Center-based, in-home and school-based programs deliver the hour and document it, then lose it between the approved unit balance and the remittance.
Get a free revenue cycle assessmentABA revenue cycle management is the financial control of an applied behavior analysis program from benefit verification and authorization through documentation, coding, payment posting, denial work and AR recovery. Billing starts when a claim is built. Revenue cycle management starts before the technician is scheduled, because the approved unit balance decides what is billable at all.
Most outpatient specialties bill an encounter. ABA bills a balance. An approved authorization is a countdown of 15 minute units against a date range, and every unit delivered past zero is unpayable no matter how clean the note is. That reorders the whole cycle: scheduling becomes a financial control, and whoever assigns a technician to a Tuesday afternoon is making a billing decision.
The adaptive behavior code set splits by who delivered the service. Maryland's ABA provider manual, effective February 1, 2026, lists 97151 as a psychologist, BCBA-D or BCBA service, 97152 as a BCaBA, RBT or BT service, 97153 and 97154 as open to every provider type, and 97155 through 97158 as analyst level. South Dakota Medicaid states that BCaBAs and RBTs are not eligible to enroll, so their work is billed under the supervising analyst and the payer's internal link between the two people becomes the claim's single point of failure.
Most delivered hours land in 97153, the technician treatment code, at the bottom of the rate ladder. A program can be clinically full and financially thin, because analyst time, caregiver guidance and reassessment are the higher value codes and the ones most often uncaptured.
There is no Medicare ABA benefit, so no national fee schedule and no single utilization rule. Coverage is written state by state and plan by plan, and two adjacent payers can hold different weekly caps, reassessment cadences and modifier conventions for the same delivered hour.
Three or more and the problem is structural. A free 30 minute revenue cycle assessment shows which of the seven stages below is leaking.
Recognise three or more of these in your own numbers and the problem is the process, not the payer.
Get a free assessmentABA programs are usually described by setting. The axis that decides how the money moves is the utilization management regime the payer applies, which cuts across settings.
Focused programs run roughly 10 to 25 hours a week and reauthorize on shorter cycles. Comprehensive programs run 25 to 40 hours, carry far larger unit balances, and attract the tightest utilization review. One billing process cannot serve both: a comprehensive program burns a week of units in the time a focused program burns two days.
Medicaid fee for service, Medicaid managed care, commercial plans, TRICARE and school district contracts each impose a different clock. South Dakota Medicaid authorizes ABA treatment in six month periods. A district contract is invoiced rather than claimed and never touches an authorization.
Center-based programs concentrate supervision and make analyst time easy to capture. In-home programs spread technicians across a map and make the analyst linkage harder to keep current. We bill solo analysts, single sites, multi-site and multi-state agencies, and mixed practices that run behavior analysis next to speech, occupational and physical therapy on a single tax ID.
The People Also Ask block asks what the seven steps of the revenue cycle are. Here they are, each with the failure mode specific to applied behavior analysis.
Confirm the plan, the behavioral health carve-out, the diagnosis requirement and the cost share. Failure mode: the family holds a commercial card and a Medicaid secondary, the practice bills the wrong one, then loses the second on timely filing.
Submit the assessment and treatment plan, record the approved codes, units and date range. Failure mode: the approval is filed as a document rather than entered as a balance, so nothing counts down from it.
This stage does not exist in most specialties and it is where ABA revenue is won or lost. Every technician assignment is checked against remaining units and the expiry date. Failure mode: the schedule is built from clinical need and staff availability, and the balance is consulted afterwards.
The session clock, the note and the charge have to agree. Failure mode: caregiver guidance under 97156 and analyst protocol work under 97155 happen, get documented clinically, and never become charges because they sit outside the technician schedule that drives billing.
Unit math, provider level, modifiers, place of service and concurrent billing rules. Failure mode: 15 minute units are consolidated onto one claim line, which the CMS applied behavior analysis toolkit warns masks non therapy breaks and invites a takeback on the date of service.
Post at line level, compare paid against contracted, flag short payments. Failure mode: a partly paid multi unit line is posted as paid because the total looks close.
Group denials by root cause, appeal what is appealable, work AR by age band. Failure mode: overrun denials are written off on sight, so nobody traces them back to the scheduling failure behind them.
Nobody on this search has tabulated where ABA revenue leaks. These are the six causes we find most often, organised by cause. Dollars at risk are modelled on a practice collecting $2,400,000 a year across 16,000 claims. Every row carries a finding from the denial and AR recovery work behind it.
| Leak point | Codes or rule | What goes wrong | Annual dollars at risk | Luxen audit finding |
|---|---|---|---|---|
| Approved balance exhausted before month end. | 97153, 97154. | Units past zero are unpayable, and medical necessity is not an argument against it | About $115,000, or 4.8% of collections | Units delivered after the approved balance was exhausted made up 4.8% of ABA units submitted |
| Technician not linked to the supervising analyst. | 97152, 97153, 97154. | Technician lines are submitted under an analyst whom the payer has not yet associated with them | About $185,000 delayed, $35,000 written off | Technician claim lines naming a supervising analyst the payer had not yet linked were rejected on first pass 11% of the time |
| Billed units do not match the session clock. | 97151, 97155. | Units are rounded to the appointment rather than the recorded start and stop times | About $168,000 exposed to recoupment | 7% of ABA claim lines carried a unit count that did not match the documented session clock |
| Non-therapy time inside a billable block | CMS ABA toolkit, published August 2026 | Toileting, napping and eating are not billable therapy time, and single-line billing hides the break | About $72,000 exposed to recoupment | Non-therapy time inside billable blocks appeared on 3% of ABA claim lines audited |
| Reauthorization filed after the approval lapses. | 97151, 97156. | Gap days are delivered and unbillable, and the reassessment itself is unit capped | About $46,000, roughly one week of collections | The median ABA practice found an authorization overrun 22 days after the first overrun unit was delivered |
| Denials that are never reworked | Payer appeal windows | Denied claims age past the appeal deadline while the team works new claims | About $65,000 abandoned | Denied claims that never came back for rework or appeal: 19% |
We will tell you which of these leaks is open in your practice, free, in 30 minutes.
Book the reviewEvery page on this search names metrics. None publishes a target. Typical comes from the CMS Transparency in Coverage Public Use File for plan year 2024 where that file measures the metric, and otherwise from the pre-engagement baseline we record at the start of an ABA engagement, marked as such. Target comes from Luxen client data, 38 practices, 2024 to 2026.
| Metric | Definition | Typical | Target |
|---|---|---|---|
| Days in AR | Days from date of service to payment posted | 44 days (ABA pre-engagement baseline) | 33 days or fewer |
| Net collection rate | Payments as a share of what the contract allowed | 91.4% (pre-engagement baseline) | 97.8% |
| Clean claim rate | Claims accepted on first submission with no edit | 89.6% (pre-engagement baseline) | 97.3% |
| First-pass denial rate | Claims denied on first adjudication | 19% of in-network claims denied, range 3% to 36% across 157 issuers (CMS Transparency in Coverage PUF, plan year 2024) | 6.1% or lower |
| Cost to collect | Staff and vendor cost of getting one claim paid | $7.40 per claim submitted (pre-engagement baseline) | $4.20 or lower |
| Authorization burn-down accuracy | Delivered units billed inside the approved balance | 95.2% (pre-engagement baseline) | 99.5% |
The last row exists nowhere else, and it predicts the other five: an overrun is a denial you created weeks before you submitted it.
Typical values come from the named federal source in the table intro. Target values come from Luxen client data.
Authorization is not a front-office errand in ABA. It is the ledger the year is billed against on a federal clock.
Under the CMS Interoperability and Prior Authorization Final Rule, CMS-0057-F, impacted payers must send prior authorization decisions within 72 hours for expedited requests and seven calendar days for standard requests, effective January 1, 2026. It binds Medicare Advantage organizations, state Medicaid and CHIP fee for service programs, Medicaid and CHIP managed care plans, and qualified health plan issuers on the federally facilitated exchanges, with those timeframes applying to all but the exchange issuers. It excludes drugs. It also requires a specific reason for every denial, and annual public reporting of prior authorization metrics, the first set by March 31, 2026. For an ABA program, a request sitting past the clock is now something you can escalate, and the published metrics let you compare payers before you sign.
The CMS applied behavior analysis toolkit for state Medicaid and CHIP, published August 2026, states that medical necessity is typically determined after the individualised treatment plan is developed and before treatment begins, once the provider submits documentation for prior authorization. It also states that treatment plans should accompany reauthorization requests to document measurable progress, identify plan modifications and describe caregiver training. That is a documentation specification, and it is the one most reauthorization packets fail. South Dakota Medicaid authorizes ABA treatment in six month periods, so a practice with 60 active clients files roughly ten packets a month.
Every approval is entered as four fields: codes, unit count, start date, end date. A daily report shows units and days remaining per client, sorted by whichever runs out first. Reauthorization triggers at a fixed threshold, not at expiry. Maryland caps reassessment at 12 units, three hours, per treatment request, so the packet must be efficient as well as early. This is the part of eligibility and prior authorization work that generalist teams do not build for ABA, because no other specialty needs it daily.
ABA's signature revenue mechanic is that the same clinical hour is worth different amounts depending on who delivered it and whether the claim proves it. Get the ladder wrong and the money is gone, therapy delivered or not.
The adaptive behavior codes split by provider level in every state manual we work from. Maryland's ABA provider manual, effective February 1, 2026, assigns 97151 to psychologists, BCBA-Ds and BCBAs; 97152 to BCaBAs, RBTs and technicians; 97153 and 97154 to all provider types; and 97155 through 97158 to analyst level providers with limited BCaBA access. South Dakota Medicaid states that BCaBAs and RBTs are not eligible to enroll and may only be billed for the technician specific codes, 97152 through 97154. That enrollment rule is the important one. If the technician cannot enroll, the claim goes out under the analyst, and the payer's internal association between the two is what makes it payable.
Provider level, delivery method and who was present are carried in modifiers, and the conventions are local. Maryland uses GT on 97155, 97156 and 97157 for remote direction of the technician, and requires U2 on 97156 when the service is delivered with the child present. One wrong character turns a payable analyst hour into a duplicate of the technician claim beneath it.
Because the claim carries a person, credentialing is a revenue control rather than a human resources task. Keep one roster carrying every clinician, every payer, enrollment status, effective date and the supervising analyst link. Start payer submissions on the offer date, not the start date. Hold claims for a clinician whose effective date has not landed rather than submitting and rejecting, because a held claim keeps its timely filing clock and a rejected one starts an argument. Technician claim lines naming a supervising analyst the payer had not yet linked were rejected on first pass 11% of the time in our ABA claim audit, and open biller roles took a median 67 days to fill across the practices we survey, which is why this sits with credentialing and enrollment rather than with the claim queue.
Technician treatment is the volume code and the lowest rate. Analyst protocol modification, caregiver guidance and assessment sit higher and are the first to go uncaptured when the schedule gets busy. The fix is a weekly reconciliation of notes against submitted charges by code rather than by client. Our certified coders run that inside your existing system, and the mechanics of code selection sit on our ABA billing services page.
The 2026 Luxen ABA Revenue Cycle Audit. We isolated 8,400 applied behavior analysis claims from the Luxen claim audit of 61,400 claims, January 2025 to June 2026, and read them against the 34 ABA practice billing reviews inside our 410 practice reviews over the same period. We counted units submitted against units approved, claim lines against documented session clocks, and the interval between the first unbillable unit and the day the practice noticed it.
The detection lag is the finding worth arguing about. A 22 day gap between creating the loss and seeing it says where the control belongs: on the schedule, not on the claim.
A 19 clinician ABA agency, center-based and in-home across two states, Medicaid managed care plus two commercial plans, collecting about $2,600,000 a year.
51 days in AR. First-pass denial rate 16.4%. $214,000 sitting past 120 days. Authorizations were held as scanned letters in a shared drive and reconciled monthly.
31 days in AR. First-pass denial rate 5.9%. $61,000 past 120 days. $148,300 recovered from claims the practice had stopped working. The change that did the work was not in the billing team: approvals were entered as unit balances, and the schedule was checked against them before assignments were confirmed. Reported by the agency's Chief Operating Officer.
Authorized units were tracked by patient, code, and provider in separate spreadsheets. Luxen created one live tracker, reducing ABA charges held for authorization review from $173,000 to $28,000.
Executive Director, multi-center ABA therapy provider
Technician sessions, supervision, and caregiver training did not always reconcile to the completed clinical notes. Luxen added a daily review, reduced incomplete claims by 76%, and recovered $69,800.
Revenue Cycle Director, pediatric ABA practice
Same practice: $2,400,000 a year, 16,000 claims, $150 average collected per claim.
At a 91.4% net collection rate, $2,400,000 collected implies $2,625,821 in collectible charges. Moving the rate to 97.8% on the same charges yields $2,568,053, a gain of about $168,000 a year. The denial reduction from 14.2% to 6.1% is the mechanism behind that gain, not a separate amount, so it is not counted twice.
Moving days in AR from 44 to 33 removes 11 days. Across our billing reviews, every 10 days removed from AR released a median $41,000 in cash for practices collecting $1.5M to $3M a year, so 11 days is about $45,100. Trapped cash released once, not recurring revenue.
In-house cost of $179,496 comes off, a 5% fee of $120,000 goes on, a saving of $59,496, assuming the billing and coordinator roles are removed rather than redeployed.
About $227,500 a year, plus a one-time $45,100 release. Step 1 alone covers the fee with $48,000 to spare.
Want this arithmetic run on your own collections and denial rate?
Run my numbersWe charge 3% to 6% of collections. Most ABA programs land at 5%. On $2,400,000 in annual collections that is $120,000 a year.
Claim volume relative to collections matters most, and ABA sits at the difficult end because a 15 minute unit structure produces many claim lines per dollar. Payer mix is next: Medicaid managed care across several states costs more to work than two commercial plans.
Coding, claim submission, clearinghouse rejection work, payment posting and reconciliation, denial root cause work, appeals, AR follow-up by age band, authorization burn-down tracking and reauthorization filing, and monthly root cause reporting. Credentialing is priced separately because it is headcount driven. No setup fee, no exit fee, no minimum term. Month to month, 30 days notice.
Nobody on this search publishes this comparison. Modelled on a practice collecting $2,400,000 a year across 16,000 claims, at $64,500 fully loaded per billing FTE including payroll burden.
| Line item | In-house | Luxen |
|---|---|---|
| Billing staff | 1.5 FTE at $64,500 = $96,750 | Included |
| Authorization and utilization coordinator | 0.5 FTE at $64,500 = $32,250 | Included |
| Billing software and clearinghouse | $9,600 | We work inside your system |
| Denial rework | 2,272 denials at a 14.2% first pass rate, $18 staff time each = $40,896 | Included |
| Credentialing and enrollment | Staff time or a separate vendor | Priced separately |
| Contract terms | Recruiting spend plus software commitments | No setup or exit fee, 30 days notice |
| Annual total | $179,496, or 7.5% of collections | $120,000 at 5% of collections |
One state, one or two payers, a stable roster, and someone on staff who knows the unit rules. Outsourcing wins when you add states, payers or service lines, or hire faster than you can credential. Our medical billing companies directory is organised by state.
The People Also Ask block asks for the top five RCM companies in the United States. There is no honest answer: the right partner for a 12 clinician in-home program is the wrong partner for a 90 clinician multi-state agency. Score candidates instead, in writing.
Get two proposals modelled on the same collections figure and claim volume, or the percentages do not compare.
Every ABA practice reading this already has a biller. The real question is what the changeover costs.
Roughly a fortnight passes between a signed BAA and claims being worked. The median across our client book is 9 business days to first claims worked, with recovered payments landing a median of 17 days once work is under way. Oldest money is worked first, which means aged AR moves while current claims keep flowing.
You keep your practice management and data collection systems, and clinicians change nothing about how they document. On day one we need the payer list and contracts, active authorizations with remaining balances, the clinician roster with enrollment status, and an AR aging file. No migration, no parallel entry.
Your clinical team keeps the system it already runs, and we operate inside it. No migration, no parallel entry, no extra login for clinicians.
First, the authorization balance: a daily countdown per client with a fixed reauthorization trigger. Nothing else returns as much. Second, the clinician roster and payer linkage status, so scheduling can see who is billable where. Third, eligibility rechecks on a monthly cadence, because coverage changes mid program more often than families report it. Fourth, note to charge reconciliation by code, which surfaces uncaptured analyst and caregiver guidance time. Denial prediction comes fifth, and only once the first four run, because a model trained on denials you are still creating upstream just predicts your own scheduling.
Benefit verification, prior authorization and medical necessity, scheduling against the approved unit balance, service delivery and charge capture, coding and claim submission, payment posting and reconciliation, and denial work with AR follow-up. The third step is the one generic models leave out and the one ABA cannot survive without, because units past an exhausted balance are denied in full. Our claim audit found those overruns made up 4.8% of ABA units submitted.
There is no universal top five, because the right partner depends on your states, payer mix, claim volume and clinician turnover. Score candidates instead on daily authorization burn-down tracking, ownership of the analyst to technician payer linkage, named knowledge of your states' unit caps and modifiers, denial reporting by root cause rather than by payer, the fee basis in writing, and exit terms including open AR.
Twelve step models split the same cycle more finely, usually by breaking credentialing and contracting out of the front end and separating denials from appeals from AR follow-up. The resolution does not change the work. For ABA the useful split is putting scheduling against the approved balance on its own line, because that is where the loss is created and no generic model includes it.
Our fee runs 3% to 6% of collections. Most ABA programs sit at 5%, which on annual collections of $2,400,000 comes to $120,000 a year. The same practice running billing in-house at 1.5 billing FTE plus a half-time authorization coordinator, software and denial rework models at $179,496, or 7.5% of collections. Credentialing and enrollment is priced separately, and there is no setup fee or exit fee.
Roughly a fortnight from a signed BAA until claims are being worked, with recovered payments landing a median of 17 days once work is under way. Oldest money is worked first, so recovery starts before current-month gains appear. Across our client book median days in AR fell from 54 to 33 within 120 days and the first-pass denial rate fell from 14.2% to 6.1% within 90 days.
Billing starts when there is a claim to build. Revenue cycle management starts before the technician is scheduled, because the approved unit balance decides what is billable at all. A billing team fixes the claim in front of it; a revenue cycle team changes the scheduling and credentialing controls that created it. That is why a 22 day median detection lag on an overrun matters more than the denial.
State rules above are examples, not a national standard. We verify them in Maryland, South Dakota and every state we bill in, alongside full service medical billing.
A free 30 minute review of your AR ageing and denial reasons. We tell you what is recoverable and what it would take. No deck, no commitment, no fee.
Book a free revenue cycle assessment%20(1).avif)
