Health+Benefits the June 2025 issue

Benchmarking Bariatric Surgery Costs for a Center of Excellence (COE) Program

Providing insights to employers through a data-driven approach.
By Cheryl Matochik Posted on May 28, 2025

The employer sought an independent total cost of care benchmark to ensure the vendor’s negotiated rates were competitive and transparent across different care settings.

Goals

  • Validate COE vendor pricing.
  • Benchmark rates against national and regional norms.
  • Enable strategic review of COE performance.

Scope of Work

  • Identify procedure trigger code and applicable CPT codes related to bariatric surgery.
  • Price procedures individually and aggregate them into bundled episode costs.
  • Stratify benchmarks by site of service (inpatient, outpatient, ambulatory surgical centers).

Methodology

1. Episode Trigger Codes. To define the bariatric surgery population, we analyzed claims provided by the COE vendor with a primary diagnosis of morbid obesity (E6601). Relevant CPT codes for bariatric procedures were extracted to build the core episode definition. Each CPT code was individually priced to create the overall “bundle” price, allowing flexibility for customization.

2. Site of Service Stratification. Recognizing that surgeries were performed in various settings, we developed separate bundled prices for:

  • Inpatient hospitals
  • Outpatient hospital departments
  • Ambulatory surgical centers (ASCs)

3. Bundle Construction Approach. Rather than using a prepackaged bundle, each CPT code was priced individually and summed to create the full episode cost. This method offers: X Transparency: Clients can see how each component contributes to total costs. X Flexibility: Clients can customize the bundle by adding or removing services based on case needs.

4. Pricing Sources. Benchmark pricing was sourced from publicly available national machine-readable file data, specifically for:

  • UnitedHealthcare
  • Aetna
  • Cigna
  • BlueCard network

Two sets of benchmarks were created:

  • National pricing averages
  • Tennessee-specific pricing, given the employer’s heavy population footprint in that state

Data Limitations and Recommendations

Vendor File Limitations. The COE vendor’s claims file lacked individual case identifiers, making it difficult to link related claims to episodes of care. Without these identifiers, it is challenging to: X Track which services belong to a specific patient’s surgery episode X Accurately validate the full bundle composition

Recommended Improvements. To enhance future benchmarking and validation efforts, we recommended the COE vendor:

  • Include a case ID or person ID to link claims at the episode level
  • Include billing NPIs to improve provider-level rate accuracy when matching to machine-readable file data

Conclusion

Through a data-driven and flexible benchmarking approach, the employer will gain critical insight into its COE vendor’s pricing structure to validate whether the negotiated rates are truly competitive, identify opportunities for cost optimization, and strengthen accountability for COE program performance.

Cheryl Matochik Managing Director/Partner, Third Horizon Read More

More in Health+Benefits

Employers Are Flocking to Level Funding
Health+Benefits Employers Are Flocking to Level Funding
These offshoots of self-funded health plans offer cost savings with less risk.
Health+Benefits Recovery @ Work
Opioid use disorder still costs U.S. employers billions of dollars, but they can...
The New Fiduciaries
Health+Benefits The New Fiduciaries
States are testing the waters on adding health insurance roles to their list of ...
Going Direct to Cut Costs
Health+Benefits Going Direct to Cut Costs
Q&A with Tom Malczewski, Head of Direct Contracting Partnerships and Vice Presid...
Data Access Opportunities for Self-Insured Employers
Health+Benefits Data Access Opportunities for Self-Insured Employers
Federal healthcare price transparency rules enable employers...
Quality of Price Transparency Data
Health+Benefits Quality of Price Transparency Data
All price data is not created equal.