How to Build Home Health Agency Software (2026)

Building software for a home health agency is harder than it looks from the outside. The scheduling problem alone is complex: match caregivers to patients by clinical skill, geographic area, continuity preference, and travel time. Add federal EVV mandates, Medicare OASIS assessments, PDGM episode billing, and 50 states each running their own EVV aggregator, and you have one of the most compliance-dense builds in healthcare IT. This article breaks down every major system you need to build, the timeline and cost, and the one technical problem that will slow you down more than any other.

According to the National Association for Home Care and Hospice, over 12,000 home health agencies operate in the United States, serving more than 5 million patients annually. That's a large market with severe software fragmentation.

What this software does

Home health agency software runs the operational backbone of agencies that send caregivers and clinicians to patients' homes. That covers skilled nursing visits (RNs and LPNs), physical and occupational therapy, personal care aides, and companion care. The software must handle patient intake, caregiver scheduling, visit verification, clinical documentation, billing, and HR compliance for caregiver credentials.

There are two fundamentally different types of home health agencies. Clinical home health agencies bill Medicare and Medicaid for skilled care. They need OASIS assessments and episode-based Medicare billing. Non-medical private duty agencies provide companion care and personal care aides billed by the hour to Medicaid or private pay. Both need EVV. The clinical agencies need considerably more.

"The integration between Electronic Visit Verification and state Medicaid Management Information Systems remains one of the most technically fragmented challenges in home care IT. Vendors that assume one aggregator works across states consistently underestimate build scope by 40-60%." -- Mark Sharp, Senior Analyst, Forrester Research, Home Care Technology Outlook 2025

Core features: MVP vs. full product

CMS data shows Medicare home health spending exceeded $16 billion in 2024. Every dollar flows through billing systems that depend on correctly collected OASIS data and properly submitted EDI 837 claims. Get the software wrong and the money stops.

An MVP for a non-medical private duty agency needs four things: patient and caregiver profiles, caregiver scheduling with shift assignment, EVV GPS clock-in/out on a mobile app, and basic invoicing for private pay or Medicaid hourly billing. That is achievable in 18-24 weeks for $180K-$300K.

A full clinical home health platform adds: OASIS assessment collection and electronic submission, care plan management (CMS-485), physician order tracking, clinical visit documentation (nursing notes, therapy notes, medication administration records, wound care records), supervisor review and clinical sign-off workflow, Medicare PDGM episode billing with 837I claims, denials management, clearinghouse integration, caregiver credential tracking, and background check integration.

The difference between the MVP and the full product is not just features. It is regulatory complexity. OASIS alone requires collecting 100+ structured data fields per patient assessment and submitting them to your state's OASIS data system in a specific format. Get this wrong and your Medicare reimbursement stops.

The architecture

The system has four distinct interfaces. The office coordinator web dashboard handles patient intake, caregiver scheduling, job assignments, clinical supervision, billing, and reporting. This is a React application connected to a Node.js API backed by PostgreSQL.

The caregiver mobile app (React Native) handles EVV clock-in/out with GPS, visit note entry, care plan access, and communication with the office. The GPS coordinates at clock-in and clock-out are the core EVV data points. The app must work offline for areas with poor connectivity and sync when the connection returns. Store pending EVV records locally with a queue and retry on reconnect.

The clinician mobile app (or a clinician-specific view of the same app) handles structured OASIS data entry, nursing and therapy assessments, medication administration records, and digital signature capture. OASIS data is complex enough that clinicians need a guided form with field-level validation, not a free-text note.

The billing module sits on the backend. It reads completed and supervisor-approved visit documentation, generates claim files in EDI 837 format, submits via clearinghouse APIs, and tracks claim status responses. Clearinghouses like Waystar return 835 remittance files when a claim is paid or denied. Parse those and update each claim record accordingly.

Store all clinical documentation in PostgreSQL with full audit trails. HIPAA requires access logs retained for 6 years. Encrypt PHI at rest (AES-256) and in transit (TLS 1.2+). Use AWS S3 with server-side encryption for any uploaded documents (wound photos, signed care plans). Sign Business Associate Agreements with every vendor that touches patient data.

The hardest technical challenge

According to Sandata Technologies, their EVV platform alone serves 28 states, yet each state has configured different submission rules, aggregator identifiers, and timing requirements. No two are identical.

The hardest problem is EVV state compliance, and it gets harder as you expand to new states.

The 21st Century Cures Act required EVV for all Medicaid-funded home care, but each state chose its own aggregator. Sandata serves many states. HHAeXchange serves others. Some states built their own. Each aggregator has its own submission format, API specification, authentication method, timing requirements, and exception handling rules.

An agency operating in 10 states needs 10 different EVV integrations. If you build those integrations as 10 separate code paths, you will spend the rest of your engineering life maintaining them. Every time an aggregator changes its API, you fix one path. Every time you add a state, you write another path from scratch.

The right approach is a modular EVV adapter layer. Define one internal interface: record a visit event with caregiver ID, patient ID, service code, location coordinates, and timestamp. Then write a separate adapter for each state aggregator that implements that interface. The scheduler and mobile app talk only to the internal interface. The adapter handles translation into whatever format the state requires.

Test every adapter against the state's sandbox environment before going live. States provide test credentials and validation rules. A submission that passes your own validation but fails the state aggregator's rules means the visit is unverified and not reimbursable. Run full end-to-end tests with real sandbox calls before you put real patient visits through any new state integration.

Build timeline and cost

RaftLabs has scoped multiple home care platforms. The pattern is consistent: agencies that treat EVV as a last-mile feature, rather than a day-one architectural requirement, spend 60-80% of their post-launch budget patching it. Build it first.

The MVP (scheduling, EVV, basic documentation) takes 18-24 weeks and costs $180K-$300K. That covers the coordinator web dashboard, the caregiver mobile app with GPS EVV, basic visit note entry, and hourly billing for private duty or simple Medicaid billing. This is the right scope for a non-medical agency or for a clinical agency that wants to prove the product before investing in Medicare billing.

The full platform (OASIS assessment, Medicare PDGM billing, clearinghouse integration, physician order tracking, credential compliance) takes 30-40 weeks and costs $380K-$600K. Infrastructure at launch runs $3K-$10K per month and scales with patient census and visit volume.

The biggest cost variables are the number of states you need EVV integrations for (add 3-6 weeks and $20K-$40K per new state aggregator beyond the first), whether you need native iOS and Android apps or can start with React Native, and how deeply you need to support the Medicare billing workflow versus starting with private pay.

Build vs. buy

A 2024 Definitive Healthcare survey found that 38% of home health agency operators cite software limitations as a top barrier to expansion. That gap is where custom builds win.

The established platforms cover most agency needs. Homecare Homebase is the enterprise standard for clinical agencies billing Medicare, at $500-$1,500+ per month depending on patient census. ClearCare (now part of WellSky) focuses on private duty at $200-$500 per month. MatrixCare serves skilled nursing and home health at $400-$800+ per month. Alayacare is common for larger agencies at $300-$700 per month.

Build custom when you have one of three situations. First: you are building a multi-state MSO (management services organization) that rolls up dozens of agencies and needs unified reporting, standardized workflows, and consolidated billing across all of them. Existing platforms are built for single agencies. Second: your state's EVV aggregator has poor API support from existing vendors, leaving your agency scrambling for a workaround. Third: your care model is unusual enough that existing platforms handle it badly, such as pediatric home health, behavioral health home care, or a hybrid skilled-plus-companion model under one roof.

Combined SaaS costs across 50 agencies at $500/month each hit $25K/month, or $300K/year. At that scale, a custom build amortizes in 18-24 months and you control the product roadmap.

Tech stack

React Native for the caregiver and clinician mobile apps. GPS is natively available and the offline-first architecture is straightforward with AsyncStorage for queued EVV records. React for the office coordinator and billing dashboard. Node.js with PostgreSQL for the backend. PostgreSQL handles relational data well: patients, caregivers, visits, schedules, and claims are all highly relational.

EVV submission uses the state aggregator's API or SFTP upload depending on what the aggregator supports. Sandata and HHAeXchange both have documented APIs. Some state-specific aggregators only accept SFTP file uploads, so build both mechanisms into your adapter layer.

Google Maps for route optimization when scheduling visit sequences. The nearest-neighbor heuristic is good enough for MVP. OR-Tools from Google is available if you need true optimization across a full day's schedule for multiple caregivers. Checkr for caregiver background checks via API. AWS S3 for document storage. Waystar or TriZetto for EDI 837 claim submission and 835 remittance parsing. Stripe for private pay patient billing.

Closing

The agencies that will pay for custom software are the ones that have outgrown the off-the-shelf tools: multi-state operators, PE roll-ups, and agencies with care models that existing platforms were not designed for. If you are building for that market, budget for the EVV adapter layer on day one. The schedule, the documentation, the billing -- all of those are solvable engineering problems. The EVV state compliance matrix is where underfunded builds fall apart. Design for it from the start.

RaftLabs builds HIPAA-compliant healthcare software for home care agencies, digital health founders, and clinical staffing platforms. See our healthcare software development work or discuss your build.