42-site ABA chain · Multi-state · 18 months engagement
Three-way match reversed $1.4M of post-pay takebacks.
Authorization, supervision ratio, and EHR session note brought to one gate before submission. Eighteen months of 97153 takebacks closed at the root. Zero unbilled auth-expired sessions for twelve months running.
Specialty
ABA Therapy
Footprint
42 sites
Geography
4 states
Engagement
18 months
Primary CPT
97153
Annual book
$28M NCR
The situationThree signals. Three sources. No gate.
TAKEAWAY
Three sources of truth, no shared gate. Takebacks landed 90 to 120 days late.
SYSTEMS OF RECORD
3
TAKEBACK LAG
90-120d
SITES IMPACTED
42
MONTHLY BLEED
$78K
Leak 01 · Authorization
Auth expired silently.
- → Balance hit zero mid-week
- → Notes kept closing
- → Takeback letter, 6 weeks later
Leak 02 · Supervision
Ratio drifted under threshold.
- → RBT hours grew faster than BCBA
- → Ratio crept under 10% in busy weeks
- → Payer audits caught it later
Leak 03 · Notes
Times did not match.
- → Billed vs EHR clock differed 7-25 min
- → Largest audit recovery source
- → Aggregated across 42 sites
Three sources · one truth
The intersection is what ships.
The fix · workflowOne gate. Three sources. Pre-submit.
TAKEAWAY
If all three agree, the claim ships. If any disagree, it holds.
What the dashboard showsActive authorizations. Live unit balance.
GREEN
Healthy
Auth balance under 60% used
AMBER
Reauth trigger
60-89% used, start reauth now
RED
Close window
90%+ used, escalation only
Active auths · 42 sites · all states
refreshed 14s ago
Beneficiary A · PhoenixRBT: M. Chen · BCBA: T. Park
97153
62%
Beneficiary B · TucsonRBT: J. Smith · BCBA: T. Park
97153
48%
Beneficiary C · MesaRBT: A. Davis · BCBA: K. Liu
97153
84%
Beneficiary D · MesaRBT: A. Davis · BCBA: K. Liu
97155
79%
Beneficiary E · TempeRBT: R. Patel · BCBA: K. Liu
97153
35%
Beneficiary F · PhoenixRBT: M. Chen · BCBA: T. Park
97153
97%
Beneficiary G · PhoenixRBT: S. Garcia · BCBA: T. Park
97153
88%
Beneficiary H · TucsonRBT: J. Smith · BCBA: T. Park
97155
42%
+842 more across 42 sites · 4 red · 27 amber
Engagement ScorecardQuarterly partner report · Q4 redacted
CONFIDENTIAL
KPI movement before vs current.
Denial rate · 97153
14.7%
3.9%
Auth-expired session takebacks
$78K/mo
$0/mo
Supervision ratio compliance
73%
98%
Net collection rate (first-pass)
86%
98%
Cash days outstanding
52 d
28 d
Time-on-claim audit recovery
11%
0.4%
Q4 · 2026
ASP-RCM Senior Partner
Recovery by CPTWhere the $1.4M came from.
TOP CPT
97153
97153 SHARE
61%
TOP 5 LINES
96%
WINDOW
18 mo
Implementation timelineFrom kickoff to steady state.
ENGAGEMENT PROFILE
DURATION
18 months
PHASES
4 gates
SITES
42
RECOVERED
$1.4M
01
Week 1 · Discovery
Map the leak.
18-month takeback log pulled. Root cause tagged per CPT. EHR, CredPro, CentralReach connection scoped.
02
Week 4 · Gate live
First catches.
Three-way match enabled at four pilot sites. Two takebacks blocked in the first week. Parallel run vs prior workflow.
03
Month 3 · Rollout
All 42 sites.
Payer-specific supervision rules added. Denial rate at 6.4 percent. Recovered claims worked back into appeals queue.
04
Month 6 · Steady
3.9% denial floor.
Zero auth-expired sessions billed. NCR at 98 percent first-pass. Cash days from 52 to 28.
We had been writing off roughly $80K a month for 18 months. The gate caught it in week four. By month six we had stopped budgeting for takebacks at all.
CFO · 42-site ABA chain (anonymized)
Operating postureHow the work actually changed.
TAKEAWAY
Coordinators stopped chasing takebacks. The gate does the work.
BEFORE · MONTH 0
7.8% post-pay takeback rate
- Auths checked manually at coding time
- Supervision ratio audited quarterly, not at submission
- EHR session notes reconciled at month-end
- Takeback letter was the first signal of a gap
- $78K per month rolled into bad-debt forecast
AFTER · MONTH 6
0.4% audit-recovery rate
- All three sources gated before submission
- Real-time supervision ratio block at session save
- Session note alignment validated at submission
- Hold queue is the only place a gap surfaces
- Takeback line item removed from forecast
OutcomesBefore. After. In numbers.
Pre-engagement · baseline
Denial rate · 97153
14.7%
Takebacks per month
$78K
Cash days outstanding
52d
NCR first-pass
86%
Annualized leak ≈ $1.0M
Steady-state · month 10+
Denial rate · 97153
3.9%
Takebacks per month
$0
Cash days outstanding
28d
NCR first-pass
98%
18-month cumulative recovery $1.4M
Capability stackWhat the gate actually runs on.
STACK
CredPro v6 service. AI Suite parser. LLM Gateway check. One audit log.
Layer 04 · AI
Coding AI · Denial Prediction · Note parser
▼
Layer 03 · LLM Gateway
Single audited choke point · cost meter · prompt registry
▼
Layer 02 · Platform
CredPro v6 · Three-way match service · Reporting cloud
▼
Layer 01 · HIPAA-eligible AWS
AES-256-GCM PHI · row-level RBAC · PHI access log
Common questionsFrequently asked: three-way match.
What is the three-way match?
A gate that confirms three things agree before a 97153 claim ships: active authorization, payer-required supervision ratio, and the EHR session note for that date, time, location, and provider. Any disagreement holds the claim.
Why is this the highest-leverage control for ABA?
The most expensive ABA takebacks happen when one of those three sources contradicts the others. Standard scrubbers do not catch silent auth expiry, ratio drift, or timestamp mismatches. The three-way match catches them deterministically before submission. The gate ships as standard scope in ASP-RCM's ABA therapy billing services.
How does this connect to CredPro v6?
CredPro v6 carries BCBA, BCaBA, RBT certification status, the BACB credential lookup, and supervising-BCBA assignment per RBT. The match pulls those facts at submission time, not from a stale spreadsheet.
Why anonymize the client?
The MSA includes reciprocal confidentiality. Numbers and timeline are real. A senior partner can host a reference call under NDA at late-stage diligence.
How long until we see results?
Visible denial-rate movement starts in week three. Steady-state shows up by week ten. The $1.4M figure is cumulative across 18 months.
What does the free audit look like for ABA?
Send 90 days of denial data with CARC and RARC, your authorization log, supervision ratio report by RBT, and provider roster. You get a 4-page written audit inside 30 days: denial taxonomy by root cause, recoverable revenue, authorization gap analysis, supervision compliance, 90-day fix plan.
Does this only work for multi-site chains?
No. The match logic is identical at any scale. Single-site practices, multi-state chains, and IDDs all run the same gate. Economics shift with volume; the discipline is the same.
Want the same audit applied to your data?
A free 30-day audit. Send 90 days of denial data and your provider roster. We return a 4-page written audit: denial taxonomy by root cause, recoverable revenue, authorization gap analysis, 90-day fix plan. Yours to keep. No SDR follow-up.