The home health plan of care has to prove, how the 485 drifts out of step with the OASIS and visit notes over a 60 day episode, and how AI review keeps them aligned.
The plan of care is the spine of the home health episode
Most documents in a home health chart describe a moment. The OASIS captures an assessment. A visit note captures a visit. The plan of care is different. It describes the whole episode, in advance, and then has to remain true for sixty days while the patient changes.
That is a harder job than it sounds, and it is why the plan of care tends to be the document that quietly stops matching everything around it.
What the plan of care is actually asserting
Strip away the form and the 485 makes a set of specific claims.
That this patient needs these services. That they need them at this frequency. That a physician reviewed and certified that judgement. That the clinical picture in the OASIS supports it. That the goals are achievable and measurable. That the orders, including medications and precautions, are current.
Every one of those claims is checkable against another document in the chart. Which is exactly how reviewers approach it.
How a plan of care drifts
Nobody sets out to let the 485 go stale. It drifts because the patient moves and the paperwork takes a moment to catch up.
Frequency drift. The plan says skilled nursing twice weekly for four weeks, then weekly. The schedule shifts after a hospitalisation, visits are added, and the delivered pattern no longer matches the certified one. Both are defensible. The mismatch is not.
Order drift. A medication changes mid-episode. The medication profile updates. The orders on the plan of care do not.
Goal drift. Goals written at start of care describe a patient who no longer exists eight weeks later, either because they improved or because they did not. A plan still aiming at a goal the patient met a month ago reads as a document nobody revisited.
Discipline drift. Therapy is added by verbal order and delivered properly, but the plan of care is never updated to reflect it. The visits happened. The authorisation for them is missing from the controlling document.
Signature timing. The content is right and the signature arrives late. On review, timing is part of the claim.
None of these is a clinical failure. Each is an administrative gap that becomes a documentation problem, then a payment problem.
Why the plan of care is hard to audit by hand
Checking a 485 properly means holding several documents open at once.
You read the certified frequency, then the visit log, and compare. You read the orders, then the medication profile, and compare. You read the goals, then the most recent OASIS and the therapy notes, and compare. You confirm the physician signature exists, is attributable and is dated appropriately.
That is slow work, and it is slow in a way that does not scale. Which is why most agencies check the plan of care thoroughly at start of care, lightly at recertification, and properly only when something has already gone wrong.
Where AI review changes the economics
Murphi treats the plan of care as a cross-document check rather than a form to proofread.
The plan of care and 485 sit in the live report library alongside coding, OASIS, PDGM, ADRs, visit notes, CTI and wound care, with more than 35 reports running across home health and hospice. Gaps and missing orders are surfaced rather than waiting to be discovered.
The ambient side closes the other half. When Murphi drafts documentation from a visit, QA flags inconsistencies against the plan of care, the physical therapy evaluation and the medications before anything is signed. A note that contradicts the controlling document is caught while the clinician who wrote it is still in the loop, and the OASIS and plan of care are synced back to the EHR together rather than drifting apart.
Pre-bill, the checks are explicit. Physician orders and the 485 are confirmed as signed. Face to face is confirmed. Homebound status is confirmed. The payment group is validated against the OASIS that produced it.
Signature handling is part of the same chain. Plans of care can be sent for physician e-signature from Murphi, signed on a phone, tablet or computer, with a time-stamped audit trail, or faxed to a physician's office with delivery confirmation saved. The lag between a correct plan of care and a signed one is often just logistics, and logistics is the easiest part of this to fix.
The human stays in the loop
Murphi drafts and flags. A certified coder or clinician approves every change before anything is written back to the EHR. Not a sample of changes. Every one. Murphi reports 98 percent accuracy with AI plus human review.
For a document that a physician certifies and an agency is paid against, that boundary is not negotiable. The people with professional responsibility for the plan of care are the people who approve what it says.
← All articles