Your clinical manager is burning 12 to 18 hours a week on chart review, and it still isn’t enough. OASIS-E1 items get scored off by one point and the case-mix weight collapses. A LUPA slips through because nobody caught that the visit plan dropped below threshold on day 14. A face-to-face encounter note sits in the chart with a signature date that predates the certifying event, and you don’t find out until the ADR letter arrives eight months later demanding $340,000 back. Meanwhile HHVBP is redistributing up to 5% of your Medicare payment based on quality data your team is entering while exhausted at 9 p.m. Every vendor in home health now has “AI” on the slide deck, and none of them will tell you plainly what it actually reads, what it actually flags, and what it quietly ignores.
This is written for agency owners, administrators, and clinical directors running Medicare-certified home health — roughly 40 to 2,000 episodes a year — who are either sitting on Axxess or WellSky and wondering if they’re leaving money on the table, or shopping a switch and refusing to eat another six-figure implementation on faith. You should already know what PDGM, LUPA, RAP/NOA, and HHVBP mean in your own P&L; we don’t re-teach the payment model. Out of scope: hospice, private duty, home care non-medical, therapy-only agencies, and anything outside US Medicare home health. This is not an implementation manual for either platform and it is not legal or compliance counsel.
Straight talk on the AI: it is genuinely good at the boring, high-volume, pattern-shaped work — surfacing internal contradictions between the OASIS and the visit notes, catching missing signatures and date sequencing problems, flagging episodes drifting toward LUPA while there’s still time to act, and ranking which charts a human should open first. It is unreliable at clinical judgment calls, at anything requiring the nurse’s actual observation of the patient, and at novel or edge-case scoring where the guidance itself is ambiguous. It will hand you confident wrong answers on M1800-series functional scoring more often than any vendor demo suggests. Final OASIS sign-off, therapy necessity decisions, F2F clinical adequacy, and anything going into an ADR packet stay under a licensed human’s name. Treat AI chart review as a triage layer that decides where your QA hours go — not as a replacement for the QA hours.
What This Guide Covers
- A plain-language definition of what “AI chart review” actually does in home health today — and the specific capabilities vendors imply but don’t deliver
- Deep-dive assessments of Axxess and its AI review layer, and WellSky Home Health with its full analytics stack, evaluated on what they do for your charts rather than what’s on the feature grid
- A complete head-to-head comparison matrix across accuracy, workflow fit, reporting depth, support responsiveness, and total cost
- Results from running matched de-identified episodes through both systems, so you can see where each one caught errors and where each one missed them
- How each platform performs on OASIS-E1 item accuracy — including which item groups produce the most false confidence
- PDGM case-mix handling and whether either system reliably surfaces clinical grouping and comorbidity adjustment opportunities you’re currently missing
- LUPA risk flagging compared side by side, with attention to whether the alert arrives early enough to change the outcome
- What each platform contributes to your HHVBP star performance, and the measures where software help is largely cosmetic
- Face-to-face validation and RAP/NOA timeliness safeguards — the two failure modes that quietly generate the largest takebacks
- Denial defense in practice: audit trail quality, documentation chain integrity, and how usable each system’s export is when you’re assembling an actual ADR response
- Real pricing structures, per-episode versus per-user models, the hidden line items, and a QA-hours-saved ROI model you can run against your own volume
- Implementation timelines, data migration risk, and a HIPAA and BAA due diligence checklist for vendor AI features — including the questions about model training on your PHI that most agencies never ask
- Documented pitfalls and case studies from agencies that adopted, over-trusted, or abandoned these tools
- A direct verdict segmented by agency size and census, plus what’s credibly coming in 2027 and what to avoid buying early
Instant online access immediately after checkout — the full guide is available the moment your payment clears. No upsell, no subscription, no follow-on course.
Reviews
There are no reviews yet.