Currently Viewing:
The American Journal of Managed Care August 2017
Health Insurance and Racial Disparities in Pulmonary Hypertension Outcomes
Kishan S. Parikh, MD; Kathryn A. Stackhouse, MD; Stephen A. Hart, MD; Thomas M. Bashore, MD; and Richard A. Krasuski, MD
Evaluation of a Hospital-in-Home Program Implemented Among Veterans
Shubing Cai, PhD; Patricia A. Laurel, MD; Rajesh Makineni, MS; Mary Lou Marks, RN; Bruce Kinosian, MD; Ciaran S. Phibbs, PhD; and Orna Intrator, PhD
The Effect of Implementing a Care Coordination Program on Team Dynamics and the Patient Experience
Paul Di Capua, MD, MBA, MSHPM; Robin Clarke, MD, MSHS; Chi-Hong Tseng, PhD; Holly Wilhalme, MS; Renee Sednew, MPH; Kathryn M. McDonald, MM, PhD; Samuel A. Skootsky, MD; and Neil Wenger, MD, MPH
What Do Pharmaceuticals Really Cost in the Long Run?
Darius Lakdawalla, PhD; Joanna P. MacEwan, PhD; Robert Dubois, MD, PhD; Kimberly Westrich, MA; Mikel Berdud, PhD; and Adrian Towse, MA, MPhil
The Hospital Tech Laboratory: Quality Innovation in a New Era of Value-Conscious Care
Courtland K. Keteyian, MD, MBA, MPH; Brahmajee K. Nallamothu, MD, MPH; and Andrew M. Ryan, PhD
Association Between Length of Stay and Readmission for COPD
Seppo T. Rinne, MD, PhD; Meredith C. Graves, PhD; Lori A. Bastian, MD; Peter K. Lindenauer, MD; Edwin S. Wong, PhD; Paul L. Hebert, PhD; and Chuan-Fen Liu, PhD
Risk Stratification for Return Emergency Department Visits Among High-Risk Patients
Katherine E.M. Miller, MSPH; Wei Duan-Porter, MD, PhD; Karen M. Stechuchak, MS; Elizabeth Mahanna, MPH; Cynthia J. Coffman, PhD; Morris Weinberger, PhD; Courtney Harold Van Houtven, PhD; Eugene Z. Odd
Currently Reading
Cost-Effectiveness Analysis of Vagal Nerve Blocking for Morbid Obesity
Jeffrey C. Yu, AB; Bruce Wolfe, MD; Robert I. Griffiths, ScD, MS; Raul Rosenthal, MD; Daniel Cohen, MA; and Iris Lin, PhD
Geographic Variation in Medicare and the Military Healthcare System
Taiwo Adesoye, MD, MPH; Linda G. Kimsey, PhD, MSc; Stuart R. Lipsitz, SCD; Louis L. Nguyen, MD, MBA, MPH; Philip Goodney, MD; Samuel Olaiya, PhD; and Joel S. Weissman, PhD

Cost-Effectiveness Analysis of Vagal Nerve Blocking for Morbid Obesity

Jeffrey C. Yu, AB; Bruce Wolfe, MD; Robert I. Griffiths, ScD, MS; Raul Rosenthal, MD; Daniel Cohen, MA; and Iris Lin, PhD
This lifetime economic analysis demonstrates vagal nerve blocking therapy to be a cost-effective alternative to conventional therapy in class 2 and 3 obesity patients.
Health-related quality of life. In order to assess health-related quality of life, we used utility equations from Ackroyd et al that were dependent on both BMI and diabetes status.19 These utility equations expressed quality of life in terms of The EuroQOL 5 Dimensions questionnaire (EQ-5D) scoring and were used to assign utilities to patients at the start of each cycle. Quality of life scores were then calculated using these EQ-5D utility values as adjustors.

Costs. Estimates for healthcare costs were derived from Arterburn et al, who reported adult per capita total healthcare expenditure by BMI category.20 First, we plotted adult per capita total healthcare expenditures against BMI and then we performed a univariate regression analysis in which we estimated the coefficient describing the association between BMI and adult per capita total healthcare expenditure. As Arterburn et al assessed total healthcare expenditures, it was assumed that costs associated with obesity sequelae were folded into the total expenditures.20 Total healthcare expenditures were inflated from 2000 to 2015 levels using the medical component of the Consumer Price Index.

The cost of vagal nerve blocking therapy was incorporated as an initial device and installation cost of $20,000 in the base-case, and as the model allowed for assumptions around device replacement, a $17,000 vagal nerve blocking device replacement cost was also incorporated at the end of year 9 for all patients receiving vagal nerve blocking therapy. These values were obtained through communication with the product manufacturer. In sensitivity analyses, this cost parameter was varied up and down by 25%. Patients were assigned only 1 vBloc device replacement in their lifetime. The likelihood of a second device replacement was variable and dependent on several factors, including age. Given the mean baseline age of 47 years, on average, patients would be 65 at the time the second device would require replacement, and it is not expected that patients would receive another device implant at that age. For simplicity in the model, we made the assumption of 1 replacement. However, we also considered that potential changes in the device’s battery life expectancy and adoption of vagal nerve blocking therapy among younger patients may ultimately result in more than 1 replacement. Device revision surgery costs for initial device implantation and replacement were also incorporated in the model.

Diabetes remission. Probabilities for diabetes remission were estimated from rates observed in Gregg et al, which provided annual probabilities of diabetes remission for patients on an intensive lifestyle intervention (ILI), starting in various BMI categories.21 Diabetes remission rates from the Gregg et al ILI cohort were deemed appropriate for the vagal nerve blocking arm, as weight loss observed in the ILI cohort was comparable to weight loss observed for vagal nerve blocking therapy in the ReCharge trial. As Gregg et al only provided data up to 12 months, the rate of diabetes remission at the end of 12 months was carried forward for months 13 to 30. After month 30, the rate of remission was set to 0 to parallel the lack of BMI change after month 30 for vagal nerve blocking therapy patients, as the assumption was that change in BMI drove diabetes remission in the model. These estimates from Gregg et al were utilized because the ReCharge trial limited the participation of those with type 2 diabetes and there were not enough patients in the trial to assess diabetes remission.10

Device revision surgery. Patients were assigned probabilities for device revision surgery and associated costs in the first and second years after initial device implantation, based on available data from the ReCharge trial and the product manufacturer. These probabilities and costs were also applied in the first and second year after device replacement.

Mortality. Probability of death was a function of both age and BMI and was applied at the end of each cycle. We employed the methodology previously used by Campbell et al,22 in which mortality rates adjusted for BMI and age were estimated by applying BMI-specific relative risk ratios23 to age-specific all-cause mortality rates from the 2010 US life tables.24 Therefore, at any given age, patients with higher BMI had higher mortality risk. In order to obtain BMI-specific RRs, we first plotted the RR of death described by Flegal et al23 against BMI, and then performed a univariate regression analysis in which we estimated the coefficient describing the association between BMI and RR of death.

Discounting. A 3.50% annual discount rate was assumed.25 This was converted to a weekly rate and applied to utilities and costs for each weekly cycle.

Obtaining weekly rates. To accommodate the weekly cycles used in the model, the data for BMI change, total healthcare expenditures, diabetes remission rates, mortality rates, and discount rates were adjusted to weekly rates using the Declining Exponential Approximation of Life Expectancy method outlined by Beck et al.26

Model Outcomes

Model outcomes were evaluated over a lifetime horizon for the class 2 and class 3 analyses. For each comparator, cumulative total healthcare costs were collected in addition to QALYs, which are a measure of both the quality and quantity of life, and were calculated by adjusting the life-years for each patient by their quality of life, as captured by literature-based EQ-5D scoring. To compare the vagal nerve blocking and conventional therapy arms, an ICER expressed as dollars per QALY gained was calculated. The ICER was estimated by dividing the difference in costs between vagal nerve blocking and conventional therapy by the difference in QALYs between the treatment arms.

RESULTS

Base-Case Results


Base-case results were generated for both the class 2 and class 3 analyses under a lifetime horizon and are presented in Table 2. Mean discounted lifetime total healthcare costs were $123,607 for vagal nerve blocking and $96,141 for conventional therapy in class 2 patients and $129,183 for vagal nerve blocking and $102,259 for conventional therapy in class 3 patients. Mean discounted QALYs were 9.27 for vagal nerve blocking therapy and 7.68 for conventional therapy in class 2 patients and 7.92 for vagal nerve blocking and 6.68 for conventional therapy in class 3 patients. The ICER for vagal nerve blocking therapy versus conventional therapy in class 2 patients was $17,274 per QALY gained and $21,713 per QALY gained in class 3 patients.

Sensitivity Analyses

To test the robustness of the base case results, multiple 1-way sensitivity analyses were conducted. Model parameters tested included costs, utility, change in BMI, proportion of class 3 patients with diabetes, diabetes remission, and mortality rates. For most parameters, the sensitivity analysis tested inputs at 75% and 125% of the base-case value. The proportion of class 3 patients with diabetes was tested using 25 percentage points below and above the base-case value. Sensitivity analysis parameters are shown in Table 3.

A series of 1-way sensitivity analyses were performed for patients with both class 2 and class 3 obesity and results are presented in tornado diagrams in Figure 2. Among both the class 2 and class 3 populations, the top 3 drivers of variation in the ICER were the EQ-5D utility values applied for patients without diabetes, BMI trajectory, and the procedural costs associated with initial vBloc device implantation. These sensitivity analyses showed the class 2 and class 3 ICERs to be robust to reasonable variation in model inputs. We performed an additional analysis, incorporating a transient utility decrement of 25% for 2 weeks after both initial device implantation and replacement for the vagal nerve blocking arm. This transient disutility was extracted from a cost-effective analysis by Comay et al of the endoscopic Stretta procedure,27 which is a minimally invasive endoscopic procedure used in gastroesophageal reflux disease. In this additional analysis, ICERs were $17,507 and $21,820 per QALY gained in class 2 and class 3, respectively. For all sensitivity scenarios, the upper limits of the ICER estimates remained below the conventionally accepted $50,000 cost-effectiveness threshold in the United States.28 Probabilistic sensitivity analysis was also performed and results were consistent with reported findings.

DISCUSSION

In this analysis, Monte Carlo modeling techniques were used to evaluate the cost-effectiveness of vagal nerve blocking therapy versus conventional therapy over a lifetime horizon of class 2 patients with diabetes and class 3 patients with or without diabetes. Assuming a $20,000 initial cost for the vBloc device and installation, the estimated ICER for vagal nerve blocking versus conventional therapy was $16,907 per QALY gained among class 2 patients and $21,424 per QALY gained among class 3 patients. Sensitivity analyses showed the cost-effectiveness results to be robust to reasonable variations in

model assumptions.

Our model design was based on previous bariatric surgery cost-effectiveness models published in the peer-reviewed literature. Prior analyses have stratified outcomes by BMI, such as in Campbell et al, which compared laparoscopic gastric banding (LAGB) and laparoscopic Roux-en-Y gastric bypass (LRYGB) to no treatment.22 Other analyses have also incorporated diabetes and diabetes remission, such as in the analysis by Hoerger et al, which also compared LAGB and LRYGB to no treatment.29 In our model design, we stratified outcomes by class 2 and class 3 obesity and incorporated diabetes and diabetes remission in the model structure.

Campbell et al observed ICERs for LAGB and LYRGB (vs no treatment) to be $5400 and $5600 per QALY gained, respectively22; and Hoerger at al observed ICERs for LAGB and LRYGB (vs no surgery) to be $11,000 to $13,000 per QALY gained and $7000 to $12,000 per QALY gained, respectively.29 Although the results of our study show that vagal nerve blocking therapy is cost-effective, the ICERs estimated from our model are higher than ICERs published for LAGB and LYRGB.

Limitations

Because of the lack of BMI data from the ReCharge trial past 30 months, BMI in the vagal nerve blocking therapy arm was assumed to remain constant beyond 30 months in order to conservatively model out a lifetime horizon. By applying this assumption in the model, additional cost savings or losses from weight loss past 30 months may not have been fully captured. More long-term data may further elucidate the value of vagal nerve blocking, and the ASMBS has encouraged participation in clinical studies and a prospective collection of outcomes for this therapy.12

We did not include AEs associated with the vagal nerve blocking procedure in the base case analysis as commonly observed AEs from the trial were not high-cost events (eg, pain in the neuroregulator site, heartburn/dyspepsia, and other types of pain). Although utility values used in the model were dependent on diabetes status, healthcare expenditures applied in the model were not explicitly dependent on diabetes status, but on BMI-specific average overall healthcare expenditures in the general obese population. Even though this allows for a conservative approach, it is possible that additional cost savings or losses experienced by patients with diabetes remission undergoing vagal nerve blocking therapy may not be fully captured in the model.

CONCLUSIONS

 
Copyright AJMC 2006-2017 Clinical Care Targeted Communications Group, LLC. All Rights Reserved.
x
Welcome the the new and improved AJMC.com, the premier managed market network. Tell us about yourself so that we can serve you better.
Sign Up
×

Sign In

Not a member? Sign up now!