Obesity-related costs and the economic impact of laparoscopic adjustable gastric banding procedures: benefits in the Texas Employees Retirement System

Abstract

Objective: To assess the return on investment (ROI) and economic impact of providing insurance coverage for the laparoscopic adjustable gastric banding (LAGB) procedure in classes II and III obese members of the Texas Employees Retirement System (ERS) and their dependents from payer, employer, and societal perspectives.

Methods: Classes II and III obese employee members and their adult dependents were identified in a Texas ERS database using self-reported health risk assessment (HRA) data. Direct health costs and related absenteeism and mortality losses were estimated using data from previous research. A dynamic input–output model was then used to calculate overall economic effects by incorporating direct, indirect, and induced impacts. Direct health costs were inflation-adjusted to 2008 US dollars using the Consumer Price Index for Medical Care and other spending categories were similarly adjusted using relevant consumer and industrial indices. The future cost savings and other monetary benefits were discounted to present value using a real rate of 4.00%.

Results: From the payer perspective (ERS), the payback period for direct health costs associated with the LAGB procedure was 23–24 months and the annual return (over 5 years) was 28.8%. From the employer perspective (State of Texas), the costs associated with the LAGB procedure were recouped within 17–19 months (in terms of direct, indirect, and induced gains as they translated into State revenue) and the annual return (over 5 years) was 45.5%. From a societal perspective, the impact on total business activity for Texas (over 5 years) included gains of $195.3 million in total expenditures, $93.8 million in gross product, and 1354 person-years of employment.

Limitations: The analysis was limited by the following: reliance on other studies for methodology and use of a control sample; restriction of cost savings to 2.5 years which required out-of-sample forecasting; conservative assumptions related to the cost of the procedure; exclusion of presenteeism; and no sensitivity analyses performed.

Conclusion: This analysis indicates that providing benefits for the LAGB procedure to eligible members of the Texas ERS and their dependents is worthy of support from payer, employer, and societal perspectives.

• bariatric surgery, economic impact, healthcare, laparoscopic adjustable gastric banding, LAGB, morbid obesity, return on investment, ROI, Texas Employees Retirement System

Introduction

The prevalence of obesity, body mass index (BMI) >30, among US adults has more than doubled within the last three decades[1]. As a result, more than one-third of the US adult population is obese with more than half the states having a prevalence between 25% and 29%[1],[2]. The World Health Organization (WHO) classifies obesity in three classes according to BMI: (1) class I, BMI 30.0–34.9 kg/m[2]; (2) class II, BMI 35.0–39.9 kg/m[2]; and (3) class III, BMI >40.0 kg/m[2]^–[3]. Class III obesity, also referenced as morbid obesity, is the fastest growing segment of obesity[4]. The prevalence of this particular obesity class increased by 50% between 2000 and 2005[4],[5].

Obesity is associated with serious health consequences (e.g., coronary heart disease, type 2 diabetes mellitus, hypertension, dyslipidemia, stroke, gallbladder disease, sleep apnea, and osteoarthritis), increased mortality, and decreased life expectancy[4],[6],[7]. Despite health consequences, there are also societal consequences related to obesity, such as absenteeism and presenteeism (productivity impairment at work), which can result in lost workdays, poor work performance, loss of earnings to workers, profits to companies, and tax revenues to the government[7],[8].

As a result of these consequences, there is significant economic burden evidenced by increased healthcare utilization and costs. Finkelstein et al. estimated the annual medical (health) costs of obesity in the US to be $147 billion per year[9]. By 2030, 1 out of every 6 healthcare dollars – up to $956.9 billion – could be spent on obesity treatment alone[10]. Employers are also burdened economically as a result of obesity. Ricci and Chee computed the annual cost of absenteeism and presenteeism attributed to obesity for employers to be $3.86 billion and $7.84 billion, respectively[7]. Trogdon and colleagues also reported similar estimates of obesity-related absenteeism costs, which ranged from $3.38 billion to $6.38 billion ($79–132 per obese person, respectively)[11]. In general, employers incur a disproportionate share of obesity-related costs in the form of absenteeism and presenteeism.

In addition, states with a higher prevalence of obesity, such as Texas, are likely to accrue even higher costs[2]. In a 2005 study by the Texas Comptroller of Public Accounts, obesity directly cost Texas employers $3.3 billion, which includes direct costs for healthcare and related indirect costs of employee absenteeism, lost productivity, and disability. This total is projected to increase to $15.8 billion annually by 2025 if the rising healthcare costs and obesity trends continue[12].

In the current economic and fiscal environment, both public and private entities are facing substantial pressure to reduce health costs. As attention has focused on increasing levels of obesity and the associated consequences, the decision of providing insurance coverage for obesity management programs and bariatric procedures has surfaced. Obesity management programs and bariatric procedures are now being considered by employers to reduce health costs and improve work productivity; however, questions still exist regarding the cost-effectiveness of these interventions[13]. Many decisions for these interventions are subjected to a cost- benefit analysis, where the return on investment (ROI) or the ratio of money saved to money consumed and payback period are critical elements of the evaluation process[14].

Bariatric surgical procedures have demonstrated to be one of the most effective obesity treatments in adults, achieving sustainable weight loss with improvement in comorbid illnesses over the long-term[5]. Laparoscopic adjustable gastric banding (LAGB) is a safe, effective, and reversible treatment for obesity and a potential option for employers to offer to their employees for weight reduction and health improvement[15].

Previous studies have examined various clinical aspects of bariatric surgery, but there is limited published literature that assesses the economic impact of bariatric surgery[16–19]. A recent study by Crémieux et al. evaluated ROI for open and laparoscopic bariatric surgery in the US and estimated all direct costs would be recouped within two and 4 years, respectively[4]. To date, no study has specifically addressed the ROI for the LAGB procedure.

This study was conducted at the request of the Texas State legislature. The purpose of this study was to assess the ROI and economic impact (from a payer, employer, and societal perspective) of providing insurance coverage for the LAGB procedure in Classes II and III obese Employees Retirement System (ERS) members and their dependents in the State of Texas.

Methods

Data source

The data source was a Texas ERS database which consisted of all State employees and retirees (n = 508,350), including dependents, based on the fiscal year of 2008. During this time period, the ERS incurred $1.9 billion in health plan expenditures, which was funded through member contributions and State appropriations.

Sample selection and study cohorts

Using the ERS database, analysis was restricted to active members and their adult dependents (greater than 18 years) identified to be in classes II and III obesity groups. Obesity was quantified in a retrospective analysis of self-reported health risk assessment (HRA) and work-loss data linked to medical claims for members with both medical and pharmacy benefits, similar to the methodology in the Durden et al. study[20]. The sample size was calculated based on national data for the proportion of individuals categorized in classes II and III obesity groups. Surgical eligibility was based on the bariatric surgical criteria (BMI ≥ 40 or ≥35 with comorbid conditions) from the National Institutes of Health (NIH) Consensus Development Conference Statement[21]. The theoretical candidates for LAGB consisted of approximately 9400 members and 4700 dependents. Although a complete demographic profile was not available, appropriate adjustments were incorporated for age, income level, and obesity patterns in Texas relative to the nation, and, in the case of dependents, labor force participation rates.

Perspectives

The economic impact of LAGB was assessed from three different perspectives: payer, employer, and society. For this analysis, the payer was the Texas ERS since it was a self-funded health plan. From the ERS perspective, only the direct healthcare costs reflected as its outlays were relevant. The employer was the State of Texas (and its taxpayers) which ultimately balances an overall budget that includes myriad components. In addition to the direct healthcare, the State incurred costs from absenteeism, lost productivity, and even foregone tax revenues from the reduced spending by employees and dependents and the production of its employees. These amounts did not flow back to the accounts of the ERS, but were quite relevant to the State legislature in its fiscal efforts. The societal perspective was defined as the economy as a whole.

Measures

An economic impact assessment tracks expenditures through an economy and measures the cumulative effects of those expenditures on various measures of economic activity. It also generates an estimate of the economic consequences of a particular initial stimulus (positive or negative) on the local economy. This study assessed the economic impact of the LAGB procedure[22]. Economic impacts can be measured by total expenditures, gross product, personal income, retail sales, permanent jobs, and person-years of employment (note that person-years were used to measure cumulative positions over multiple periods or those that are temporary in nature). Measures of economic impacts are illustrated in Figure 1.

Figure 1. Measures of economic impacts.

Total expenditures reflect the overall interplay of all industries in the economy and some key fiscal variables (e.g., sales taxes); it essentially represents every transaction that occurs between economic agents (e.g., consumers, producers) as a result of the process being evaluated. Gross product represents the regional equivalent of Gross Domestic Product (GDP) and is defined as the value of all final goods produced in a given region for a specific period of time. One property of gross product is that it eliminates any double counting from total expenditures. Personal income is the income received by individuals in the form of wages, salaries, interest, dividends, proprietors’ profits, or other sources. Retail sales represent the component of total expenditures which occurs in retail outlets and measures consumer activity. Permanent jobs anticipate that the relevant positions will be maintained on a continuing basis. Finally, person-years of employment reveals the annual full-time equivalent jobs (one person working for 1 year) generated by an activity.

There are different forms of economic activity: direct, indirect and induced. Direct, indirect, and induced impacts were analyzed for the fiscal year of 2008 to assess the economic impact of the LAGB procedure within the Texas ERS. In the context of this study, impacts comprise both costs and benefits. Direct impacts encompass the costs of preventive, diagnosis and treatment services related to the health condition (e.g., hospital care, physician services, and prescriptions)[23]. Indirect impacts comprise the value of lost output due to cessation or impairment of productivity (e.g., absenteeism, presenteeism) caused by morbidity and mortality[23]. This measure (indirect impacts) also includes the multiple rounds of spending which are lost due to direct spending reductions (or gained upon reversal of the trend). Induced impacts refer to the reductions in consumer spending as a result of lost wages associated with obesity and the corresponding gains when such patterns are reversed[22]. The summation of these impacts equal the total economic impact of a treatment service, which would all be specific to the treatment of obesity with the LAGB procedure (Figure 2). Examples of these inputs and their relationship to each perspective (i.e., payer, employer, and society) are listed in Figure 3.

Figure 2. The concept of total economic impact.

Figure 3. Summary of different inputs associated with different perspectives in evaluating return on investment (ROI) for the laparoscopic adjustable gastric banding (LAGB) procedure.

The direct impact of the LAGB procedure (including expenses pre- and post-procedure) was calculated based on two different sources: (1) a cost-effectiveness model and (2) a paired sample study by Crémieux et al.[4],[24],[25]. The methodologies of these sources are described within their studies. The values from both these sources were seen to be extremely comparable (within 0.6%) and the higher of the two (which generates the smallest net benefit) was employed. Adjustments were made for the typical copayment arrangement in the ERS health plan. In addition, the cost savings for future periods were also derived from the Crémieux et al. study[4]. As noted earlier, the current analysis also included the higher costs associated with untreated obese members, as these outlays reflect the likely incremental payments by ERS in the absence of LAGB coverage.

Indirect impacts (which are treated as direct inputs to the modeling process because they represent the first stage of a multiplier process) consisted of obesity-related absenteeism and productivity. The levels of absenteeism were derived from the Durden et al. and Combs studies[12],[20]. Durden and colleagues defined absenteeism as an absence from work with pay (including sick time) and found that employees in all three classes of obesity missed greater than 30 work days over a 2-year period. Combs and the Texas Comptroller of Public Accounts Division found that class II obese male and female employees missed 2 and 3 days more work than normal weight employees due to obesity, respectively[20]. All relevant work life and life expectancy (age) parameters were matched to the characteristics of ERS active members.

For this present analysis, absenteeism by State employees and the associated lost productivity were considered a cost to the State, whereas for the dependents, absenteeism was seen as a cost to their respective employers. The costs associated with reduced absences were adjusted to reflect current average salaries of State employees according to the US Department of Commerce, whereas incremental productivity was reflected by the ratio of gross product to income for the State government sector in Texas. Losses from absenteeism for the dependents sample were allocated across the other sectors of the economy based on private employment patterns. Productivity costs were estimated by multiplying the lost compensation by the ratio of output to compensation in the relevant sector (government). As absenteeism is reduced, the employer receives the full benefit of the worker being present which exceeds compensation; therefore, an adjustment was performed for this phenomenon. Although presenteeism is also an issue, the current published literature at the time of the analysis did not adequately address this topic for inclusion in this model.

Induced impacts from changes in consumer spending were estimated based on standard patterns of savings, taxpaying, and spending which were available from major national and regional surveys. These costs are often referred to as having a dynamic effect. A dynamic effect looks beyond the immediate effect and incorporates other indirect influences. For example, if someone is paid less because of their absence, that individual spends less, and hence, the State collects less in sales tax. The dynamic effect from both ERS members and their dependents were used in the analysis to provide the appropriate assessments from the perspectives of the employer (Texas government and taxpayers) and society (overall economy). Sources for information used to derive induced costs included the US Bureau of the Census, the US Bureau of Labor Statistics (BLS), the Regional Economic Information System of the US Department of Commerce, and other public and private sources. In addition, the Consumer Expenditure Survey by BLS and the American Chamber of Commerce Researchers’ Association (ACCRA) Cost of Living Index were utilized to apply the appropriate weights to these induced costs. For this study, induced impacts were simulated through the US Multi-Regional Impact Assessment System (USMRIAS), which is an economic impact model further discussed in the analysis section.

Analysis

The impacts of a policy decision to provide benefits for LAGB procedures to ERS members and their dependents are the subject of this analysis. Dynamic input–output modeling was selected to assess the economic impact of the LAGB procedure after quantification of direct impacts, incidence of obesity, and the potential improvements (e.g., weight loss, reduced absenteeism, increased productivity, improved mortality) from the ERS members and their dependents. The technique of dynamic input–output modeling allows estimation of the regional economic impacts (i.e., indirect and induced impacts) of decisions in one sector on the full range of other industries (i.e., the effect of changes in one industry on others and by consumers, government, foreign suppliers on the economy). It accounts for inter-industry relationships within regions and can help determine how regional economies are likely to respond to changes[26]. Examples of well-known input–output models include the Input–Output Model of the United States[27] and Regional Input–Output Modeling System (RIMS II)[26]. Both of these models are simple, static systems maintained by the US Department of Commerce that are widely used for routine impact assessments in a variety of contexts.

However, the input–output model utilized in this study was USMRIAS, due to the nature of the research objective. The USMRIAS is a proprietary model developed and maintained by one of the authors and has been published and described within the scientific literature[28]. It is constantly updated and has been used for numerous evaluations of healthcare funding issues, including wellness, mental health and substance abuse initiatives, cancer research and treatment outlays, and spending for Medicaid and the Children’s Health Insurance Program (CHIP)[29–33].

The USMRIAS estimates economic impacts using a variety of direct-effect multipliers (i.e., estimates of how much additional economic activity will result from an investment in the economy) through feed-forward calculations[22]. The USMRIAS is much more dynamic and extensive than the Input–Output Model of the United States and RIMS II and contains a number of significant enhancements. For example, the system provides: (1) comprehensive 500-sector coverage for any county, multi-county or urban region of the US (extended to other countries on several occasions); (2) calculation of total expenditures and value-added by industry and region; (3) direct expenditure estimation for multiple basic input choices (expenditures, output, income, or employment); (4) extensive parameter localization; (5) price adjustments for real and nominal assessments by sectors and areas which permit dynamic adjustments over time as relative prices evolve; (6) measurement of the induced impacts associated with payrolls and consumer spending which recognize the unique role of the household sector in the economy; and (7) comprehensive linkage and integration with a wide variety of econometric, real estate, occupational, and fiscal impact models. In particular, the linkage with an econometric system provides a full dynamic character to the system.

It should be noted that, because the system derives consumer spending patterns from the Consumer Expenditure Survey, it fully captures the fact that people at different income levels have varying spending patterns. Moreover, because input–output coefficients are non-stochastic in nature, results are not conducive to traditional confidence intervals (relevant USMRIAS coefficients are provided in the Appendix). It has been demonstrated, however, that parameter estimates across advanced economies with a natural resource endowment are relatively consistent, thus providing evidence of model stability[34]. In addition, in dynamic models with price sensitivity and a fully specified and interactive household sector, such as USMRIAS, applications of a dynamic nature have been shown to produce statistically significant results in appropriate applications[35],[36].

In calculating the economic impact of the LAGB procedure and its ROI estimate, the following adjustments were made: (1) appropriate age, income, and labor force participation calibrations, (2) expression of monetary values in constant 2008 US dollars using Consumer Price Index for Medical Care for health expenditures and appropriate consumer and industrial indices for other sectors or spending categories (e.g., Personal Consumption Deflator by category for the various elements of household spending and the Implicit Price Deflator for various categories of industrial outlays), and (3) discounting of cost-savings and associated impacts by 4% ‘real’ (inflation-adjusted) discount rate to express values at the time of procedure. This rate was selected to reflect a conservative valuation of the benefits to the State of Texas relative to its cost of funds. All costs were also expressed on a monthly basis.

The assumptions for the economic impact simulation included: (1) a 5-year time horizon; (2) a 5% participation rate (percentage of eligible members and dependents who actually underwent the procedure); (3) both ERS members and their dependents being typical of individuals who received bariatric surgery; (4) members having 100% probability of employment; (5) dependents earning the average private compensation, working in typical private sector production categories, and having a probability of working equal to the ratio of the employed work force to the adult population; (6) no adjustments for potential tradeoffs between the LAGB procedure and other bariatric procedures (as these procedures are not currently covered under the ERS benefits structure); and (7) the incremental resources spent in a manner typical of consumer spending patterns in Texas at the appropriate income levels.

Results

The final study sample included a total of approximately 14,100 employees and dependents (9400 employees and 4700 dependents) who met the NIH criteria for bariatric surgery[24]. The median age of the sample was 43.7 years.

Cost of classes II and III obesity

To provide an initial perspective, the USMRIAS was employed to estimate the overall costs of classes II and III obesity among ERS members and their dependents to include on the Texas economy. This analysis revealed losses of $4.59 billion in total expenditures, $2.20 billion in gross product, and 31,807 person-years of employment over a 5-year horizon to the Texan economy. These results are summarized for several aggregate indicators of economic activity in Figure 4.

Figure 4. Projected economic impact (losses) due to classes II and III obesity within Employees Retirement System of Texas (ERS) members and their dependents over 5 years.

The effects of providing benefits for the LAGB procedure: Payer perspective

The impact on the payer (ERS) was restricted to the direct health costs relative to the initial outlays surrounding the procedure. Table 1 provides a summary of the mean monthly direct health cost savings associated with LAGB. The economic impact assessment indicated a payback period of approximately 23–24 months for both classes II and III obese employee members and dependents, as well as, an annual rate of return (over 5 years) of 28.8% (26.1% on a net present value basis). A summary of the payback period and return on investment (average annual ROI) is provided in Table 2.

Table 1.  Summary of mean monthly health cost savings associated with the LAGB procedure for a Texas ERS member or dependent.

Post-procedure, $	Monthly ($)	NPV monthly ($)
Months 1–2	0	0
Months 3–6	−192.16	−189.98
Months 7–12	−591.66	−575.46
Months 13–18	−1109.94	−1058.56
Months 19 and longer	−1154.24	−1018.37

LAGB, laparoscopic adjustable gastric banding; ERS, Employees Retirement System; NPV, net present value

Table 2.  Health cost savings and benefits associated with providing LAGB coverage to a ‘typical’ eligible employee of the ERS or adult dependent from payer and employer perspectives expressed in payback period and average annual ROI (over 5 years).

	Payback period (months)	Average annual ROI (years 1–5)
Payer perspective (ERS)
Employee or dependent
Direct health costs savings	23.15	28.85%
NPV direct health costs	23.89	26.07%
Employer perspective (State of Texas)
Employee
Total State revenue benefits	17.33	45.56%
NPV State revenue benefits	18.80	40.06%
Dependent
Total State revenue benefits	21.93	35.00%
NPV State revenue benefits	22.92	29.90%

LAGB, laparoscopic adjustable gastric banding; ERS, Employees Retirement System of Texas (ERS); ROI, return-on-investment; NPV, net present value.

The effects of providing benefits for the LAGB procedure: employer perspective

The impact on the employer (the State of Texas) includes not only the cost savings from reduced direct health costs, but also the increased tax receipts stemming from the positive economic effects. The economic impact assessment of the dynamic revenue to Texas indicated the cost of the LAGB procedure for classes II and III obese employee members was recouped in 17–19 months (in terms of direct, indirect, and induced gains as they translated into State revenue), with a net gain of approximately $12,700 (∼$10,600 on a net present value basis) over a 2-year period. During a 5-year period, Texas would receive an annual rate of return of 45.6% (40.1% on a net present value basis) from the LAGB procedure on classes II and III obese members. As shown in Table 2, the cost of the LAGB procedure for classes II and III obese adult dependents was recouped in 22–23 months with a net gain of approximately $2761 ($1351 on a net present value basis) across 2 years; the annual rate of return (over 5 years) was 35.0% (29.9% on a net present value basis).

A class II obese employee member had a 6.8% reduction in total costs relative to the average costs of for all qualified obese members (classes II and III), which implies a payback period increase of approximately 7.3% relative to the above results. The cost for a Class III obese employee member was approximately 24% higher, implying a payback period that was shorter by approximately 19.4%. These changes were likely exaggerated in the dependent group, as the variations are generally higher for direct health costs than for absenteeism.

The economic impact of the LAGB procedure: societal perspective

Simulation of the USMRIAS model found that providing benefits for the LAGB and the associated reduction in direct health costs paid by the ERS, enhancement of overall productivity, and improvements in mortality yield a notable positive effect on the Texas economy. As seen in Figure 5, the net gains in business activity (over a 5-year horizon) were forecast to include $195.3 million in total expenditures, $93.8 million in gross product, and 1354 person-years of employment. While these findings provide an overall societal perspective, they do not account for the additional benefits such as gains in the quality of life of the affected individuals and their families.

Figure 5. Projected economic impact (gains) of providing laparoscopic adjustable gastric banding (LAGB) benefits to class II and III obese Employees Retirement System of Texas (ERS) members and their dependents over 5 years.

Discussion

Obesity is a critical health concern in the US and is associated with excessive direct impacts, which are increasing in obesity classes II and III. In addition, there are indirect impacts of obesity including absenteeism, lower productivity, disability, and other aspects of organizational efficiency (e.g., work performance) that have economic impact, as well as induced impacts. These obesity-related costs affect payers, employers, and society, as seen in the analysis of the State of Texas employees and their dependents.

This study provided information on obesity and its overall economic impact, as well as the potential benefits of providing insurance coverage for the LAGB procedure which could counteract and reduce the economic consequences of obesity. The economic impact assessment of the LAGB procedure appeared to have a beneficial impact on a direct, indirect, and induced basis. Although the LAGB procedure involves initial upfront costs, they are eventually offset by downstream savings associated with reduced healthcare needs and improved productivity. In fact, the payback periods are relatively short in comparison to many investments.

The payback period in this study was attained earlier than other previous studies that computed an ROI estimate for bariatric surgeries. Crémieux demonstrated that bariatric surgeries recouped their costs in 2 and 4 years for laparoscopic and open surgeries, respectively[4]. However, indirect and induced impacts (costs) were not included in that analysis. To date, the present study is the only study that has generated an ROI estimate for the LAGB procedure through an economic impact model which incorporates direct, indirect, and induced factors.

Reduced health costs and increased economic activity were the drivers for the cost-savings seen in this study. Cost reduction in direct and indirect health expenses may be attributable to health benefits, risk reduction, reduction in absenteeism, surgical experience, improved technology, and dedicated bariatric facilities[4],[13],[37]. Induced expenses may increase economic activity further through income recovery and psychosocial issues (e.g., self-esteem, body esteem). This study complements other published cost-effectiveness analyses that depict the LAGB procedure to be cost effective. For instance, Salem et al. and Keating et al. conclude that the LAGB procedure is cost effective for morbid obesity and obesity-related diabetes, respectively[38],[39].

Besides LAGB being cost effective on a direct basis, this study also showed the potential of the LAGB procedure to alleviate the burden of obesity on employers. Currently, only non-surgical obesity management work place interventions (e.g., diet, exercise, behavior programs) have been instituted nationwide to help employees reduce their weight, improve their health, and increase levels of productivity[13],[40],[41]. Employees have been incentivized directly through rewards and indirectly through subsidization of program costs. To date, these programs have not generally resulted in the sustained weight loss necessary to recoup overall costs[14],[42]. However, they are still supported by employers and recommended by healthcare providers[43]. In contrast, the weight loss from bariatric surgery has shown to be both dramatic and prolonged, as well as achievable in patient populations with multiple comorbid illnesses, yet they are not offered to employees due to lack of insurance coverage from payers[44].

This study was an attempt to examine whether there was a business case for insurance coverage of the LAGB procedure for the ERS (payer), State of Texas (employer), and society through the use of a dynamic input–output model. Analyses examining the net economic consequences of procedures with these types of models, such as the LAGB procedure, provide both novelty and relevance for two reasons: (1) it provides a vehicle to assess the effects of improved outcomes in a comprehensive manner that allows varying types of gains to be evaluated in a unified and consistent structure, and (2) it permits full consideration of the secondary (indirect and induced) impacts which emanate from improved outcomes.

Limitations

This study was based on a large Texas ERS database that includes limited information on demographics. BMI, work loss, and cost information were derived from the Durden et al. and Crémieux et al. studies (adjusted for known information regarding age, workforce participation, and general obesity patterns in Texas). The data cited from these studies have limitations. The Durden study, which provided the BMI and work loss data, was limited by the use of voluntary, self-reported data and relied on paid time off as the sole measure of work loss[22]. The Crémieux study was used to calculate the cost of the surgical procedure and is limited by: (1) the ROI estimates being driven by rising costs in the matched control group rather than by a reduction in costs from the bariatric patient group, and (2) ‘savings’ equating to the difference in costs between the procedure and the control groups in the post-procedure period, which is highly dependent on the accuracy of their matching process[4],[45]. Based on these limitations, there is a possibility that the data utilized in this study could be under- or over-estimated. It should be again noted, however, that the increased costs in the control group are relevant in the current context, as they are reflected on the potential outlays in the absence of funding the procedures.

There are also limitations directly associated to this study. First, this analysis was not a controlled design, which limits its ability to assess what costs would have been for non-surgical members. In this regard, the study relied on the patterns observed in the Crémieux study. The second limitation involves the use of claims data which does not allow the assessment of direct measures of disease severity (e.g., BMI). The third limitation is that some of the perspectives did not account for certain factors. For instance, this study did not include non-reimbursed health costs, such as copayments and over-the-counter drugs within the payer perspective. This approach (payer) sought to capture the typical cost-sharing patterns in ERS coverage, which would likely be different for other employers. In addition, the societal perspective did not account for the additional benefits such as gains in health-related quality of life (HRQoL). An avenue for future research would be to explore this aspect (i.e., HRQoL), perhaps using the methods set forth in the well-known work of Murphy and Topel[46].

The fourth limitation is that data are based on a limited follow-up period of 2.5 years post-procedure, with the economic impact analysis predicated on sample predictions. All savings, or lack thereof, are assumed to extend beyond the study period, although this assumption may not be accurate. Other methodologies may have also led to a conservative under-estimation of cost savings, such as adjusting for demographics of the ERS population, using national rates of classes II and III obesity (adjusted to overall Texas patterns), and assuming the spending on incremental productive resources is similar to consumer patterns within the State.

Finally, there were no sensitivity analyses incorporated in the present study because the model was characterized by fixed coefficients. The testing of alternatives would be limited to assumptions regarding underlying health costs, absenteeism, and similar factors associated with the treatment of obesity. Considering LAGB research in these areas is only beginning to develop, it would be beneficial to incorporate sensitivity analyses as more information becomes available

Despite these limitations, this analysis provides a novel approach to estimate an ROI for the LAGB procedure. Since this study is an initial effort to determine an ROI estimate from an economic impact model through the use of direct, indirect and induced impacts (costs and benefits), further research is warranted to validate this concept. Other studies may be needed with a similar methodology to strengthen the business case to use economic impact models, as well as ROI estimates, in healthcare.

Conclusion

An economic impact assessment of the LAGB procedures in classes II and III obese ERS members and their dependents demonstrated significant net health cost reductions and economic activity increases in Texas. This analysis indicates that providing benefits for the LAGB procedure to eligible members of the Texas ERS and their dependents is worthy of support from payer, employer, and societal perspectives.

Transparency

Declaration of funding: This study was funded by Allergan Inc.

Declaration of financial/other relationships: M.R.P. and V.G. have disclosed that they are employees of the Perryman Group, a company that received funding from Allergan in support of this study.

Appendix

The summary coefficients of the US Multi-Regional Impact Assessment System associated with measurement of the net overall impact of a program to provide laparoscopic adjustable gastric banding (LAGB) surgery benefits for severe obesity within the active membership and adult dependents of the Employees Retirement System of Texas (ERS) benefit program on the economy of Texas – gains over a 5-year period assuming a 5% acceptance rate.

Table

Explanatory variable sectors	Expenditure coefficients		Gross product coefficients
	Direct & indirect (2008 dollars)	Induced (2008 dollars)	Direct & indirect (2008 dollars)	Induced (2008 dollars)
Agricultural products & services	0.037975	0.013882	0.010325	0.003774
Forestry & fishery products	0.000959	0.000351	0.001006	0.000368
Coal mining	0.005132	0.001876	0.001482	0.000542
Crude petroleum & natural gas	0.027475	0.010043	0.006002	0.002194
Miscellaneous mining	0.000661	0.000242	0.000284	0.000104
New construction	0.000000	0.000000	0.000000	0.000000
Maintenance & repair construction	0.049652	0.018150	0.025445	0.009301
Food products & tobacco	0.077649	0.028384	0.019973	0.007301
Textile mill products	0.001021	0.000373	0.000233	0.000085
Apparel	0.014155	0.005174	0.007827	0.002861
Paper & allied products	0.012246	0.004476	0.005419	0.001981
Printing & publishing	0.017237	0.006301	0.008728	0.003190
Chemicals & petroleum refining	0.064412	0.023545	0.009754	0.003565
Rubber and leather products	0.009356	0.003420	0.004023	0.001471
Lumber products & furniture	0.003867	0.001414	0.001356	0.000496
Stone, clay, & glass products	0.005023	0.001836	0.002784	0.001018
Primary metal	0.004045	0.001479	0.001119	0.000409
Fabricated metal products	0.010232	0.003740	0.003674	0.001343
Machinery, except electrical	0.006324	0.002312	0.002535	0.000927
Electric & electronic equipment	0.006131	0.002241	0.003228	0.001180
Motor vehicles & equipment	0.004936	0.001804	0.001111	0.000406
Transport equip., except motor vehicles	0.002145	0.000784	0.000918	0.000336
Instruments & related products	0.001726	0.000631	0.000708	0.000259
Miscellaneous manufacturing	0.003958	0.001447	0.001562	0.000571
Transportation	0.058092	0.021235	0.039934	0.014598
Communication	0.060582	0.022145	0.037432	0.013683
Electric, gas, water, sanitary services	0.147545	0.053934	0.033121	0.012107
Wholesale trade	0.062359	0.022795	0.042198	0.015425
Retail trade	0.386273	0.141200	0.320081	0.117004
Finance	0.026276	0.009605	0.013878	0.005073
Insurance	0.036096	0.013195	0.021604	0.007897
Real estate	0.360292	0.131703	0.035389	0.012936
Hotels, lodging places, amusements	0.030654	0.011205	0.016145	0.005902
Personal services	0.085457	0.031238	0.053154	0.019430
Business services	0.080597	0.029462	0.047444	0.017343
Eating & drinking places	0.167597	0.061264	0.098152	0.035879
Health services	0.107012	0.039118	0.075848	0.027726
Miscellaneous services	0.076528	0.027974	0.029943	0.010945
Households	0.002439	0.000891	0.002439	0.000891

The summary coefficients of the US Multi-Regional Impact Assessment System associated with measurement of the net overall impact of a program to provide laparoscopic adjustable gastric banding (LAGB) surgery benefits for severe obesity within the active membership and adult dependents of the Employees Retirement System of Texas (ERS) benefit program on the economy of Texas – gains over a 5-year period assuming a 5% acceptance rate.

Table

Explanatory variable sectors	Personal income coefficients		Employment coefficients
	Direct & indirec (2008 dollars)	Induced (2008 dollars)	Direct & indirect (2008 dollars)	Induced (2008 dollars)
Agricultural products & services	0.007032	0.002571	0.000000	0.000000
Forestry & fishery products	0.000373	0.000136	0.000000	0.000000
Coal mining	0.001562	0.000571	0.000000	0.000000
Crude petroleum & natural gas	0.002768	0.001012	0.000000	0.000000
Miscellaneous mining	0.000167	0.000061	0.000000	0.000000
New construction	0.000000	0.000000	0.000000	0.000000
Maintenance & repair construction	0.020968	0.007665	0.000000	0.000000
Food products & tobacco	0.010203	0.003730	0.000000	0.000000
Textile mill products	0.000197	0.000072	0.000000	0.000000
Apparel	0.003966	0.001450	0.000000	0.000000
Paper & allied products	0.002450	0.000896	0.000000	0.000000
Printing & publishing	0.005697	0.002083	0.000000	0.000000
Chemicals & petroleum refining	0.004580	0.001674	0.000000	0.000000
Rubber and leather products	0.002352	0.000860	0.000000	0.000000
Lumber products & furniture	0.000967	0.000353	0.000000	0.000000
Stone, clay, & glass products	0.001456	0.000532	0.000000	0.000000
Primary metal	0.000833	0.000304	0.000000	0.000000
Fabricated metal products	0.002372	0.000867	0.000000	0.000000
Machinery, except electrical	0.001811	0.000662	0.000000	0.000000
Electric & electronic equipment	0.001930	0.000706	0.000000	0.000000
Motor vehicles & equipment	0.000722	0.000264	0.000000	0.000000
Transport equip., except motor vehicles	0.000600	0.000219	0.000000	0.000000
Instruments & related products	0.000538	0.000197	0.000000	0.000000
Miscellaneous manufacturing	0.001077	0.000394	0.000000	0.000000
Transportation	0.026411	0.009654	0.000000	0.000000
Communication	0.015981	0.005842	0.000000	0.000000
Electric, gas, water, sanitary services	0.014453	0.005283	0.000000	0.000000
Wholesale trade	0.024332	0.008894	0.000000	0.000000
Retail trade	0.191398	0.069965	0.000006	0.000002
Finance	0.008081	0.002954	0.000000	0.000000
Insurance	0.012916	0.004721	0.000000	0.000000
Real estate	0.005702	0.002084	0.000000	0.000000
Hotels, lodging places, amusements	0.010592	0.003872	0.000000	0.000000
Personal services	0.041355	0.015117	0.000001	0.000000
Business services	0.038702	0.014147	0.000001	0.000000
Eating & drinking places	0.052222	0.019089	0.000003	0.000001
Health services	0.064130	0.023442	0.000001	0.000000
Miscellaneous services	0.025958	0.009489	0.000001	0.000000
Households	0.002387	0.000873	0.000000	0.000000

Dependent variable: direct benefits from providing LAGB coverage.

Source: US Multi-Regional Impact Assessment System, The Perryman Group.