Skip to content
Reeve
ReeveFor Texas home care agenciesRead only

Texas home care margin recovery.

You delivered the visits.

In Texas, an EVV visit that does not match the claim on all six data elements gets denied, and a claim past 95 days cannot be filed at all. Reeve reads your own billing export, right in your browser, and shows you the authorized, delivered, billed, and paid columns side by side, so the gaps have a record attached.

The four numbers

Every dollar of personal care in Texas passes through four states of being, and they do not always agree.

Authorized

The payer approved a number of units. In Texas that is your STAR+PLUS, STAR Kids, or STAR Health managed care authorization, or a fee-for-service auth under Primary Home Care, Community Attendant Services, Family Care, or a waiver like CLASS or DBMD. The clock and the ceiling both live here.

Delivered

The people who deliver the care worked the visit and it landed in Electronic Visit Verification. Texas runs EVV through a state Aggregator operated by TMHP, with HHAeXchange as the state-funded system or an approved proprietary system. Either way, the visit becomes an accepted EVV transaction, or it does not.

Billed

A claim went out to TMHP or the managed care organization with a Medicaid ID, date of service, NPI or API, a HCPCS code, its modifiers, and a unit count. Where the unit sits is the thing to get right in Texas, because the two halves of the program do not agree. Texas fee for service bills attendant services on state assigned G codes, and the HHSC schedule prints G0702 at an hourly rate. Texas managed care bills the same service on S5125 in fifteen minute units. A reconciliation that resolves a Texas book on the code alone is wrong by a factor of four in one direction or the other, so Reeve resolves the unit from the program before it counts anything.

Paid

The claim matched, cleared its other edits, and remitted. Or it hit a mismatch code and denied.

The leaks live in the gaps between those four columns, and Texas has two hard rules that turn a gap into lost money: EVV claims matching is a hard edit that denies any claim that does not match an accepted visit on all six data elements, and the 95-day filing window closes quietly. That window is not only a TMHP rule. Uniform Managed Care Manual 2.0 puts the same 95 days from the date of service on every Texas Medicaid managed care organization, with 120 days from the plan disposition date to appeal, https://www.hhs.texas.gov/handbooks/uniform-managed-care-manual, so a STAR+PLUS line is timed rather than left unclocked. Texas also fixes prompt pay at 30 days for a clean electronic claim under Insurance Code Chapter 843 and 28 TAC 21.2807, and Reeve prints that as a disclosure rather than as a finding.

  • 01
    Delivered but never billed
  • 02
    EVV visit that does not match the claim
  • 03
    Billed fewer units than the visit shows
  • 04
    Denied for a modifier or ID mismatch
  • 05
    Aged past the 95-day filing window
Illustrative findings

Every number is illustrative and synthetic. Because your own export may not carry a rate for every code, findings on this page are expressed in hours, units, and records to check against your own rate sheet, never in dollars Reeve made up.

Finding A
EVV06 unit mismatch, silent shortfall

14 claims paid clean this quarter where the accepted EVV visit shows more delivered quarter-hour units than the claim billed. Example: a visit accepted for 18 units of S5125, claim billed 16. The claim matched and paid, so nothing flagged, but 2 units per visit across 14 visits is 28 units of delivered care that never turned into a claim line. Records attached, ready to check against your contracted rate.

Finding B
EVV05 service mismatch, denied and not reworked

16 denied claim lines carrying an EVV05 result, a HCPCS or modifier disagreement between the visit and the claim. Example: the visit coded with one modifier, the claim submitted with another. Twelve of the sixteen are inside the 120-day appeal window and reworkable today. The other four are past it.

Finding C
Authorization ceiling versus delivery

2 clients where delivered and accepted EVV units exceed the authorized units on file. This is not recoverable, and Reeve will not pretend it is. It surfaces separately as over-delivery, care the people who deliver it gave that was never authorized, so you can see the exposure and fix the auth going forward rather than quote it as money owed.

Finding D
Timely-filing risk, the aging shelf

23 delivered-and-accepted EVV visits with no matching claim found in the export, sorted by date of service against the 95-day clock. 11 are still inside the window and billable now. 8 are between 95 and 365 days, appeal-or-lose territory. 4 are past 365 and gone. The list is ranked by days remaining, so the billable ones are at the top.

Finding E
EVV04 provider or attendant ID mismatch

5 visits where the attendant identifier on the accepted EVV transaction does not line up with the provider identifier on the claim. Each one is a denial waiting to happen or one that already did, with the specific visit and claim record paired so your biller can see exactly which field to correct.

The blind spot

Why an annual audit or the EMR alone misses this.

01

A once-a-year audit looks back over twelve months. In Texas the money has a 95-day clock. By the time an annual review names a denied line, the appeal window may be shut and the filing window long past. An annual audit is history. The 120-day appeal window is not waiting for it.

02

The EMR is not built to catch it either. Your EMR knows what it billed. The EVV Aggregator knows what visit was accepted. The remittance knows what paid. Those three live in three places, and the mismatch is in the seam between them. An EVV01 clean match shows green in your billing screen even when the visit carried more units than the claim, because the claim matched the units it declared, not the units delivered.

03

The gap only shows when you put the accepted EVV transaction, the claim, and the remittance in the same view and read them together. That is the one thing Reeve does. It reads across the seam.

That is what Reeve reads for in Texas: what HHAeXchange accepted, what the authorization allowed, what you billed, and what actually paid, read together rather than audited for rate.

Questions

Plain answers, on the record.

Straight answers on how the Texas Margin Review works, what it touches, and what it will not claim.

No made up numbers. Reeve prices findings off the rate on your own export or the rate your own remittances actually paid most often, and where it cannot find a rate in your data it gives you the hours, the units, and the exact records instead of a dollar figure. A rate Reeve has not verified is a rate Reeve will not quote.

No. The Margin Review runs in your own browser on an export you provide. Client names, member identifiers, and the names of the people who deliver the care are replaced with a coded reference locally before any math runs. Nothing with a name on it is uploaded or sent anywhere. Reeve reads. It does not write to your EMR, does not file a claim, and does not move money.

The EVV claims matching hard edit. Since April 1, 2024, an EVV-required claim that does not match an accepted EVV visit transaction on all six data elements is denied for payment. Reeve lines up the visit, the claim, and the remittance and shows you where those six fields disagree, using the same result-code logic the Aggregator uses.

The personal care and attendant world. Fee-for-service Primary Home Care, Community Attendant Services, Family Care, and waivers like CLASS and DBMD, plus managed care LTSS under STAR+PLUS, STAR Kids, and STAR Health. Common codes include S5125 and T1019, both billed in 15-minute units, which is where a lot of the small unit mismatches hide.

Keep it. Whether you use the state-funded HHAeXchange system, the TMHP Aggregator flow, or an approved proprietary EVV system, Reeve does not replace any of it. It reads the export those systems produce and reconciles it against your claims and remittances. It sits above your stack, read-only.

Reeve sorts findings by the clock. Delivered visits still inside 95 days are billable now and go to the top. Denials inside the 120-day appeal window are workable. Lines past 365 days are gone and Reeve marks them as gone rather than dressing them up. You see what is recoverable and what is a lesson, clearly separated.

You do, yourself, for free. There is no sales call required to try it, no automated nightly scanning of your systems, and no team quietly holding your data. You provide an export, the review runs locally, and you keep the findings whether or not we ever work together.

Two lanes, priced separately

Collect is money you never captured. Cover is money a payer can still take back.

Reeve reports the two separately and never adds them together, because only one of them is yours to go and get. The Margin Review reads both on your own export and costs nothing.

Margin Review
Free

One pass over your own export, in your browser. The findings are yours to keep, with no obligation.

Collect
$750 per branch per month

The recovery lane. Care you delivered and never billed, units short of what was authorized, lines paid under the published rate.

Cover
$1,000 per branch per month

Everything in Collect, plus the exposure lane. Retired codes, authorizations at the end of their period, care delivered past what was approved.

Month to month, no annual contract. Read-only in every tier. Run the free review.

Start with a Margin Review

Try it yourself. Run a free Margin Review.

No sales call. No account to create. Export your data, open the review in your browser, and see your authorized, delivered, billed, and paid columns lined up in about the time it takes to read this page.

  • 01

    Read-only. Reeve reads your export and does not write to your EMR, does not file a claim, does not move money.

  • 02

    Names coded locally. Names and member identifiers are replaced with a coded reference in your browser before any analysis runs.

  • 03

    Your file is processed in your browser, and no name is written into any output. No upload of identified data, no server holding your client list.

  • 04

    Yours to keep. The findings are yours whether or not we ever work together.

Reeve reads your billing read-only and replaces its names with coded references before any math runs. Every example figure is illustrative and synthetic, method rather than results, and Reeve makes no guarantee of recovery.