Roger Healthcare Review: Home Health AI Charting, Intake and Coding

Roger’s charting product brings referral context into the visit, captures the encounter and accepts additional dictation afterward. The vendor describes both OASIS and routine documentation, with prompts for missing information and support for medication pictures and wound images. Its integration page describes proprietary robotic process automation, or RPA, for transferring information into the EHR.

The important evaluation is the whole chart, not just a fluent narrative. Ask a nurse and therapist to review the same demonstration from their own disciplines. Count fields requiring correction, unsupported statements removed and time spent finding the relevant source. A short draft-generation time is only useful if the resulting record is easy to inspect, correct and finalize.

Roger also publishes a dedicated intake product. It describes referral-document extraction, eligibility checks, a shared intake dashboard and creation of pre-populated patient records in the EHR. This gives an admissions team a reason to evaluate Roger beyond the scribe. The public workflow includes identifying details such as face-to-face documentation, orders and signatures.

Use a referral with a duplicate attachment, an unclear payer response and a missing signature in the demo. Follow each exception until a person resolves it, then inspect the record created in the EHR. Establish who can authorize acceptance, correct extracted information and prevent duplicate patients. The purchasing question is whether the team can reliably handle difficult packets, rather than how quickly a clean sample becomes a summary.

Roger’s coding product generates diagnosis recommendations from provider documentation and presents explanations with editing controls. Its public page describes a patient overview, documentation-gap flags and primary and secondary diagnosis suggestions. These are meaningful capabilities to compare with a separate coding application or an outsourced service; the page’s reimbursement claims are vendor marketing, not results established by this review.

Bring the coding lead into procurement. Ask them to reject a suggested diagnosis, follow its supporting evidence and request additional documentation. Agree on responsibility for the final code set and any unresolved queries before comparing prices. A recommendation should remain useful when it lowers an unsupported code or leaves a question open. Evaluate defensibility and correction time alongside throughput, without treating a higher reimbursement estimate as proof of better coding.

Roger’s published positioning centers on connecting to the EHR an agency already uses. Our inference is that the first buying decision should be which team owns the initial rollout: admissions, field clinicians or coding. Starting with a named workflow makes it possible to measure whether the implementation removes work or moves it elsewhere.

Request a written map of the supported EHR version, forms, access requirements and recovery process after an EHR update. For RPA-based chart transfer, demonstrate a changed field, interrupted session and retry rather than assuming every connection is a supported API. Establish an exception queue that your staff can see. In the pilot, track completed and signed charts, rework, duplicate entry and the age of unresolved transfers. Expand to another module only after the initial team can operate it predictably.