CHIP reads the wound-care note your provider wrote — typed, scanned, or photographed — and turns it into the record a payer expects: measurements, healing trajectory, coverage criteria, and the products that match them.
Wound-care practices billing Medicare · nothing leaves our server
No new charting system, no second place to type. CHIP works from the documentation your providers already produce.
PDF, scan, or a phone photo of a printed chart. Text extraction first, on-device OCR when the page is an image. Multiple visits of one patient merge into a single healing timeline.
Every finding is matched against the coverage rules that govern it, with the LCD or article named. What the note supports is marked. What it doesn't is flagged, not filled in.
Confirm the dressing or graft, and CHIP writes the clinical note for signature and carries the plan into a referral, HealKit, or NPWT order — quantities already calculated.
CHIP charts every measured visit and reads the trajectory the way a reviewer will. The software that supports a skin substitute when the record does will tell you just as plainly when it doesn't.
– Not todayArea down 39% in eight weeks of documented standard care. A wound that is healing has not failed standard care — and does not need advanced therapy.
✓ Criterion metNo measurable change across the last four weeks of documented standard care — the failed response a reviewer looks for. CHIP then checks every other criterion before anything is offered.
| Week | Wound A | Wound B |
|---|---|---|
| 0 | 100 | 100 |
| 2 | 86 | 97 |
| 4 | 74 | 96 |
| 6 | 66 | 96 |
| 8 | 61 | 95 |
Submitted as-is, this is a denial. CHIP shows it now — your denial letter, ten weeks early — and names the one line that closes the gap.
If the note doesn't document it, CHIP says it isn't documented.
An undocumented depth is not zero. An unrecorded wound bed is not clean. A venous ulcer is not assumed painful because venous ulcers usually are.
Software that quietly fills a gap produces a record that reads well and cannot be defended. CHIP does the opposite: it shows you the gap while it can still be closed, and where a finding is true but rarely written down, it asks the provider to attest it — stored separately, so an auditor can tell what was read from what a clinician asserted.
Not "AI-powered compliance." Specific policies, cited on the page, so your team can check CHIP's reasoning instead of trusting it.
CHIPLCD L33831Surgical dressingsPrimary and secondary selection by thickness and exudate, with the covered change frequency and the per-code allowance.LCD L33821 · PA A52511Negative pressure wound therapyQualifying wound type, the programme that must have failed first, and the four conditions that exclude coverage.Article A56175Non-contact ultrasound (97610)All eight qualifying criteria, and the codes it cannot share a date of service with.NCD 270.3Platelet-rich plasma (G0465)Diabetic indication, the 20-week limit, and the wound-bed condition that has to hold before it applies.CHIP reads the denial letter, answers each finding from the record, and names what is missing — before anyone writes a word. Your staff starts from a draft, not a blank page, and the treating provider signs it.
Medical necessityNot reasonable and necessaryThe record does not support the need.
Safe & effectiveInvestigational / experimentalNot proven safe and effective for this use.
Standard careFailure of standard care not shownNo documented trial of conservative care.
UnitsUnits exceed the woundMore product billed than the wound supports.
Each deadline is computed from the decision date on the letter, so the clock is on the page, not in someone's head.
Drop in a patient's complete record. CHIP reads every page and says what would pass, what is at risk and what would fail — and what to gather while there is still time.
Each line opens to the reviewer's checklist: what the record shows, why a reviewer would fail it, and the document that closes it.
A retrospective, within-wound study of your product — built from charts that already exist, generated on demand, with every exclusion printed.
CHIP reads the visit notes your customers' providers already write. For a manufacturer, that becomes real-world evidence without new enrollment or a survey: every wound your product touched, followed backwards in time, each one compared with its own standard-care weeks before your product.
48 excluded — each with its reason
Change in area
Change per week
A per-week rate sits beside every total, so a longer window cannot flatter a result. 45 of 48 wounds reduced faster on the product; the 3 that did not are reported as non-responders.
| Week | Your product | Category median |
|---|---|---|
| 0 | 100 | 100 |
| 2 | 81 | 88 |
| 4 | 64 | 78 |
| 6 | 54 | 70 |
| 8 | 47 | 64 |
| 10 | 42 | 60 |
| 12 | 39 | 57 |
Split by wound type — diabetic foot, venous leg, pressure. Comparators are other products in the same category and are never named. The comparison a payer will make anyway: make it first, from charts, with every exclusion printed.
Illustrative figures: the method is real; the numbers are not yours yet. No claim of superiority over any other product is made or implied. Patient data never leaves our server, and what a manufacturer sees is aggregated and de-identified.
Extraction, coverage checking and the clinical language model all run on our server, inside our own environment. That is an architectural property, not a policy promise — there is no outbound call to switch off.
Bring a real chart — the messier the better. The most useful demo is the one where CHIP finds something your last denial was about.