG-specific comorbidity resolution data, clinical data were instead drawn from a study reporting detailed outcomes by trajectory for patients undergoing RYGB.⁶ Data from that study were used to model resolution after surgery as a function of time and changing body mass index. To account for the fact that mechanisms of resolution differ between RYGB and SG, a meta-analysis of 11 studies derived from other meta-analyses (eTable 3, eFigure 1, eFigure 2, and eFigure 3 in the Supplement) was performed to estimate the odds ratio of comorbidity resolution between SG and RYGB to adjust the resolution rates observed in the RYGB study for SG, thereby incorporating differences in resolution efficacy and mechanism (eAppendix 3 in the Supplement).

Model Design

Modeling parameters^{4,23,27,28,29,30,31} are presented in Table 1. Baseline demographic characteristics for the 100-patient cohort used in the model were drawn from a Canadian study.²³ The standard of care pathway has patients receiving surgery at 3.5 years after referral and thereafter experiencing the modeled weight trajectory that best represents published Canadian data. The comparator improved care pathway has surgery at 1 year after referral and thereafter patients experiencing the best identified weight trajectory. Comorbidity incidence and overall prevalence are calculated at 6-month intervals and converted into costs using reported annual treatment costs. Values from source literature were inflated to 2018 Canadian dollars using Statistics Canada consumer price index data for health care items.

Table 1. Model Parameters.

Parameter	Source	Base Case, Mean (SD)	Notes
Age, y	Padwal et al,²³ 2014	43.6 (9.2)	Canadian study; we used population demographic data characteristics of the wait-listed population (150 patients)
Body mass index^a	Padwal et al,²³ 2014	49.4 (8.2)
Female, %	Padwal et al,²³ 2014	88.2 (1.4)
Baseline comorbidities, %
Type 2 diabetes	Padwal et al,²³ 2014	50.0 (4.1)
Hypertension	Padwal et al,²³ 2014	66.0 (3.9)
Dyslipidemia	Padwal et al,²³ 2014	59.3 (4.0)
Dropout rate, %	Padwal et al,²³ 2014	Improved: 6.3 (2.5)	Values correspond to base-case surgical times but are changed during sensitivity analyses where improved and standard care pathways have different surgical times
Dropout rate, %	Padwal et al,²³ 2014	Standard: 19.6 (2.7)
Time of surgery, y	NA	Improved: 1.0	Times are postreferral from primary to specialist care
Time of surgery, y	NA	Standard: 3.5	Times are postreferral from primary to specialist care
Cohort size, No.	NA	100	Example cohort
Discount rate, %	CADTH guidelines²⁷	1.5	Fourth-edition guidelines of the CADTH
Cost, CAD
Type 2 diabetes
Year 1^b	Rosella et al,²⁸ 2016	Male: 4061 (609)	Ontario, Canada; base value uncertainty taken as ±15%
Year 1^b	Rosella et al,²⁸ 2016	Female: 4017 (603)	Ontario, Canada; base value uncertainty taken as ±15%
Year 2 onward^b	Rosella et al,²⁸ 2016	Male: 828 (123)	Ontario, Canada; base value average costs years 2-8 in study
Year 2 onward^b	Rosella et al,²⁸ 2016	Female: 1023 (124)
Hypertension^b	Weaver et al,²⁹ 2015	2163 (227)	Canada-wide
Dyslipidemia^b	Conly et al,³² 2011	79 (8)	Alberta, Canada; final value includes only laboratory costs for patients receiving statins minus costs for patient time and travel
Sleeve gastrectomy odds ratio
Type 2 diabetes resolution	Huang et al,³⁰ 2018	0.73 (0.09)	Odds ratios determined from meta-analysis of meta-analyses, expressed as odds of comorbidity resolution after sleeve gastrectomy vs Roux-en-Y gastric bypass; uncertainty taken as one-quarter of the 95% CI^c
	Shoar et al,⁴ 2017
	Li et al,³¹ 2016
Hypertension resolution	Shoar et al,⁴ 2017	0.79 (0.11)
Hypertension resolution	Li et al,³¹ 2016	0.79 (0.11)
Dyslipidemia resolution	Shoar et al,⁴ 2017	0.58 (0.11)
Dyslipidemia resolution	Li et al,³¹ 2016	0.58 (0.11)

Open in a new tab

Abbreviations: CAD, Canadian dollars; CADTH, Canadian Agency for Drugs and Technologies in Health; NA, not applicable.

^{^a}

Calculated as weight in kilograms divided by height in meters squared.

^{^b}

Costs in 2018 CAD, inflated from source data using Statistics Canada consumer price index data for health care items.

^{^c}

Details of meta-analysis comparing sleeve gastrectomy with Roux-en-Y bypass appear in eAppendix 3 in the Supplement.

Sensitivity Analyses

The base-case analysis included improvements to both presurgical care (earlier surgery) and postsurgical outcomes (better weight trajectory). The main sensitivity analysis for the base case saw sampling of the model input parameters using the normal distribution for all but the odds ratios of comorbidity resolution between SG and RYGB, which were sampled using the log-normal distribution. The assumption of presurgical weight gain was tested against scenarios of no change (weight stability) and presurgical weight loss, and decreased rates of developing comorbidities were also tested. Presurgical degree of improvement was tested for sensitivity by sampling combinations of the time of delivery in the improved pathway (6 months to 2.5 years) with time of delivery in the standard care pathway (2.5-4.5 years). Postsurgical improvement was tested between the Canada-attributed trajectory group and the remaining trajectory groups.

Statistical Analysis

Results were determined as medians and 95% credibility intervals (CrIs) from repeated sampling (10 000 replicates for main analysis, 2500 replicates for each combination in sensitivity analyses). Significance may be considered as nonoverlapping of CrIs, or exclusion of 0 (no difference) when considering cost or patient-years of treatment outcomes. For meta-analyses of comorbidity resolution outcomes between SG and RYGB, significance in overall association and intragroup differences was taken to be 2-tailed P < .05. No formal claims are offered regarding significance of observed differences; data are presented with uncertainty ranges to allow readers to interpret in their own context what degree of difference may be relevant. All calculations were performed using the R statistical language version 3.6 (R Project for Statistical Computing).

Results

Plotting of trajectories from 101 studies reporting mid- to long-term weight loss outcomes around the world demonstrated a range of outcomes (eAppendix 1, eTable 1, and eTable 2 in the Supplement; Figure 1A). Studies with the longest follow-up times appeared to have better total weight loss outcomes. Eleven of the recovered reports were from Canada (eTable 2 in the Supplement), and these fell in the lower half of trajectory results collected across the world.

After processing, 68 studies remained for trajectory analysis (Figure 1B), roughly evenly distributed among the 5 trajectory groups, although slightly more studies fell within the 2 groups with the poorest trajectories (19.1% in group 4 [G4] and 29.4% in G5, for a total 48.5% of studies) compared with the 2 groups with the best trajectories (17.6% each G1 and G2, for 35.2% of studies). Shapes were overall similar, except for G3, which was characterized by early higher weight loss followed by more rapid weight regain. Comparison of the Canadian data with modeled groups suggested that the poorest trajectory group, G5, was the most representative of Canadian results (Figure 1C). For the economic analysis, G5 was used to represent the postsurgical standard of care for Canada, with G1 used as the improved care pathway.

For the main analysis, the 100-patient cohort (mean [SD] 88.2% [1.4%] female; mean [SD] age, 43.6 [9.2] years; mean [SD] body mass index, 49.4 [8.2]; and mean [SD] comorbidity prevalence of 50.0% [4.1%], 66.0% [3.9%], and 59.3% [4.0%] for diabetes, hypertension, and dyslipidemia, respectively) analyzed used the wait-list patient demographic characteristics of Padwal et al²³ because in this evaluation, patients began at referral, had a waiting period during which comorbidity and body mass index may have changed, and then underwent surgery.

The burden of comorbidities includes initial baseline prevalence of comorbidities plus new-onset cases that developed during the wait for surgery, and this prevalence became the new baseline for surgical resolution for patients remaining in the surgical program. A reduction in treatment burden was found for diabetes, hypertension, and dyslipidemia (Figure 2A). Ten-year results saw the improved pathway associated with 324 (95% CrI, 249-396), 245 (95% CrI, 163-356), and 255 (95% CrI, 169-352) fewer patient-years of treatment for diabetes, hypertension, and dyslipidemia, respectively, compared with the standard of care. Considering the new-onset cases separately from those not resolved by surgery, more new cases of comorbidities (with the greatest increase seen for hypertension) were estimated to occur in the standard care pathway (Figure 2B). Conversion to costs yielded a considerably different impact to public payers, despite each comorbidity having reductions in patient-years of treatment on a similar order of magnitude (Figure 2C). Notably, dyslipidemia public payer costs were lower than the others over the time horizon as they are driven by medications,³² which are not typically covered by public health insurance programs in Canada. Total savings over the 10 years following referral, determined by earlier surgical delivery and better postsurgical trajectory, amounted to $900 000 (95% CrI, $630 000 to $1.2 million).

Figure 2. — A, Cumulative curves (lines) with 95% credibility intervals (shaded regions) depicting differences in patient-years of treatment (savings as treatment in the standard care case minus treatment in the improved care pathway case) for total prevalence of type 2 diabetes, hypertension, and dyslipidemia. B, Of those prevalent cases, the total numbers of incident cases in the 2 pathways over the 10-year time horizon are shown, with error bars depicting the 95% credibility intervals. C, Patient-years of treatment were converted to costs, stratified by individual comorbidity and for total public payer burden. CAD indicates Canadian dollars.

Absolute totals and relative differences were also calculated for the delayed and improved pathways (Figure 3; eAppendix 4 and eTable 4 in the Supplement). In the base case, the total 10-year savings in the improved care pathway of $900 000 thus occurred on a background estimated cost of $3.1 million, corresponding to savings of 29% (95% CrI, 20%-42%), or $900 of $3000 per average patient per year. Cumulative curves for patient-years of treatment and costs reveal the time-dependent context in which the observed savings occur (eFigure 4 in the Supplement).

In sensitivity analyses, the first assumption tested was that of presurgical weight gain (eAppendix 5 and eFigure 5 in the Supplement). Results indicate no significant association with total costs if presurgical weight remains stable, or if patients achieve weight loss. Similarly, a test of slower comorbidity development in the presurgical wait period did not considerably reduce the total 10-year costs (eFigure 6 in the Supplement). To sample results across a range of surgical wait time improvements, combinations of various improved and standard care surgery delivery times were tested (Table 2). All intervals excluded 0, suggesting every combination of surgical delivery improvement from 6 months up to 4 years earlier yielded savings. The savings estimated where both improved and standard care pathways deliver surgery at 2.5 years ($150 000 [95% CrI, −$10 000 to $320 000]) are indicative of the contribution afforded by the included improvement in postsurgical trajectory. When results are assessed for different postsurgical weight trajectories achieved (G1 to G4; Figure 1B), the lowest median savings observed among the 4 potential trajectory groups was $780 000 (95% CrI, $540 000 to $1.1 million) if trajectory G3 were achieved relative to standard of care (eFigure 7 in the Supplement), still inclusive of reduction in surgical delivery time.

Table 2. Sensitivity Analysis of Total Costs 10 Years After Referral^a.

Improved Care Pathway Time From Referral to Surgery	Standard Care Pathway Time From Referral to Surgery
	2.5 y		3.0 y		3.5 y		4.0 y		4.5 y
	Cost, Median (95% CrI), CAD	Cost-Saving Replicates, %	Cost, Median (95% CrI), CAD	Cost-Saving Replicates, %	Cost, Median (95% CrI), CAD	Cost-Saving Replicates, %	Cost, Median (95% CrI), CAD	Cost-Saving Replicates, %	Cost, Median (95% CrI), CAD	Cost-Saving Replicates, %
0.5 y	890 000 (630 000 to 1.23 million)	100	1000 000 (730 000 to 1.38 million)	100	1.12 million (810 000 to 1.52 million)	100	1.22 million (890 000 to 1.66 million)	100	1.32 million (950 000 to 1.79 million)	100
1.0 y	670 000 (440 000 to 950 000)	100	780 000 (550 000 to 1.09 million)	100	900 000 (630 000 to 1.24 million)	100	1.00 million (710 000 to 1.38 million)	100	1.09 million (770 000 to 1.52 million)	100
1.5 y	470 000 (280 000 to 700 000)	100	580 000 (380 000 to 850 000)	100	700 000 (470 000 to 990 000)	100	800 000 (550 000 to 1.13 million)	100	890 000 (600 000 to 1.28 million)	100
2.0 y	300 000 (130 000 to 500 000)	100	410 000 (240 000 to 640 000)	100	530 000 (340 000 to 770 000)	100	630 000 (410 000 to 910 000)	100	720 000 (460 000 to 1.06 million)	100
2.5 y	150 000 (–10 000 to 320 000)	97	270 000 (110 000 to 450 000)	100	380 000 (210 000 to 580 000)	100	480 000 (290 000 to 720 000)	100	570 000 (340 000 to 870 000)	100

Open in a new tab

Abbreviations: CAD, Canadian dollars; CrI, credibility interval.

^{^a}

Results are shown for total costs 10 years after referral for a 100-patient cohort in an improved care pathway (with surgery offered at times along the farthest left column) vs the standard of care pathway (with surgical times at the top of the second through sixth columns). The pathways include the postsurgical trajectory improvement, such that the standard care pathway has patients following the poor trajectory and the improved pathway has patients following the good trajectory. Values shown are median cost savings, the 95% CrI of cost savings, and the percentage of replicates that were cost saving when comparing costs in the standard care pathway with those in the improved surgical delivery pathway. Savings are expressed in 2018 CAD rounded to the nearest $10 000.

Given the combination of earlier surgery and better postsurgical trajectory in the main analysis, further sensitivity analyses were conducted to test for the independence of these contributions to the overall estimated benefit. Assuming both the improved and standard pathways had the same postsurgical weight trajectory (G5), various combinations of wait time reductions were sampled (eTable 5 in the Supplement). Again, the smallest improvement of 6 months advancement of surgery yielded savings ($120 000 [95% CrI, −$10 000 to $270 000). If surgery is taken to occur at the same time in both care pathways, but with the postsurgical trajectory comparisons between the standard of care (G5) and improved pathways (G1-G4), then savings decrease as the improved group moves from G1 to G4 (eFigure 8 in the Supplement). If both care pathways achieve surgery at the standard of care 3.5 years after referral, estimated savings if the best trajectory (G1) is achieved are $110 000 (95% CrI, −$30 000 to $270 000), with 94% of simulations being cost saving. The proportion of cost-saving simulations for groups G2, G3, and G4 were 88%, 76%, and 76%, respectively (eFigure 8 in the Supplement).

Discussion

Ideally, audits of real-world data representative of the clinical experience provide the best measure of the outcomes associated with changes to health care provision. In the absence of such data, implementing change can be fraught with uncertainty. Estimates from appropriately designed models may give an indication of potential outcomes of such changes without the time, expense, or risk to patients involved with obtaining real-world data. In the present study, the care pathway in Canada for patients with obesity undergoing SG surgery was found to have considerable room for improvement. Wait times were sufficiently long that some patients were likely to drop out of the program,²³ and of those who remained, total weight loss trajectories after surgery were among the poorest observed across the world.

Considering practices and outcome results from other settings provides a basis for proposed realistic improvements, as they have been achieved elsewhere and are not purely theoretical. If surgical wait times could, on average, be shortened and postsurgical weight trajectories improved toward a group with sustained long-term weight loss, patients in the Canadian setting would be estimated to incur fewer patient-years of comorbidity treatment for diabetes, hypertension, and dyslipidemia. The first 2 of these represent considerable public payer costs, while the impact of the latter may incur other burden beyond the scope of the present analysis. The presence of dyslipidemia may increase the risk of devastating cardiovascular events, such as stroke or myocardial infarction.

Despite its increasing use, few authors in Canada have published long-term results after SG. Only 1 study³³ was identified with 5-year follow-up, while most others^{18,19,34,35,36,37,38,39} stopped at 2 to 3 years, which, from inspection of the global summary data, is around the time of maximum weight loss or the beginnings of weight regain (Figure 1). More long-term data regarding outcomes are required, as these patients may continue to engage with health care services beyond this time. Addressing presurgical issues in Canada, efforts have been made to increase capacity for more bariatric surgery. The volume increased almost 4-fold between 2007 and 2013 (1578 to 5989 surgical procedures)¹³ but, more recently, growth has slowed (8539 surgical procedures by year ending 2016, representing a 42% increase from 2013).⁹ Demand still, however, outstrips capacity, as suggested by persistently lengthy wait times for surgery and issues of whether surgery is offered by province according to a recent report.¹⁴ A survey of surgeons in the province of Quebec, Canada, suggested that while most considered a wait time of up to 6 months acceptable, only a third achieved this goal, citing access to surgical rooms and access to other health care specialists, such as dedicated bariatric nutritionists, as impediments.⁴⁰ Improved capacity, throughput, or organization of multidisciplinary care may address these issues. The target of 1 year for surgery in a publicly funded system, as proposed for the improved care pathway in the present study, is not unreasonable and has been reported in Spain (average nationwide wait time, 397 days)⁴¹ and observed in a program in Toronto, Ontario (wait time, 445 days).²⁵ Where earlier surgery cannot be achieved, an Australian study⁴² of a public system suggested increased peer and health professional support to help counteract the development or worsening of depression or other obesity-related comorbidities, and ultimate presurgical weight gain during the wait, despite weight loss efforts.

Numerous studies have attempted to address factors predicting or influencing postsurgical weight loss outcomes, including genetics, sex, and race.^6,8 These parameters, however, are not readily modifiable, nor can we yet apply these associations to inform changes to clinical practice.⁴³ Potential changes to clinical management that can be applied independently of patient demographic characteristics are required. Improved weight loss outcomes have been associated with closer care, including attendance at postsurgical appointments.^44,45,46 Especially in the early postsurgical period, close follow-up may improve later outcomes, as early weight loss after gastric bypass or SG has been shown to be independently associated with maximal weight loss, allowing for identification of nonresponders and intervention to manage nonresponse.⁴⁷ Patient perception also plays a role, as interviews among patients at a Canadian center revealed beliefs that further follow-up was unnecessary or too impersonal, thereby presenting barriers to motivation to engage with specialist care.⁴⁸ It is otherwise unclear why published Canadian weight loss outcomes are generally worse than those in other settings; a future investigation may examine differences in training or surgical technique as potential contributors.

Limitations

As a model, the presented results relate to hypothetical improvements in care. Individual jurisdictions must assess what range of improvement is feasible based on available resources, which vary greatly among the provinces in Canada. The study population used for baseline model parameters, for example, is from Alberta, Canada, and may not be representative of the entire country. Sensitivity analyses probing a range of improvements (such as shortening wait time by as little as 6 months and as much as 4 years) at least partially address this issue. The fitting of equations, for example, dropout rate from survival analysis or modeling comorbidity incidence, is an assumption. In the latter case of comorbidity evolution, the data used may not be ideal given the inclusion of patients who were denied surgery, not just those awaiting surgery, potentially indicating an unhealthier group.²⁴ We used a parsimonious approach, using simpler models (linear, exponential, power) in the absence of specific data to avoid overfitting of the limited data available, and uncertainty was captured in the sensitivity analyses. More data on trajectories and comorbidity outcomes after different types of surgical procedure are needed, especially data in the presurgical period while patients await surgery. The principle of the analysis is applicable to any locale where there may be barriers to access for bariatric surgery such as capacity or insurance-mandated prerequisites, or where postsurgical outcomes need to be improved. Detailed costs, however, may not be broadly generalizable, as it has been noted that outcomes may differ between publicly funded programs as in Canada, and privately funded ones such as in the United States, as the contributing populations may be composed of differing socioeconomic backgrounds.¹⁹

Conclusions

Limited data are available regarding long-term outcomes after SG in Canada, but published reports suggest there is scope for improvement compared with results from other locations. Shortening of surgical wait times and increased weight loss in postsurgical trajectory in the present analysis were associated with fewer patient-years of treatment for the common comorbidities diabetes, hypertension, and dyslipidemia and considerable savings to the public payer in their associated costs. The modeled improvements are reasonable, based on what has been achieved elsewhere in similarly publicly funded bariatric programs. The estimated savings may, therefore, provide a basis for discussions around investments to improving bariatric surgical care in Canada.

Supplement.

eAppendix 1. Sleeve Gastrectomy Weight Loss Outcomes Literature

eAppendix 2. Statistical Analysis and Model Fitting to Sleeve Gastrectomy Data

eAppendix 3. Comparative Sleeve Gastrectomy Versus Gastric Bypass Co-morbidity Resolution Meta-analysis

eAppendix 4. Raw Totals and Relative Change of Co-morbidity Treatment and Costs Between Standard of Care and Improved Pathways

eAppendix 5. Sensitivity Analyses

eReferences.

eTable 1. Sleeve Gastrectomy Weight Loss Studies Extracted From Systematic Reviews and Meta-analyses

eTable 2. Sleeve Gastrectomy Weight Loss Studies Obtained From Supplemental Searches

eTable 3. Co-morbidity Resolution Meta-analyses and Study Data of Sleeve Gastrectomy Versus Gastric Bypass

eTable 4. Relative Change in Co-morbidity Treatment and Associated Costs

eTable 5. Sensitivity Analysis Reducing Wait Time for Surgery With Standard of Care Post-surgical Weight Trajectory

eFigure 1. Meta-analysis Results: Type 2 Diabetes Resolution, Sleeve Gastrectomy Versus Gastric Bypass

eFigure 2. Meta-analysis Results: Hypertension Resolution, Sleeve Gastrectomy Versus Gastric Bypass

eFigure 3. Meta-analysis Results: Dyslipidemia Resolution, Sleeve Gastrectomy Versus Gastric Bypass

eFigure 4. Stratified Costs for Delayed and Improved Care Pathways Around Sleeve Gastrectomy Surgery

eFigure 5. Sensitivity Analysis of Overall Costs According to Pre-surgical Weight Trajectory

eFigure 6. Sensitivity Analysis of Overall Costs According to Pre-surgical Weight Trajectory

eFigure 7. Sensitivity Analysis of Basecase Varying Post-surgical Weight Trajectory With Improved Surgical Delivery Time

eFigure 8. Sensitivity Analysis Varying Post-surgical Weight Trajectory With Identical Times of Delivering Surgery

Click here for additional data file.^{(1.3MB, pdf)}

References

1.Khorgami Z, Shoar S, Saber AA, Howard CA, Danaei G, Sclabas GM. Outcomes of bariatric surgery versus medical management for type 2 diabetes mellitus: a meta-analysis of randomized controlled trials. Obes Surg. 2019;29(3):-. doi: 10.1007/s11695-018-3552-x [DOI] [PubMed] [Google Scholar]
2.Kodama S, Fujihara K, Horikawa C, et al. Network meta-analysis of the relative efficacy of bariatric surgeries for diabetes remission. Obes Rev. 2018;19(12):1621-1629. doi: 10.1111/obr.12751 [DOI] [PubMed] [Google Scholar]
3.Schauer PR, Bhatt DL, Kirwan JP, et al. ; STAMPEDE Investigators . Bariatric surgery versus intensive medical therapy for diabetes—5-year outcomes. N Engl J Med. 2017;376(7):641-651. doi: 10.1056/NEJMoa1600869 [DOI] [PMC free article] [PubMed] [Google Scholar]
4.Shoar S, Saber AA. Long-term and midterm outcomes of laparoscopic sleeve gastrectomy versus Roux-en-Y gastric bypass: a systematic review and meta-analysis of comparative studies. Surg Obes Relat Dis. 2017;13(2):170-180. doi: 10.1016/j.soard.2016.08.011 [DOI] [PubMed] [Google Scholar]
5.Wang Y, Song YH, Chen J, et al. Roux-en-Y gastric bypass versus sleeve gastrectomy for super obese and super obese: systematic review and meta-analysis of weight results, comorbidity resolution. Obes Surg. 2019;29(6):1954-1964. doi: 10.1007/s11695-019-03817-4 [DOI] [PubMed] [Google Scholar]
6.Courcoulas AP, King WC, Belle SH, et al. Seven-year weight trajectories and health outcomes in the Longitudinal Assessment of Bariatric Surgery (LABS) study. JAMA Surg. 2018;153(5):427-434. doi: 10.1001/jamasurg.2017.5025 [DOI] [PMC free article] [PubMed] [Google Scholar]
7.Zhou K, Wolski K, Malin SK, et al. Impact of weight loss trajectory following randomization to bariatric surgery on long-term diabetes glycemic and cardiometabolic parameters. Endocr Pract. 2019;25(6):572-579. doi: 10.4158/EP-2018-0522 [DOI] [PubMed] [Google Scholar]
8.de Toro-Martín J, Guénard F, Tchernof A, Pérusse L, Marceau S, Vohl MC. Polygenic risk score for predicting weight loss after bariatric surgery. JCI Insight. 2018;3(17):122011. doi: 10.1172/jci.insight.122011 [DOI] [PMC free article] [PubMed] [Google Scholar]
9.Anvari M, Lemus R, Breau R. A landscape of bariatric surgery in Canada: for the treatment of obesity, type 2 diabetes and other comorbidities in adults. Can J Diabetes. 2018;42(5):560-567. doi: 10.1016/j.jcjd.2017.12.007 [DOI] [PubMed] [Google Scholar]
10.Jackson C, Snyder J, Crooks VA, Lavergne MR. “I didn’t have to prove to anybody that I was a good candidate”: a case study framing international bariatric tourism by Canadians as circumvention tourism. BMC Health Serv Res. 2018;18(1):573. doi: 10.1186/s12913-018-3385-2 [DOI] [PMC free article] [PubMed] [Google Scholar]
11.Imbus JR, Voils CI, Funk LM. Bariatric surgery barriers: a review using Andersen’s Model of Health Services Use. Surg Obes Relat Dis. 2018;14(3):404-412. doi: 10.1016/j.soard.2017.11.012 [DOI] [PMC free article] [PubMed] [Google Scholar]
12.Wharton S, Serodio KJ, Kuk JL, Sivapalan N, Craik A, Aarts MA. Interest, views and perceived barriers to bariatric surgery in patients with morbid obesity. Clin Obes. 2016;6(2):154-160. doi: 10.1111/cob.12131 [DOI] [PubMed] [Google Scholar]
13.Canadian Institute for Health Information . Bariatric Surgery in Canada. Ottawa, ON: Canadian Institute for Health Information; 2014. [Google Scholar]
14.Canadian Obesity Network . Report card on access to obesity treatment for adults in Canada 2017. https://obesitycanada.ca/publications/canadians-lack-access-obesity-treatments-support-report-card/. Accessed April 3, 2019.
15.Christou NV, Efthimiou E. Bariatric surgery waiting times in Canada. Can J Surg. 2009;52(3):229-234. [PMC free article] [PubMed] [Google Scholar]
16.Khan S, Rock K, Baskara A, Qu W, Nazzal M, Ortiz J. Trends in bariatric surgery from 2008 to 2012. Am J Surg. 2016;211(6):1041-1046. doi: 10.1016/j.amjsurg.2015.10.012 [DOI] [PubMed] [Google Scholar]
17.Karmali S, Schauer P, Birch D, Sharma AM, Sherman V. Laparoscopic sleeve gastrectomy: an innovative new tool in the battle against the obesity epidemic in Canada. Can J Surg. 2010;53(2):126-132. [PMC free article] [PubMed] [Google Scholar]
18.Hoogerboord M, Wiebe S, Klassen D, Ransom T, Lawlor D, Ellsmere J. Laparoscopic sleeve gastrectomy: perioperative outcomes, weight loss and impact on type 2 diabetes mellitus over 2 years. Can J Surg. 2014;57(2):101-105. doi: 10.1503/cjs.024212 [DOI] [PMC free article] [PubMed] [Google Scholar]
19.Gill RS, Majumdar SR, Rueda-Clausen CF, et al. Comparative effectiveness and safety of gastric bypass, sleeve gastrectomy and adjustable gastric banding in a population-based bariatric program: prospective cohort study. Can J Surg. 2016;59(4):233-241. doi: 10.1503/cjs.013315 [DOI] [PMC free article] [PubMed] [Google Scholar]
20.Clapp B, Wynn M, Martyn C, Foster C, O’Dell M, Tyroch A. Long term (7 or more years) outcomes of the sleeve gastrectomy: a meta-analysis. Surg Obes Relat Dis. 2018;14(6):741-747. doi: 10.1016/j.soard.2018.02.027 [DOI] [PubMed] [Google Scholar]
21.Juodeikis Ž, Brimas G. Long-term results after sleeve gastrectomy: a systematic review. Surg Obes Relat Dis. 2017;13(4):693-699. doi: 10.1016/j.soard.2016.10.006 [DOI] [PubMed] [Google Scholar]
22.Brethauer SA, Aminian A, Romero-Talamás H, et al. Can diabetes be surgically cured? long-term metabolic effects of bariatric surgery in obese patients with type 2 diabetes mellitus. Ann Surg. 2013;258(4):628-636. doi: 10.1097/SLA.0b013e3182a5034b [DOI] [PMC free article] [PubMed] [Google Scholar]
23.Padwal RS, Rueda-Clausen CF, Sharma AM, et al. Weight loss and outcomes in wait-listed, medically managed, and surgically treated patients enrolled in a population-based bariatric program: prospective cohort study. Med Care. 2014;52(3):208-215. doi: 10.1097/MLR.0000000000000070 [DOI] [PubMed] [Google Scholar]
24.Al Harakeh AB, Burkhamer KJ, Kallies KJ, Mathiason MA, Kothari SN. Natural history and metabolic consequences of morbid obesity for patients denied coverage for bariatric surgery. Surg Obes Relat Dis. 2010;6(6):591-596. doi: 10.1016/j.soard.2010.08.012 [DOI] [PubMed] [Google Scholar]
25.Diamant A, Cleghorn MC, Milner J, et al. Patient and operational factors affecting wait times in a bariatric surgery program in Toronto: a retrospective cohort study. CMAJ Open. 2015;3(3):E331-E337. doi: 10.9778/cmajo.20150020 [DOI] [PMC free article] [PubMed] [Google Scholar]
26.Alvarez R, Bonham AJ, Buda CM, Carlin AM, Ghaferi AA, Varban OA. Factors associated with long wait times for bariatric surgery. Ann Surg. 2019;270(6):1103-1109. [DOI] [PubMed] [Google Scholar]
27.Canadian Agency for Drugs and Technologies in Health . Guidelines for the Economic Evaluation of Health Technologies: Canada. 4th ed. Ottawa, ON: Canadian Agency for Drugs and Technologies in Health; 2017:19. [Google Scholar]
28.Rosella LC, Lebenbaum M, Fitzpatrick T, et al. Impact of diabetes on healthcare costs in a population-based cohort: a cost analysis. Diabet Med. 2016;33(3):395-403. doi: 10.1111/dme.12858 [DOI] [PMC free article] [PubMed] [Google Scholar]
29.Weaver CG, Clement FM, Campbell NR, et al. ; Alberta Kidney Disease Network and the Interdisciplinary Chronic Disease Collaboration . Healthcare costs attributable to hypertension: Canadian population-based cohort study. Hypertension. 2015;66(3):502-508. doi: 10.1161/HYPERTENSIONAHA.115.05702 [DOI] [PubMed] [Google Scholar]
30.Huang X, Liu T, Zhong M, Cheng Y, Hu S, Liu S. Predictors of glycemic control after sleeve gastrectomy versus Roux-en-Y gastric bypass: a meta-analysis, meta-regression, and systematic review. Surg Obes Relat Dis. 2018;14(12):1822-1831. doi: 10.1016/j.soard.2018.08.027 [DOI] [PubMed] [Google Scholar]
31.Li J, Lai D, Wu D. Laparoscopic Roux-en-Y gastric bypass versus laparoscopic sleeve gastrectomy to treat morbid obesity-related comorbidities: a systematic review and meta-analysis. Obes Surg. 2016;26(2):429-442. doi: 10.1007/s11695-015-1996-9 [DOI] [PubMed] [Google Scholar]
32.Conly J, Clement F, Tonelli M, et al. ; Alberta Kidney Disease Network . Cost-effectiveness of the use of low- and high-potency statins in people at low cardiovascular risk. CMAJ. 2011;183(16):E1180-E1188. doi: 10.1503/cmaj.101281 [DOI] [PMC free article] [PubMed] [Google Scholar]
33.Caron M, Hould FS, Lescelleur O, et al. Long-term nutritional impact of sleeve gastrectomy. Surg Obes Relat Dis. 2017;13(10):1664-1673. doi: 10.1016/j.soard.2017.07.019 [DOI] [PubMed] [Google Scholar]
34.Biertho L, Lebel S, Marceau S, et al. Laparoscopic sleeve gastrectomy: with or without duodenal switch? a consecutive series of 800 cases. Dig Surg. 2014;31(1):48-54. doi: 10.1159/000354313 [DOI] [PubMed] [Google Scholar]
35.Elbahrawy A, Bougie A, Loiselle SE, Demyttenaere S, Court O, Andalib A. Medium to long-term outcomes of bariatric surgery in older adults with super obesity. Surg Obes Relat Dis. 2018;14(4):470-476. doi: 10.1016/j.soard.2017.11.008 [DOI] [PubMed] [Google Scholar]
36.Garofalo F, Denis R, Pescarus R, Atlas H, Bacon SL, Garneau P. Long-term outcome after laparoscopic sleeve gastrectomy in patients over 65 years old: a retrospective analysis. Surg Obes Relat Dis. 2017;13(1):1-6. doi: 10.1016/j.soard.2016.05.020 [DOI] [PubMed] [Google Scholar]
37.Gregory DM, Twells LK, Lester KK, et al. Preoperative and postoperative assessments of biochemical parameters in patients with severe obesity undergoing laparoscopic sleeve gastrectomy. Obes Surg. 2018;28(8):2261-2271. doi: 10.1007/s11695-017-3007-9 [DOI] [PubMed] [Google Scholar]
38.Lemus R, Karni D, Hong D, Gmora S, Breau R, Anvari M. The impact of bariatric surgery on insulin-treated type 2 diabetes patients. Surg Endosc. 2018;32(2):990-1001. doi: 10.1007/s00464-017-5777-5 [DOI] [PubMed] [Google Scholar]
39.Strain GW, Saif T, Gagner M, Rossidis M, Dakin G, Pomp A. Cross-sectional review of effects of laparoscopic sleeve gastrectomy at 1, 3, and 5 years. Surg Obes Relat Dis. 2011;7(6):714-719. doi: 10.1016/j.soard.2011.08.021 [DOI] [PubMed] [Google Scholar]
40.Andalib A, Bouchard P, Bougie A, Loiselle SE, Demyttenaere S, Court O. Variability in bariatric surgical care among various centers: a survey of all bariatric surgeons in the province of Quebec, Canada. Obes Surg. 2018;28(8):2327-2332. doi: 10.1007/s11695-018-3157-4 [DOI] [PubMed] [Google Scholar]
41.Arteaga-González IJ, Martín-Malagón AI, Ruiz de Adana JC, de la Cruz Vigo F, Torres-García AJ, Carrillo-Pallares AC. Bariatric surgery waiting lists in Spain. Obes Surg. 2018;28(12):3992-3996. doi: 10.1007/s11695-018-3453-z [DOI] [PubMed] [Google Scholar]
42.Sharman MJ, Venn AJ, Jose KA, et al. The support needs of patients waiting for publicly funded bariatric surgery—implications for health service planners. Clin Obes. 2017;7(1):46-53. doi: 10.1111/cob.12169 [DOI] [PubMed] [Google Scholar]
43.Aasbrenn M, Schnurr TM, Have CT, et al. Genetic determinants of weight loss after bariatric surgery. Obes Surg. 2019;29(8):2554-2561. doi: 10.1007/s11695-019-03878-5 [DOI] [PubMed] [Google Scholar]
44.Schwoerer A, Kasten K, Celio A, Pories W, Spaniolas K. The effect of close postoperative follow-up on co-morbidity improvement after bariatric surgery. Surg Obes Relat Dis. 2017;13(8):1347-1352. doi: 10.1016/j.soard.2017.03.024 [DOI] [PubMed] [Google Scholar]
45.Shilton H, Gao Y, Nerlekar N, Evennett N, Ram R, Beban G. Pre-operative bariatric clinic attendance is a predictor of post-operative clinic attendance and weight loss outcomes. Obes Surg. 2019;29(7):2270-2275. doi: 10.1007/s11695-019-03843-2 [DOI] [PubMed] [Google Scholar]
46.Spaniolas K, Kasten KR, Celio A, Burruss MB, Pories WJ. Postoperative follow-up after bariatric surgery: effect on weight loss. Obes Surg. 2016;26(4):900-903. doi: 10.1007/s11695-016-2059-6 [DOI] [PubMed] [Google Scholar]
47.Manning S, Pucci A, Carter NC, et al. Early postoperative weight loss predicts maximal weight loss after sleeve gastrectomy and Roux-en-Y gastric bypass. Surg Endosc. 2015;29(6):1484-1491. doi: 10.1007/s00464-014-3829-7 [DOI] [PMC free article] [PubMed] [Google Scholar]
48.Aarts MA, Sivapalan N, Nikzad SE, Serodio K, Sockalingam S, Conn LG. Optimizing bariatric surgery multidisciplinary follow-up: a focus on patient-centered care. Obes Surg. 2017;27(3):730-736. doi: 10.1007/s11695-016-2354-2 [DOI] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials