Source hierarchy
GlobalMed ranks sources and always prefers the highest available tier for a given price.
- P0: the hospital's official machine-readable price transparency file.
- P1: the hospital's own price transparency page or published price list.
- P2: public government data, such as CMS releases.
- P3: a written price from the hospital billing office.
- P4: other public sources, used only when nothing above is available.
One internal code per test
Hospital files describe the same test with different billing codes and hundreds of text variations. GlobalMed maps each one onto a single internal code so that a comparison always puts the same test on both sides. Mapping confidence is recorded per row, and a low-confidence mapping keeps a price out of the public catalog.
GlobalMed does not redistribute the AMA's CPT code descriptions. Internal codes are GlobalMed's own, with mapping handled at the data layer.
The publication gate
Before a price reaches the catalog it must clear two thresholds at the same time: a source quality score of at least 70 out of 100, and a mapping confidence of at least 0.80. A price that clears one but not the other stays out of the public comparison and goes to review instead. The verification date and the source URL travel with the row.
Freshness policy and its current limit
GlobalMed's data policy marks a price as possibly out of date once it is more than 180 days old, and excludes prices older than 365 days from automatic comparison. That policy is not yet enforced in the comparison query layer. Separately, a price reaches the public catalog only if it carries a source link and a verification status, and a China price read from a government catalog is shown as pending confirmation until the hospital confirms it. That is stated plainly here rather than buried, because a price with unclear provenance is worse than no price.
The practical rule for a reader is the same in both cases: treat every figure as an estimate, and get the final number from the hospital in writing before committing to anything.
What GlobalMed will not do
GlobalMed does not diagnose, does not recommend or rule out treatment, does not prescribe, does not predict disease, and does not publish hospital rankings, doctor rankings or success-rate predictions. Those are out of scope by design, not by omission.
Where a comparison produces a negative saving, the result is shown as negative. A comparison tool that cannot say 'this does not save you money' is not a comparison tool.