Every EHR eventually gets asked for billing — it's the stickiest, highest-margin module you can ship, and the one whose absence pushes customers toward all-in-one competitors. There are four ways to add claims capability, and the right one depends on five questions, not on any vendor's pitch. Here's the honest map, with published numbers.
The Four Options
| Option | What it is | Cost model | You control denials/UX? |
|---|---|---|---|
| 1. Traditional clearinghouse | Waystar, Availity, Optum — SFTP/batch + newer APIs, sales-led contracts | Quote-based; per-EHR agreements | You build everything above the pipe |
| 2. API-first clearinghouse | Stedi, Claim.MD — REST/JSON, self-serve | Published: ~$0.10–$0.30/claim, no minimums | You build everything above the pipe |
| 3. Embedded RCM-as-a-service | Candid Health, Apero — billing operations behind an API | Subscription / % of collections | Limited — their workflow, their data exhaust |
| 4. Build with a partner | Specialist engineering firm builds YOUR billing layer on option 1/2 rails | Project + your per-claim costs | Full — you own the code and the data |
Option 1: Traditional Clearinghouses
Maximum payer reach and enterprise credibility — Waystar alone touches thousands of payer connections. The realities from having shipped one: contracts are per-EHR and sales-led, connectivity is often SFTP batch, enrollment runs weeks per payer, and the acknowledgement handling (999 vs 277CA) is entirely your problem. Our Waystar integration guide walks the whole loop step by step.
Option 2: API-First Clearinghouses
The transport layer, modernized: REST/JSON, real documentation, self-serve signup, and published pricing — Stedi lists claims at $0.10–$0.30 and eligibility at $0.08–$0.30 per transaction with no monthly minimum; Claim.MD similar. What they deliberately don't solve: scrubbing rules, denial workqueues, posting logic — the product layer is still yours to build.
Option 3: Embedded RCM Platforms
Fastest to revenue: billing operations arrive as a service. The trade-offs compound at scale — percentage-of-collections pricing grows with your GMV, the denial workflow is their roadmap not yours, and the remit/denial data that would train your rules and AI features accumulates in their warehouse. Right answer for some stages; expensive moat-transfer for others.
Option 4: Build (Usually with a Partner)
Realistic scope, sourced: a basic billing module runs 3–5 months with 3–4 engineers; a fuller platform 8–12 months — plus clearinghouse credentialing and payer-enrollment calendar time (see our enrollment timeline guide). The part every “build” article omits: you still need a clearinghouse underneath — nobody connects to thousands of payers directly. And one architecture lesson the industry paid $2.5B/week to learn: the 2024 Change Healthcare outage took down roughly 40% of US claims flow through a single clearinghouse (HFMA) — owning your claims layer above a swappable clearinghouse is now a resilience argument, not just a control argument.
The Five Questions That Settle It
- Volume: at $0.10–$0.30/claim, transaction fees are noise — the fork is % of collections vs owning the workflow.
- Specialty complexity: single-specialty professional claims → API rails nearly suffice; institutional/multi-specialty → scrubbing depth dominates.
- Denials control: if denial-management UX is part of your product's value (for an EHR, it is), renting it caps your roadmap.
- Data ownership: remit and denial data is what trains your future rules and AI features.
- Margin: at scale, owned billing is the highest-margin module an EHR sells.
The Hybrid Most Teams Land On
Own the claims data model and the denials UX; rent the pipes. API-first clearinghouse underneath, your product on top — built either by your team over 6–12 months or with a specialist partner who has shipped the acknowledgement chains and posting engines before (that's the difference between two sprints of surprises and a plan — see the seven traps). It's faster than a solo build and, unlike RCM rental, everything you learn accrues to you.
We're the partner in that sentence: claims engines, posting, and denial workflows built inside EHRs, on whichever rails fit. If you're at this decision, an hour mapping your five answers onto the four options is the cheapest de-risking available. See our RCM software development services or talk to our team.
