What Is the Best Software for a Non-Medical Home Care Agency in 2026?

The direct answer

The best software for a non-medical home care agency runs scheduling, EVV, and billing on the same record so that visit data matches claims exactly. Manual edits trigger state audits, and systems that don't talk to each other produce denials you won't see until payment stops.

You need a system where the caregiver's clock-in becomes the visit record, the visit record becomes the EVV submission, and the EVV submission becomes the claim—without re-entry, without export files, and without hoping three vendors stay in sync.

Why this matters now

States have moved from soft enforcement to hard edits. Under soft edits, claims with minor EVV data errors were paid with a warning. Hard edits trigger automatic denial with no grace period.

The 21st Century Cures Act mandated that states implement EVV for all Medicaid personal care services by January 1, 2020 and home health services by January 1, 2023. States that fail to comply face incremental FMAP reductions that begin at 0.25 percentage points and eventually reach 1 percent. For PCS, reductions are: 0.25% (2020), 0.5% (2021), 0.75% (2022), 1% (2023 and beyond). For HHCS: 0.25% (2023), 0.5% (2024), 0.75% (2025), 1% (2026 and beyond). That penalty pushes states to enforce strictly.

State auditors now treat a high manual edit rate as a primary fraud indicator. Pennsylvania's 15% threshold is the most publicized.

What specifically goes wrong

Forgotten clock-in. Caregiver enters the home, starts care, remembers to clock in twenty minutes later. The timestamp does not match the scheduled visit window.

GPS drift. Apartment buildings, rural addresses, and weak cell service produce coordinates that do not match the patient's registered location.

Manual edit cascades. A coordinator fixes the timestamp manually. The visit gets paid, but the agency's unmodified EVV compliance score drops, and the state flags the pattern.

Service code mismatches. Once a state enforces hard edits, claims will be automatically denied if the shift does not match the EVV data exactly, including mismatches in time, service code, or visit verification status. Agencies running scheduling, EVV, and billing in three separate systems do not see these problems until the denial comes back.

Visit data doesn't match the claim: the EVV system records a different time, location, or service than what was billed. High manual entry rates made after a missed clock-in are treated as an audit indicator. Aggregator transmission failures mean visit data captured by the EVV system never reaches the state aggregator before the claim is submitted. Eligibility lapses generate denials even when every other data element is accurate.

What the law actually requires

The Cures Act requires six data elements for every visit:

1. Type of service performed
2. Individual receiving the service
3. Date of service
4. Location of service delivery
5. Individual providing the service
6. Time the service begins and ends

CMS notes that there is no requirement to use GPS, but it is one approach for implementation.

EVV requirements do not apply when the caregiver and beneficiary live together. Personal care or home health services rendered by an individual living in the residence does not constitute an "in-home visit."

Services consisting only of instrumental activities of daily living—chore and homemaker services—are not subject to EVV requirements. Services offered in 24-hour settings are separate from the home visits subject to EVV.

How this differs by state

Each state chooses an EVV model, an aggregator, and enforcement rules. Here are three large states:

| State | Model | Aggregator | Key Dates | Specific Rules | |-------|-------|------------|-----------|----------------| | California | Open Model—providers use their own EVV system or CalEVV | Sandata | PCS: Jan 1, 2022<br>HHS: Jan 1, 2023 | Live-in caregiver exemption form required | | Texas | Provider choice with state-funded option (HHAeXchange) | TMHP Aggregator | PCS: Jan 1, 2021<br>HHS: Jan 1, 2024 | 80% EVV Usage Score required | | New York | Provider Choice Model | eMedNY | PCS: Jan 1, 2021 (with good faith extension from Jan 2020)<br>HHS: Jan 1, 2023 | Must integrate with eMedNY |

Texas has explicit compliance requirements: Program providers and FMSAs must achieve and maintain a minimum EVV Usage Score of 80% rounded to the nearest whole percentage, each state fiscal year quarter. Visit maintenance (corrections) must be completed within 95 calendar days from the date of service delivery.

California follows the Open Model as defined by DHCS, allowing providers to use either the state-run CalEVV system or a certified third-party solution, as long as it integrates with Sandata.

New York allows provider choice but requires all systems to meet NYSDOH technical specifications and successfully integrate with eMedNY. New York received a one-year good faith effort extension for sole practitioner private duty nurses (PDNs), who became subject to attestation and data submission requirements beginning January 1, 2024.

What experienced operators do

Every Friday, review the past week's visits against your claim submissions. Are there visits not billed? EVV records that don't match billing records? Fix mismatches immediately. This weekly habit catches the majority of denials before they happen.

Pre-submission checks are run against every batch, not just the ones that have obvious issues. EVV exceptions are resolved before claims are built.

Coordinators working from a compliance dashboard fix errors before they become denials. After submission, the same errors require a resubmission workflow and, in some states, additional supporting documentation.

What people commonly get wrong

Assuming manual edits are fine. In the early days of EVV, retroactive edits were common and accepted. That is no longer true. States now treat manual edits as an indicator of poor compliance or potential fraud. Pennsylvania specifically requires providers to keep manual edits below 15%, with formal non-compliance notices issued after exceeding this threshold for two consecutive quarters.

Thinking GPS is required. CMS explicitly states there is no requirement to use GPS. Address verification, geofencing, or other location methods are acceptable. GPS is useful in rural areas but problematic in apartment buildings and areas with weak cell service.

Believing EVV equals automatic claim denial. This is technically incorrect but practically irrelevant. CMS does not mandate that states deny claims for EVV mismatches. However, CMS prefers it, and nearly all states have implemented hard edits that produce automatic denials. Operating as if EVV mismatches equal denials is the correct practical approach.

Using three systems. Scheduling in one system, EVV in another, billing in a third. Each handoff introduces errors. The visit time changes. A service code gets retyped wrong. An eligibility check happens at the wrong moment. By the time you realize the claim was denied, the visit is weeks old and the caregiver does not remember the details.

What to look for in software

Single record from schedule to claim. The shift you schedule should become the EVV visit and then the claim without re-entry. If your scheduler does not write directly to your billing file, you will spend hours reconciling.

Built-in clearinghouse. Eligibility checks, claim submission, and remittance should happen in the same system. If you are downloading 835 files and importing them, you are doing work the software should do.

Exception dashboard. You need to see every visit that does not have a matching clock-in and clock-out before you build claims. Fixing exceptions after submission costs time and money.

Aggregator transmission. The system should transmit visit data directly to your state aggregator—Sandata, TMHP, eMedNY, or whichever your state uses. If the vendor says they submit to the aggregator but you have to check manually, that is not good enough.

No per-seat fees. Your coordinator count will change. Your caregiver count will definitely change. Per-seat pricing makes growth expensive and creates an incentive to share logins, which breaks audit trails.

The actual workflow

Here is the order of operations:

1. Check eligibility before you schedule. Not after the visit. Not when you build the claim. Before you put the visit on the calendar.

1. Schedule the visit with the correct service code, the correct time window, and the correct location. This record becomes the EVV visit template.

1. Caregiver clocks in and out. The EVV app captures the six required data elements and writes them to the same visit record.

1. Coordinator reviews exceptions daily. Missed clock-ins, GPS mismatches, time variances. Fix them while the visit is recent.

1. Run pre-submission checks. Before you build claims, verify that every visit has a matching EVV record and that all required fields are complete.

1. Submit to aggregator and clearinghouse simultaneously. EVV data goes to the state aggregator. Claim goes to the clearinghouse. Both should happen at the same time from the same record.

1. Monitor remittance. When the 835 comes back, denials should be rare. If they are common, you have a workflow problem, not a billing problem.

When your software runs all of it

AgenciesForge runs charting, scheduling, EVV, and billing on one record, with the clearinghouse included. See pricing here.