Pennsylvania is a hard state to reconcile. It did not mandate one rigid EVV vendor. It left the model open and then split it down the middle. A single agency can be feeding visits into Sandata on the fee-for-service side and into HHAeXchange for three different Community HealthChoices managed-care plans at the same time. Each plan brings its own configuration, its own authorization rules, and its own idea of what a clean claim looks like. For a thin-margin personal-assistance book, the reconciliation load across those systems is where the margin goes.
This page is about where money leaks in Pennsylvania specifically. The Sandata and HHAeXchange split, the three CHC MCOs, the 85% compliance enforcement. Catching it means reading your data against each payer's rules instead of trusting that a verified visit got paid.
Pennsylvania Medicaid home-care, at a glance
- Personal-care programs
- Community HealthChoices (CHC), statewide managed long-term services and supports since 2020; Personal Assistance Services (PAS) under the CHC Waiver; participant-directed “Services My Way”
- EVV model
- “Open Choice.” Providers may use the state system or an alternate that interfaces with the aggregator. But the system splits: Sandata on the fee-for-service side, HHAeXchange for the CHC MCOs
- CHC plans
- Three MCOs run EVV through HHAeXchange: AmeriHealth Caritas, PA Health & Wellness, and UPMC Community HealthChoices
- Compliance rule
- From January 1, 2025, providers must meet an 85% EVV compliance threshold (no more than 15% manual edits); from 2026, two consecutive quarters over 15% triggers a formal notice and a Corrective Action Plan
One open model, two aggregators
Pennsylvania uses an Open Choice EVV model: an agency can use the state's free system or an alternate third-party system that interfaces with the state aggregator. What “the aggregator” means in practice is where it gets complicated. On the fee-for-service side, the state EVV system is Sandata. But all three Community HealthChoices MCOs (AmeriHealth Caritas, PA Health & Wellness, and UPMC) selected HHAeXchange for their EVV and claims. So a CHC provider is reconciling Sandata for FFS and HHAeXchange for managed care, and within HHAeXchange, three separate plan configurations.
That split is the defining Pennsylvania reconciliation headache. A provider running its own EVV system has to configure feeds individually to each contracted MCO's HHAeXchange instance, and the state's own guidance is to bill only after a success response appears on the aggregation report. A visit can be captured perfectly and still stall because it did not post a success to the right plan's portal.
The 85% rule made manual edits expensive
Pennsylvania's EVV system went live for personal care in 2019 and expanded to home health care services in 2022, with home-health claims denying for dates of service on or after January 1, 2024 without a matching EVV visit. Then the state added a data-quality standard on top. From January 1, 2025, providers must hit an 85% EVV compliance threshold, no more than 15% manual edits. From 2026, the state issues alerts when a provider's manual-edit rate exceeds 15% for a quarter, and two consecutive quarters over the line triggers a formal notice of noncompliance and a required Corrective Action Plan.
The practical effect is that the manual fixes agencies have always relied on to rescue a missed clock-in are now both a billing risk and a compliance risk. A visit patched by hand still has to match the authorization. Too many patches put the agency's standing, and its claims, under scrutiny.
Where the margin actually leaks in Pennsylvania
From the way Pennsylvania splits EVV across Sandata and three HHAeXchange plan instances, the recoverable losses cluster in a few predictable places:
- Cross-system reconciliation gaps. Visits that post a success on one system, whether Sandata FFS or one MCO’s HHAeXchange, but never reconcile against the right payer’s claim and authorization.
- Per-MCO configuration errors. A feed misconfigured to one of the three CHC plans’ HHAeXchange instances repeats the same denial across every claim to that plan.
- Authorization drift across plans. A member moving between fee-for-service and a CHC plan, or between CHC plans, carries different authorization rules and unit limits. Claims outside the current authorization deny.
- Manual-edit exposure. Visits rescued by hand that still fail to match the authorization, plus the compliance risk of crossing the 15% manual-edit line for two quarters.
- Silent underpayments. Claims that pay below the contracted or authorized amount across any of the four payer paths. They surface only when payment received is compared to payment expected, line by line.
None of these are visible from the scheduling view. The schedule says the visit happened; one portal says it posted a success. It is only when you reconcile the EVV transactions against each payer's claim lines, authorizations, and remittances that the gap appears.
Why a read-only recovery layer is the right tool for this
Reeve is built for exactly this kind of cross-system reconciliation. It sits read-only over whatever EMR and EVV export an agency already runs, whether that's WellSky, AxisCare, HHAeXchange, AlayaCare, or anything else, and compares what was delivered against what each payer authorized against what was actually paid. For a Pennsylvania agency, that means lining up the Sandata FFS transactions, the three CHC plans' HHAeXchange data, the claim lines, and the prior authorizations, then surfacing every place they fail to reconcile: the cross-system gaps, the per-MCO configuration errors, the authorization drift, and the silent underpayments.
Reeve is read-only and neutral across every EMR and every payer, so it has no stake in which system you run, and it never writes to your billing workflow without your control. You get a ranked list of recoverable dollars with the reason attached: the unreconciled visit, the misconfigured plan feed, the lapsed authorization. The ones still inside each payer's filing window are the ones you can rebill now.
This is the same engine described across the rest of the site. For the mechanics of how EVV gaps become denials, see EVV billing for home care. For the broader map of revenue loss, see where home-care margin leaks. And for the authorization side, see home-care authorization tracking.
What the free Pennsylvania Margin Teardown does
The way to find out whether the Sandata and HHAeXchange split is draining your margin is to look. On a real, de-identified slice of your own data, before you spend a dollar. The Margin Teardown is a one-time, read-only read of where margin is leaking in your book: the cross-system gaps, the per-MCO configuration errors, the authorization drift, and the underpayments. It is free, and it is yours to keep whether or not you ever work with Reeve. It carries the same 3×-or-free guarantee the rest of the engine does. If Reeve does not surface at least three times its monthly fee in recoverable margin you agree is real, you do not pay.
A free, de-identified Margin Teardown reconciles your EVV, authorizations, and claims and shows you exactly what slipped. Read-only. Yours to keep.