Methodology

How we score every price

A price by itself is just a number. We rate every price on five signals so you know which ones to trust.

1 · Methodology trust (max 0.30)

Hospitals encode rates in several ways — as exact dollars, as a percentage of billed charges, or as algorithmic contract logic. Exact-dollar fee schedules score highest. “Percent of billed” rates score lowest because the final cost depends on a billed-charge number we don't know in advance.

2 · Cross-validation (max 0.25)

We compare each hospital's rate to the same payer's Transparency-in-Coverage file. If both sides agree within 10% (the Berkowitz et al. 2024 baseline), the rate is cross-verified. Same payer brand + plan-name fuzzy match within band yields a strict verification.

3 · Volume of supporting rates (max 0.20)

One payer's contract for one CPT is a single data point. Twenty payers' contracts for the same CPT at the same hospital is a much stronger signal. We use a log scale so the 5th rate matters less than the 1st.

4 · Freshness of the MRF snapshot (max 0.15)

Hospitals update their Machine-Readable Files monthly under CMS rules. We score full marks for ≤90 days, half for ≤180 days, less beyond.

5 · Plausibility (max 0.10)

We compare each rate to the published Medicare benchmark for the same CPT. Anything between 0.5× and 8× Medicare is plausible; outside that band is usually a parsing error or a mis-encoded methodology.

What this doesn't tell you (yet)

The negotiated rate is the dollar amount the hospital and the payer agreed on. It's not what you, the patient, will owe — your plan's deductible, coinsurance, and out-of-pocket maximum all change the final number. That patient-specific cost is what we're working on next.