About this site
VerifiedCareData answers one question: what do you actually need to buy before surgery, and what will it cost you? Not what a brand paid to put in front of you, and not a list scraped from other lists.
Who writes it
Erin Rose is a healthcare price-transparency analyst. She builds the Medicare coverage, fee-schedule and device-safety datasets this site runs on, and reviews every recommendation against them before it ships.
She is not a clinician, and this site never pretends otherwise. It publishes equipment, cost and coverage facts. It does not tell you whether to have surgery, when it is safe to bear weight, how to manage pain, or anything about medication — those belong to your surgeon and your physical therapist, and the pages say so where the line gets close.
Corrections, disputes and “you got this wrong” all go to erinrose451@gmail.com. Corrections get noted on the page, not quietly overwritten.
How we know what we know
There is no hands-on product testing on this site, and we do not claim any. What there is instead is three independent evidence layers, cross-referenced — each one checkable by you, with a link.
Government data
What Medicare actually pays, and whether it pays at all. The CMS DMEPOS fee schedule, the HCPCS Level II quarterly file, national utilization by code, DME MAC policy articles, and the PDAC product-classification list — queried directly from CMS, never through a third party's summary.
Named clinical instructions
What hospitals actually send patients home with. Every clinical claim on a page quotes a named institution's own discharge instructions or patient-education material — UNC Otolaryngology, Memorial Hermann, UC Davis and others — and links to it. Where sources disagree, the page says they disagree rather than picking one.
A corpus of first-person patient accounts
What people who had the surgery say they actually used. We read 4,513 posts from 683 different people across 560 discussion threads, and count how many independently raise each item. Counts are per person, so a prolific poster counts once. A recommendation here has to survive that count — and where almost nobody raises an item, the page says so.
The corpus is built per procedure, because the answers differ sharply between them. A sling is raised by 77% of shoulder patients and by nobody at all in the other two; a walker by 50% of hip patients and 6% of shoulder patients; a cold-therapy machine by 42% of knee patients and 12% of hip patients. Reading one procedure's accounts onto another is exactly the error this layer exists to catch, so each has its own:
- knee replacement — 265 people, 1,768 posts
- hip replacement — 277 people, 1,993 posts
- shoulder surgery — 141 people, 752 posts
- Other procedures — not yet. Those pages say so rather than implying a sample they do not have.
What we do not do is have a machine decide whether someone was recommending a thing or dismissing it. We tried it, checked it against the posts, and it misread plain English: “something you didn’t need for the hip recovery, but might want for this one’ came back as a rejection. So mentions are counted; verdicts are read.
None of those layers is sufficient alone. Medicare data tells you what is covered but not what helps. Hospital instructions tell you what is clinically sensible but name almost no products. Patient accounts tell you what people reached for at 2am but include plenty of things that did not help anyone. The recommendation is what survives all three.
What we will not do
- No clinical advice. No weight-bearing timelines, no rehab protocols, no wound care, no medication.
- No star ratings or scores. A number invented by us is not evidence.
- No paid placement, ever. No brand has paid to appear here and none can.
- No invented facts. If a number cannot be traced to a source on the page, it gets cut rather than estimated. Where we know a gap exists, the page discloses the gap.
How the site makes money
Some product links are Amazon affiliate links, which pay us a small commission if you buy. That is the whole business model — there is no advertising, no sponsorship, no lead generation, and no data sold.
Three rules keep it from bending the recommendations:
- Coverage pages carry no affiliate links at all. The
/medicare/and/fsa-hsa/pages — the ones people arrive at with money questions — are commercially clean by design, as is every research dataset. - “Borrow it” and “you don't need this” are real answers and appear on the pages, including for items we could have linked.
- Every pick names its own flaw. If we could not find a real one, the pick is not researched enough to publish.
How to check us
The methodology page lists every data source, refresh cadence and honesty rule in detail. The underlying datasets are published under CC BY 4.0 at /research/ — free, no account, no affiliate parameters — so you can check the arithmetic yourself rather than taking our word for it.