Revenue Cycle Management in Healthcare: The 2026 Guide to Rules, Automation and Denial Prevention

Key Takeaways
-
Revenue cycle management (RCM) covers every step between scheduling a patient and collecting the last dollar for the visit: eligibility, authorization, coding, claims, payments, denials and patient billing.
-
Denials are the most expensive failure point. Hospitals reported that nearly 15% of claims to private payers were initially denied, and fighting a denial cost an average of $43.84 per claim.
-
Most denials trace back to data and rules: eligibility, missing authorization, coding edits, missing information and payer-specific requirements. Those are problems a rules layer can catch before a claim leaves the building.
-
Machine learning adds prediction: which claims are likely to be denied, which accounts to work first, which appeals are worth filing. The working pattern is a model that scores and rules that decide what happens next.
-
Billing and claims software is not a medical device under federal law, but RCM still operates under HIPAA, CMS rules and a growing set of state rules on automated decisions.
-
2026 brought new prior authorization timelines and denial-reason requirements for many payers, and a prior authorization model in traditional Medicare in six states. Payer rules change faster than ever, so the ability to update rules quickly matters.
What Is Revenue Cycle Management in Healthcare?
Revenue cycle management is the set of processes a healthcare organization uses to get paid for the care it provides. It starts before the patient arrives, when the organization confirms who the patient is and who will pay. It ends when the account balance reaches zero, either through payer payments, patient payments, or a documented write-off.
RCM is usually divided into three parts. The front end covers scheduling, registration, eligibility verification, prior authorization and patient financial clearance. The middle covers charge capture, clinical documentation and coding. The back end covers claim submission, payment posting, denial management, appeals and patient collections.
Each step depends on the one before it. A wrong insurance ID entered at registration becomes a rejected claim three weeks later. A missing authorization becomes a denial that takes hours of staff time to appeal. That is why the most effective RCM programs focus on getting decisions right early, rather than fixing them after payment fails.
What Are the Stages of the Healthcare Revenue Cycle?
| Stage | What happens | Typical decisions |
|---|---|---|
| 1. Scheduling and pre-registration | Appointment booked; demographic and insurance data collected | Is the data complete? Which payer is primary? |
| 2. Eligibility and benefits verification | Coverage checked with the payer, usually through an electronic eligibility transaction | Is the patient covered on the date of service? What are the copay, deductible and coinsurance? |
| 3. Prior authorization | Approval requested for services that need it | Does this payer require authorization for this service, site and diagnosis? |
| 4. Patient estimate and financial clearance | Expected patient responsibility calculated; payment options offered | What will the patient owe? Is a Good Faith Estimate required? |
| 5. Charge capture | Services documented and turned into charges | Were all billable services captured? |
| 6. Coding | Diagnoses and procedures coded (ICD-10-CM, CPT, HCPCS) | Do codes match documentation? Do they pass coding edits? |
| 7. Claim scrubbing and submission | Claim checked against edits and payer rules, then sent through a clearinghouse | Will this claim pass payer edits? Is anything missing? |
| 8. Payment posting | Remittance advice received and posted | Was the claim paid as contracted? |
| 9. Denial management and appeals | Denials analyzed, corrected and resubmitted or appealed | Is this denial worth appealing? Who should work it? |
| 10. Patient billing and collections | Statements sent; payment plans set up | Which payment plan fits? When to send to collections? |
| 11. Reporting and analytics | Performance measured across the cycle | Where do we lose revenue, and why? |
Most of these stages involve a decision that follows rules: payer policies, coding guidelines, contract terms, regulatory requirements and the organization’s own financial policies.
Why Is Revenue Cycle Management Getting Harder?
Denials are high and expensive. In Premier’s survey of 516 hospitals covering 2022 claims, nearly 15% of claims submitted to private payers were initially denied. 54.3% of those denials were eventually paid, but only after rounds of appeals, and providers spent an average of $43.84 per claim fighting a denial. Premier estimated total hospital spending on denials at about $19.7 billion a year, roughly $10.6 billion of it on claims that were ultimately paid (Premier, March 2024).
More organizations see double-digit denial rates. In Experian Health’s State of Claims 2025 survey of 250 revenue cycle leaders, more than 40% said at least one in ten claims is denied, up 11% since 2022. Half named missing or inaccurate claims data as the top cause of rising denials, and 35% cited prior authorization (Experian Health, November 2025).
Payer behavior varies widely. KFF’s analysis of CMS data found that insurers on HealthCare.gov denied 19% of in-network claims in 2024, about 85 million of 451 million claims. Insurer-level denial rates ranged from 3% to 36%, and fewer than 1% of denied claims were appealed (KFF, March 2026).
Automation is incomplete. The 2025 CAQH Index estimates that an additional $21 billion in savings remains available through further automation of administrative transactions between providers and health plans (CAQH Index 2025).
Rules change constantly. ICD-10-CM codes update every year, CPT codes update every year, CMS updates National Correct Coding Initiative (NCCI) edits quarterly, and every commercial payer publishes its own policies on authorization, medical necessity, modifiers and documentation. A revenue cycle team is, in practice, maintaining thousands of rules from dozens of sources.
What Causes Claim Denials, and Which Can Be Prevented?
| Denial cause | Example | Can a rule catch it before submission? |
|---|---|---|
| Eligibility and coverage | Coverage terminated before date of service; wrong payer billed as primary | Yes, if eligibility is checked close to the date of service and results are applied by rules |
| Missing or invalid prior authorization | Service required authorization; none on file, or authorization expired | Yes: payer × service × site rules flag missing or mismatched authorizations |
| Coding edits | Procedure-to-procedure conflict under NCCI; units above medically unlikely edit limits | Yes: coding edits are explicit rules |
| Missing or inaccurate data | Missing referring provider, invalid subscriber ID, wrong place of service | Yes: completeness and format rules |
| Medical necessity | Diagnosis does not support the procedure under payer policy or Medicare coverage determinations | Partly: rules can check diagnosis–procedure combinations; clinical documentation review stays with people |
| Timely filing | Claim submitted after the payer’s filing limit | Yes: deadline rules by payer |
| Duplicate claims | Same service billed twice | Yes: duplicate detection rules |
| Coordination of benefits | Secondary payer billed without primary payment information | Yes: payer sequence rules |
| Bundling and contract terms | Service included in a global or bundled payment | Mostly: contract rules, if contracts are modeled |
| Non-covered or excluded services | Service excluded under the patient’s plan | Partly: requires benefit details from eligibility response |
The pattern is clear. Most denial causes are failures to apply a known rule at the right moment. The rule existed, in a payer manual, a coding guideline or a contract, but it was not checked before the claim went out.
What Changed in 2026 for Prior Authorization and Claims?
CMS Interoperability and Prior Authorization Final Rule (CMS-0057-F). Medicare Advantage organizations, state Medicaid and CHIP fee-for-service programs, Medicaid managed care plans, CHIP managed care entities and Qualified Health Plan issuers on the federally facilitated exchanges must, starting in 2026, send prior authorization decisions within 72 hours for expedited requests and seven calendar days for standard requests, give a specific reason for denied prior authorization requests, and publicly report prior authorization metrics each year. API requirements for prior authorization and data exchange generally apply from January 1, 2027 (CMS, January 2024).
WISeR Model in traditional Medicare. CMS’s Wasteful and Inappropriate Service Reduction (WISeR) Model launched on January 1, 2026 in Arizona, New Jersey, Ohio, Oklahoma, Texas and Washington. Technology companies participating in the model use AI and machine learning tools to perform prior authorization and pre-payment review for selected services, such as nerve stimulators, epidural steroid injections and skin substitutes. Participation is mandatory for providers in those states for the covered services (DLA Piper, January 2026). For providers, that means a new authorization workflow, with its own service list, for one payer.
Limits on algorithmic coverage decisions. CMS has told Medicare Advantage plans that an algorithm or AI tool may assist coverage decisions, but the decision must be based on the individual patient’s circumstances, and “an algorithm that determines coverage based on a larger data set instead of the individual patient’s medical history” would not comply (CMS FAQ, February 6, 2024). California’s SB 1120, in effect since January 1, 2025, requires that a denial, delay or modification of care based on medical necessity be decided by a licensed physician or qualified health care provider (XiFin, 2025).
Patient estimates. Under the No Surprises Act, providers must give uninsured and self-pay patients a Good Faith Estimate of expected charges for scheduled services. Estimate logic depends on the same charge and contract rules as billing.
Security. HIPAA’s Privacy and Security Rules apply to every system that handles protected health information. HHS proposed significant updates to the Security Rule in January 2025; organizations should track the status of that rulemaking.
The direction is consistent: faster decisions, explicit reasons, and human accountability for anything that affects access to care. All three reward organizations that keep their rules explicit, current and logged.
Which RCM Decisions Should Use Rules, and Which Should Use Machine Learning?
| Decision | Rules | Machine learning | Typical combination |
|---|---|---|---|
| Eligibility interpretation | Map eligibility responses to coverage status and patient responsibility | — | Rules |
| Prior authorization requirement | Payer × service × site × diagnosis requirements | Predict likely approval to prioritize submissions | Rules decide whether authorization is needed; model helps prioritize |
| Patient estimate | Contract rates, benefits, deductible status | Predict likelihood of payment to offer plans | Rules calculate; model informs payment plan offers |
| Coding validation | NCCI edits, payer-specific modifier rules | Suggest codes from documentation for coder review | Model suggests; rules validate; coder decides |
| Claim scrubbing | Completeness, format, payer edits, timely filing | Predict denial risk for claims that pass all edits | Rules block known errors; model flags hidden risk for review |
| Work queue prioritization | Deadlines and dollar thresholds | Predict collectability and effort | Model ranks; rules set hard deadlines |
| Denial management | Map denial codes to next actions | Predict appeal success | Rules route by denial reason; model ranks which appeals to file first |
| Underpayment detection | Contract terms per payer | Detect unusual payment patterns | Rules compare with contract; model finds outliers |
| Payer-side utilization management | Coverage criteria and policy | Flag cases for clinical review | Model and rules triage; clinicians decide any denial |
A useful test: if the payer, an auditor or a patient asks “why?”, can you answer by pointing to a policy, a code edit or a contract term? If yes, that step belongs in rules. If the step is an estimate, it belongs in a model, and the action taken on that estimate belongs in rules.
How Do Rules and Machine Learning Work Together in Revenue Cycle Management?
Three integration patterns cover most RCM use cases.
1. The model scores, the rules decide. A denial prediction model estimates the probability that a claim will be denied. Rules turn that score into an action: submit, hold for coder review, request missing documentation, or route to a specialist queue. The thresholds belong to revenue cycle leadership and can change without retraining the model.
2. Rules as guardrails. Some actions should never happen automatically, whatever a model says. On the provider side, that might mean no automatic write-off above a dollar amount, or no claim submitted with a code the model suggested but a coder has not confirmed. On the payer side, CMS guidance and state laws such as California’s SB 1120 make human review of medical necessity denials a requirement. Guardrail rules make those limits explicit and testable.
3. Rules filter, then the model ranks. Rules first remove claims with known errors and route them for correction. The model then ranks the remaining work, for example which accounts in accounts receivable are most likely to be collected with one more call.
Two practices make any pattern safer. Log the model version and the rule version with every decision, so you can reconstruct it later. And define a fallback path in rules for when the model is unavailable or its inputs are missing, so claims keep moving.
For the general version of this pattern, see Rules Engines vs. Machine Learning.
What Does a Pre-Submission Claim Decision Look Like?
A simplified flow for an outpatient claim before it leaves for the clearinghouse:
| Step | What happens | Owner |
|---|---|---|
| 1. Data | Claim data from the practice management system, eligibility response, authorization record | Patient access, billing |
| 2. Completeness rules | Required fields, valid identifiers, place of service, referring provider | Billing |
| 3. Coding edits | NCCI procedure-to-procedure and medically unlikely edits; payer modifier rules | Coding |
| 4. Payer rules | Authorization required and on file for this payer, service and site; timely filing window | Revenue cycle |
| 5. Model score | Denial risk model returns 0.62 | Data science |
| 6. Decision rules | All edits passed; risk above 0.50 and charge above $1,000 → hold for coder review; otherwise submit | Revenue cycle leadership |
| 7. Record | Inputs, rule versions, model version and outcome logged | IT, compliance |
Thresholds are illustrative. Steps 2–4 catch the errors that are certain. Step 5 finds the risk that rules cannot see. Step 6 is where the organization decides how much risk to accept, and it can change that decision in hours.
What Does a Payer-Specific Rule Look Like?
Payer rules are often stored as PDFs, spreadsheets and staff knowledge. In a rules engine, the same requirements become a decision table:
| Payer | Service (CPT range) | Site of service | Authorization required | Additional requirement |
|---|---|---|---|---|
| Payer A (commercial) | Advanced imaging | Outpatient hospital | Yes | Authorization number on claim |
| Payer A (commercial) | Advanced imaging | Freestanding imaging center | Yes | — |
| Payer B (Medicare Advantage) | Advanced imaging | Any | Yes | Decision within plan timeline |
| Payer C (Medicaid managed care) | Advanced imaging | Any | Yes, unless emergency | Emergency indicator if applicable |
| Payer D (commercial) | Advanced imaging | Freestanding imaging center | No | — |
Payers and rules are illustrative. Each row has a source (the payer policy and its effective date), an owner and a version. When a payer changes its policy, the team updates the row, tests it against recent claims for that payer, and publishes the new version on the effective date. Every later claim decision references the version that was in force.
Is Revenue Cycle Software Regulated by the FDA?
Generally, no. Under section 520(o)(1)(A) of the Federal Food, Drug, and Cosmetic Act, added by the 21st Century Cures Act, a software function intended “for administrative support of a health care facility, including the processing and maintenance of financial records, claims or billing information, appointment schedules, business analytics, information about patient populations, admissions, practice and inventory management, analysis of historical claims data to predict future utilization or cost-effectiveness, determination of health benefit eligibility, population health management, and laboratory workflow” is not a medical device (FDA guidance, September 2019).
That covers the core of RCM: eligibility, billing, claims, analytics on claims data and administrative workflow. Pre-market submissions such as a 510(k) do not apply to these functions.
The boundary matters when software moves from administration into clinical care. A tool that analyzes clinical data to recommend a diagnosis or treatment falls under a different analysis, including FDA’s clinical decision support guidance, revised in January 2026. For that side, see Rules Engines in Healthcare: Supporting Care Decisions Without Replacing the Clinician.
Outside FDA rules, RCM software still operates under HIPAA, CMS program rules, payer contracts and state laws on automated decisions. Whether a specific product is a device is a regulatory question for its developer and counsel.
Where Does a Rules Engine Fit in the RCM Technology Stack?
| Layer | What it does | Typical systems |
|---|---|---|
| EHR and practice management | Clinical documentation, scheduling, registration, charges | EHR and PM systems |
| Eligibility and authorization tools | Electronic eligibility checks, authorization submission and tracking | Payer portals, clearinghouse services, authorization platforms |
| Coding tools | Code assignment and validation, computer-assisted coding | Encoders, coding software |
| Claim scrubber and clearinghouse | Formats claims, applies edits, transmits to payers, returns remittances | Clearinghouses and claim editing software |
| Rules layer | Holds the organization’s payer rules, authorization matrices, financial policies and decision thresholds; combines them with model scores | Business rules engine |
| Analytics and ML | Denial prediction, collectability scoring, underpayment detection | Data platforms, data science tools |
| Patient financial engagement | Estimates, statements, payment plans | Patient billing and payment tools |
Many RCM platforms include rules of their own, and clearinghouses apply standard edits. A separate rules layer is useful when the organization needs rules that are specific to its payers, contracts and policies, wants business teams to own those rules, and wants one set of rules to serve several systems.
Should You Build or Buy RCM Rules Automation?
Most organizations start with the rules their PM system, EHR or clearinghouse provides. That works for standard edits. It becomes limiting when:
- payer-specific rules pile up in custom scripts and spreadsheets;
- every rule change waits for a vendor release or an IT ticket;
- nobody can show all active rules for a payer in one place;
- testing a change means waiting to see what gets denied;
- the same rule must be applied in several systems, such as scheduling, authorization and billing.
A dedicated rules engine moves that logic out of application code without replacing the EHR or the clearinghouse. IT keeps ownership of the platform and integrations; revenue cycle teams own the rules. This removes backlog from the IT queue and gives IT more time for architecture and data quality.
RCM vendors face the same choice from the other side. A vendor serving many clients needs client- and payer-specific rules without forking its code base. A rules layer lets each client’s rules live as configuration, versioned and tested separately.
What Should You Look for in a Rules Engine for RCM?
- Business ownership. Can revenue cycle analysts read, change and test payer rules without writing code?
- Decision tables for payer matrices. Can authorization requirements, modifiers and filing limits be stored as tables with effective dates?
- Testing on historical claims. Can you run a rule change against last month’s claims and see which would have been held, corrected or submitted?
- Versioning and audit trail. Can you show which rule version applied to a claim decided months ago, and who approved it?
- Model support. Can your own denial prediction models run inside the decision, with rules around them?
- Performance. Can the engine check claims in real time at registration and before submission, at peak volume?
- Integration. REST API or SDK access from PM, EHR, clearinghouse and RCM platforms.
- Security and deployment. HIPAA-aligned deployment options, including on-premise or private cloud, single sign-on for business users and authenticated APIs.
- Visibility. Can you see which rules fire most often and which never fire?
How Do You Implement Rules and Automation in Your Revenue Cycle?
- Start with your denial data. Group the last 6–12 months of denials by reason code, payer and service. Pick the three to five causes that cost the most and are most preventable.
- Find the rule behind each cause. For each, identify the payer policy, coding edit or internal policy that would have prevented it, and where it should be checked: registration, authorization, coding or pre-submission.
- Model the rules as decision tables. One table per decision, with a named owner, a source and an effective date.
- Test on historical claims. Run past claims through the new rules. Measure how many denied claims the rules would have caught, and how many clean claims they would have held unnecessarily.
- Run in parallel. For a few weeks, let the rules flag claims without blocking them, and compare flags with real outcomes.
- Add a model where rules run out. Once known errors are caught, add a denial prediction model for the remaining risk, with thresholds owned by revenue cycle leadership.
- Set up governance. A monthly review of new payer policies, rules that never fire, and rules that hold too many clean claims.
- Measure. Track the KPIs below before and after each change.
Which KPIs Show Whether RCM Automation Works?
| KPI | What it measures | Why it matters for rules |
|---|---|---|
| Clean claim rate | Share of claims that pass edits without manual correction | Shows how much rework rules remove |
| First-pass yield / first-pass resolution | Share of claims paid on first submission | The most direct measure of denial prevention |
| Initial denial rate | Share of claims denied on first adjudication, by reason and payer | Shows which rules are missing or stale |
| Denial overturn rate | Share of appealed denials that are paid | Shows whether appeals are well targeted |
| Days in accounts receivable | Average time to collect | Falls when fewer claims need rework |
| Cost to collect | Revenue cycle cost relative to cash collected | Falls when less staff time goes to rework |
| Authorization-related denials | Denials for missing or invalid authorization | Direct measure of authorization rules |
| Rule hit rate | How often each rule fires, and how often a hold was justified | Prevents rules that slow down clean claims |
Define each KPI once, using your organization’s standard definitions, and measure it the same way before and after a change.
Where Does Higson Fit?
Higson is a business rules engine built by Decerto. It was designed for property and casualty insurers and financial services companies, where claims, underwriting and pricing rules face similar demands: frequent changes, many sources of rules, and audit questions about every decision. Higson is not an RCM platform, a clearinghouse or a medical device, and we do not claim healthcare customers.
What it offers is the rules layer described in this guide:
- Decision tables that business teams maintain. Payer matrices, authorization requirements, filing limits, financial policies and decision thresholds live in decision tables that analysts edit and test in Higson Studio. Tables prepared in Excel are imported through CSV.
- Bring your own model. Higson’s ONNX runtime (since version 4.2) runs your own machine learning models, such as a denial prediction model, inside rule execution. Your data science team keeps ownership of its models. The score and the rules around it are evaluated in one decision and logged together.
- Testing before production. The tester and mass tester run a rule change against historical cases, so teams see how many decisions change before publishing.
- Versioning, audit trail and granular permissions. Every change is versioned with rollback. Higson 4.3 LTS adds role-based permissions at the level of business domains, sessions and operations, with a full audit trail.
- Visibility. The Diagnostic Panel shows flamegraphs of rule calls, heatmaps of decision table rows and P50/P95/P99 percentiles, so teams can find rules that fire too often or never.
- Real-time performance. Rule execution takes 0.23 ms per decision, with throughput of 9,000 requests/second.
- Deployment and security. Cloud (AWS, Azure, GCP), on-premise Kubernetes or hybrid. The Runtime API supports JWT, SSL and basic authentication; Higson Studio supports SSO through Active Directory, SAML and CAS. Any deployment that processes protected health information must meet HIPAA requirements.
- AI-ready. An official MCP server with 50+ tools lets AI agents work with rules, and the AI Assistant in Higson Studio (beta) drafts functions in Groovy or Python and decision tables from natural-language descriptions.
The closest fit is on the claims side, where Higson already works in insurance. BNP Paribas Cardif consolidated claims rules from 5 product lines into one Higson engine in 3 months, with one audit trail. At TUW, actuaries publish a new motor rate in 2 hours instead of 2 weeks. Health plans and RCM vendors face the same problem: many rule sources, frequent changes and a need to prove which rule made each decision.
See How Payer Rules Would Look in a Rules Engine
Bring one denial cause, such as missing authorizations for one payer, and we will show you how the rule would look as a decision table in Higson, tested on your own sample claims.
Download the Business Rules Engine Comparison Guide 2026
Related Articles
Rules Engines in Healthcare: Supporting Care Decisions Without Replacing the Clinician


