You've seen the demo. A vendor walks into your office, pulls up a sleek calculator, types in your total medical spend, and a 30% savings number appears. The slides after that are full of green arrows and happy faces. Everyone around the table nods. The CFO smiles.
The part nobody says out loud is that this calculator is the least transparent piece of the entire reference-based pricing (RBP) pitch. I've spent two decades inside benefits administration systems: claims repricing engines, compliance rule engines, the works. Those calculators are built to sell, not to inform.
Underneath the interface is a simple formula, and the distortion comes from the shortcuts vendors take.
How the Savings Number Gets Built
Every RBP calculator follows the same basic formula:
Projected RBP Cost = (Your Claim Volume × Assumed RBP Multiplier on Medicare) + Vendor Fees
That multiplier is usually 120% to 200% of Medicare rates. The trickery lives in how the repricing gets simulated. There are two methods:
- Method 1: averaged repricing. The tool takes your average allowed amount per claim, compares it to a generic Medicare benchmark, and returns a savings percentage. Fast and cheap, and often badly wrong.
- Method 2: procedure-level simulation. The tool maps your actual claims to CPT codes, applies Medicare fee schedules, then aggregates. Far more accurate, but it requires a full claims data export and a real pricing engine.
Most vendors use Method 1 because it's easy to build into a web calculator. In the proposals I've audited, Method 1 inflated projected savings by 5 to 10 points. Averaging hides the fact that some procedures reprice much worse than others.
Three Levers Behind the Interface
I've audited over 40 RBP proposals. Every single calculator had three silent levers behind the interface:
1. The Settlement Rate Assumption
RBP is about more than the initial payment. The real test is what happens when a provider pushes back. The calculator assumes a certain share of claims will be paid at the RBP rate without appeal, negotiation, or litigation. In the proposals I've audited, vendors defaulted that settlement assumption to 85% to 95%. For groups with heavy out-of-area utilization or large hospital systems, the number drops toward 60% to 70%. Most calculators don't let you adjust it.
2. Balance Billing and Member Out-of-Pocket Costs
Most calculators completely ignore what happens when a member gets a balance bill. Imagine a $50,000 bill: the plan pays 150% of Medicare, and the member is stuck with the rest. Those balance-billed amounts do not count toward the ACA out-of-pocket maximum, so the plan never has to cover them. The member does. The calculator pretends this never happens, and that optimism raises fiduciary questions under ERISA section 404(a).
3. Network Disruption
The calculator assumes your members will keep seeing the same doctors and that those doctors will accept the new RBP rates. In reality, some providers stop accepting the plan. If 10% of your members switch to higher-cost out-of-network providers, the calculator never shows the effect. That's a blind spot.
Garbage In, Golden Projections Out
Even the most sophisticated calculator is only as good as the data you feed it. Three common data problems:
- Claim lag: you hand over 12 months of claims that are 4-6 months old. The calculator assumes today's utilization, but your workforce might have changed: new hires, maternity leaves, chronic condition clusters. The numbers shift.
- Pharmacy carve-outs: many RBP tools only handle medical claims. If your pharmacy sits under a separate pharmacy benefit manager (PBM), the calculator may double-count total spend or miss medical drug costs that reprice differently.
- Stop-loss carve-outs: if you have a stop-loss policy with a $50,000 attachment point, the savings on claims above that threshold belong to the reinsurer, not you. The calculator often ignores this, inflating the number.
A good vendor runs the calculator on your actual claims file and gives you a sensitivity analysis. A bad vendor gives you a generic link where you type in 'Total Medical Spend' and get a rosy result.
Compliance Risks the Calculator Never Models
RBP has serious legal and regulatory risks. Yet I've never seen a calculator model any of these:
- ERISA section 404(a) exposure: when a member picks a doctor who balance-bills them and the plan does not step in, the plan's fiduciaries face questions about whether they met their duty of prudence. The calculator is silent on that.
- State rate-setting and surprise-billing rules: Maryland's all-payer hospital system means hospitals charge state-approved uniform rates, so a calculator built on national Medicare rates overstates savings there. Several states also cap or regulate what providers can charge out-of-network patients, which changes the balance-billing picture a calculator assumes away.
- ACA out-of-pocket maximums: balance-billed amounts from non-network providers do not count toward the annual out-of-pocket maximum. A member can hit the cap and still owe thousands. The calculator rarely models that member exposure.
- No Surprises Act gaps: federal rules effective January 1, 2022 ban balance billing for emergency care and for out-of-network providers at in-network facilities, but they do not reach elective out-of-network care. Members can still be balance-billed in full for planned care at an out-of-network provider, and the calculator assumes that never happens.
These aren't edge cases. They can shrink a 20% savings projection to 5%, or end it in a lawsuit.
What the Calculator Hides From Employees
The calculator counts plan savings and stops there. It never counts the cost that lands on your employees, and that cost is part of the design. In a 2025 California court ruling involving UC Davis Medical Center, a reference-based plan paid about $74,500 of a roughly $397,500 hospital bill, and the court rejected the hospital's attempt to make the plan pay the remaining balance. The member was left carrying the difference. That outcome follows from how repricing works when a hospital refuses the reference price: the unpaid portion moves onto the employee's statement, and the employee who gets that bill tells coworkers. Cost shifted onto your own workforce is not the kind of savings a CFO wants to defend in an all-hands meeting.
The Calculator Is a Sales Tool
Every element of that calculator is tuned to win you over.
The interface tells the story. You never input a settlement rate or a litigation expense. The output is a single, crisp percentage, not a range. The visualization is a big green bar shooting up next to a sad red bar, making your current plan look wasteful. There's no comparison to your existing network discount.
Vendors know the calculator is often the first thing a CFO sees. They optimize it for conversion, not accuracy.
None of this makes RBP a bad idea. It means the calculator's output is a starting point for a risk-adjusted model, not gospel.
Your Practical Checklist
Before you sign anything, demand answers to these seven questions:
- Does the calculator use procedure-level repricing or aggregate averages?
- Can I upload my actual claims data for a personalized projection?
- What is the assumed settlement rate? Can I adjust it?
- Does the model account for stop-loss reimbursement?
- Are balance billing costs and member out-of-pocket impacts included?
- Does the tool output a range (pessimistic, base, optimistic) or a single number?
- Can the vendor replicate the projection after one year of real claims experience? (I call this the retrospective honesty test.)
If a vendor balks at any of these, walk away. A pricing system should be auditable, and its projections should survive an audit.
Trust, but Verify the Algorithm
Reference-based pricing can produce real savings when executed well. The savings calculator is the vehicle that gets you onto that road. Until vendors treat their calculators as risk-assessment tools instead of sales brochures, the gap between projected and actual savings will stay wide.
Print this article. Hand it to your broker before the next RBP demo. Ask them to explain those three levers. If they can't, you'll know exactly how much trust to put in that green bar.
The author is a veteran health benefits systems consultant who has designed claims data models for Fortune 500 employers, TPAs, and health plans. He has audited over 40 RBP proposals and never met a calculator he didn't want to reverse-engineer.
This article is for general information only and is not legal, tax, or medical advice. Employers should consult their own advisors.