06-reference/research

louisiana act 233 aba denial series

2026-11-17·research-brief·source: deep-research·by Ray Data Co (deep-research synthesis)
behavioral-healthdenial-ratesmedicaid-managed-carefamily-ventureaba

Louisiana's ABA denial rate holds below the non-behavioral-health rate in all seven quarters that exist, but the series starts in Q4 2023, not 2019, and the 7.2x payer spread is a small-denominator artifact sitting on top of a real ~2x one

The question

Verbatim: "Does the Louisiana ABA denial rate hold across the full Act 233 quarterly series back to calendar 2019, and does the 7.2x per-MCO spread persist or was Q2 2025 an outlier?"

Context: [[2026-10-17-state-medicaid-behavioral-health-denial-rates-by-category]] read one quarter (Q2 2025) of the Healthy Louisiana Claims Report and found Applied Behavior Analysis (ABA) denied at 9.31% against 16.79% for non-behavioral-health categories, with a 7.2x spread across six managed care organizations (MCOs). This brief reads the whole series. Acronyms: ABA = Applied Behavior Analysis; BH = behavioral health; MCO = managed care organization; LDH = Louisiana Department of Health. MCO codes as printed: ABH = Aetna Better Health, ACLA = AmeriHealth Caritas Louisiana, HBL = Healthy Blue, HUM/HHH = Humana Healthy Horizons, LHCC = Louisiana Healthcare Connections, UHC = UnitedHealthcare.

What we already know (from the vault)

What the web says

All figures below are from 26 LDH report PDFs that I downloaded from ldh.la.gov and parsed (pdftotext -layout, then scripted extraction; the Q2 2025 parse reconciles to the parent brief to the claim). Quarters read: every quarter from 2018 Q2 through 2025 Q2 (29 quarters), no sampling. 2018 Q2 to 2023 Q3 come from the rolling four-quarter tables in 19 Act 710-format reports; 2023 Q4 to 2025 Q2 come from Table 1.8 of the seven Act 233-format reports. Statewide aggregates are my derivation from published counts; the reports print per-MCO percentages only.

Quarter ABA denied / adjudicated ABA BH - Residential All Other Specialized BH All BH Non-BH ABA minus non-BH
2023 Q4 7,089 / 89,085 7.96% 7.56% 16.14% 14.76% 14.84% -6.9 pts
2024 Q1 7,864 / 88,589 8.88% 9.49% 16.59% 15.37% 17.21% -8.3 pts
2024 Q2 6,627 / 102,855 6.44% 9.75% 17.08% 15.59% 17.20% -10.8 pts
2024 Q3 6,892 / 98,583 6.99% 10.73% 18.07% 16.59% 16.40% -9.4 pts
2024 Q4 9,136 / 102,097 8.95% 10.92% 15.62% 14.72% 16.06% -7.1 pts
2025 Q1 9,174 / 96,479 9.51% 10.57% 14.72% 13.87% 16.47% -7.0 pts
2025 Q2 10,241 / 109,991 9.31% 9.12% 15.38% 14.18% 16.79% -7.5 pts

ABA ranges 6.4% to 9.5%; non-BH ranges 14.8% to 17.2%. Net of claims denied as exact duplicates (a row the report also publishes), ABA is 4.9% to 7.9% and the ordering is unchanged. Q2 2025 was not an outlier on the level; it is the second-highest ABA quarter of seven. The drift is mildly upward: the last three quarters (8.95 / 9.51 / 9.31) sit above the first four (7.96 / 8.88 / 6.44 / 6.99). Confidence: high on arithmetic, medium on interpretation for the reasons under methodology flags.

Quarter Mental/Behavioral Rehab Mental/Behavioral Other BH combined Non-BH basket
2018 Q2 10.38% 14.01% 11.75% 12.82%
2018 Q3 12.45% 13.75% 12.99% 13.62%
2018 Q4 14.73% 16.33% 15.52% 14.19%
2019 Q1 16.02% 14.91% 15.37% 15.04%
2019 Q2 21.87% 16.70% 18.62% 14.36%
2019 Q3 14.76% 21.38% 19.34% 14.28%
2019 Q4 14.62% 17.37% 16.52% 12.22%
2020 Q1 13.11% 16.17% 15.27% 12.67%
2020 Q2 11.48% 15.77% 14.68% 13.83%
2020 Q3 9.74% 16.06% 14.62% 13.00%
2020 Q4 10.60% 17.42% 15.95% 12.77%
2021 Q1 13.36% 22.20% 20.31% 14.23%
2021 Q2 11.46% 14.58% 13.98% 13.95%
2021 Q3 20.29% 19.01% 19.27% 13.99%
2021 Q4 9.76% 16.99% 15.67% 13.40%
2022 Q1 9.89% 14.54% 13.71% 13.52%
2022 Q2 12.59% 15.84% 15.24% 12.82%
2022 Q3 11.38% 15.85% 15.10% 13.42%
2022 Q4 12.03% 13.70% 13.43% 12.94%
2023 Q1 20.96% 17.04% 17.67% 16.84%
2023 Q2 19.79% 15.13% 15.86% 15.75%
2023 Q3 20.54% 15.86% 16.60% 15.38%

The first Act 710 report says so in words: "The claims denial rates for behavioral health services are slightly higher than those found for most acute care services." BH combined ran 11.8% to 20.3%, the non-BH basket 12.2% to 16.8%. The BH lines are also spikier: Rehab jumped to roughly 20-22% in 2019 Q2, 2021 Q3 and all of 2023 Q1-Q3, each time without a matching move in non-BH. The two eras are not one continuous series (different categories, different formula, different non-BH basket), so the apparent flip from "BH slightly higher" to "BH slightly lower" at 2023 Q4 cannot be read as a real change. Confidence: high on the figures, low on any cross-era comparison.

Quarter ABH ACLA HBL HUM LHCC UHC Min-max, all six Spread Spread, three MCOs with 10k+ lines
2023 Q4 0.6% (165) 5.6% (12,932) 7.4% (13,781) 23.1% (143) 8.6% (61,254) 4.7% (810) ABH - HUM 38x 1.5x
2024 Q1 2.7% (112) 6.0% (11,435) 11.4% (14,509) 33.8% (130) 8.8% (60,901) 6.8% (1,502) ABH - HUM 13x 1.9x
2024 Q2 21.3% (422) 5.2% (15,530) 6.7% (15,907) 8.9% (79) 6.5% (69,779) 8.5% (1,138) ACLA - ABH 4.1x 1.3x
2024 Q3 10.3% (416) 3.8% (13,982) 7.2% (15,552) 48.3% (29) 7.6% (67,679) 3.6% (925) UHC - HUM 13x 2.0x
2024 Q4 0.3% (385) 4.7% (14,318) 6.7% (16,818) 47.2% (53) 10.2% (69,174) 16.5% (1,349) ABH - HUM 182x 2.2x
2025 Q1 7.2% (278) 2.8% (13,834) 9.0% (16,126) 39.6% (96) 10.9% (64,593) 16.5% (1,552) ACLA - HUM 14x 3.9x
2025 Q2 11.0% (454) 5.0% (15,843) 11.4% (20,183) 34.7% (838) 8.6% (70,564) 35.9% (2,109) ACLA - UHC 7.2x 2.3x

Gross rate, adjudicated ABA lines in parentheses. Humana's rate rests on 29 to 838 lines, Aetna's on 112 to 454, UnitedHealthcare's on 810 to 2,109. ACLA, HBL and LHCC adjudicate 97% to 99% of all ABA lines every quarter, and LHCC alone 64% to 69%. Among those three the spread is 1.3x to 3.9x, median about 2x. Confidence: high.

Convergences and contradictions

Synthesis for RDCO

The finding that was worth protecting holds. ABA is denied at 6% to 10% on Louisiana Medicaid managed care in every quarter anyone has published, against 15% to 17% for the non-behavioral-health categories in the same table, across roughly 688,000 adjudicated ABA lines. That is enough to quote, with three qualifiers that should travel with it every time: seven quarters and one state, MCO-self-reported, and last-adjudication-in-period rather than final. The vendor "15-30% ABA initial denial" claim is still contradicted. What cannot be said is that this is a six-year result. ABA was not a reported line before Q4 2023, and the entire seven-quarter series was first published between June 2025 and February 2026.

The payer-variance thesis from the parent brief needs to be rebuilt on a smaller number. "7.2x" should be retired; it is one quarter's ratio between a payer with 15,843 ABA lines and one with 2,109, and the same calculation gives 182x one quarter and 4.1x another. What is real is narrower and probably more useful. First, among the three MCOs that process 97-99% of Louisiana ABA, the best payer runs about half the denial rate of the other two in every one of seven quarters. A persistent 2x under one state's rules is a payer-rules effect a provider cannot see from inside one practice, which is the original product argument at honest scale. Second, the volume distribution is itself a finding: one MCO carries two-thirds of the state's ABA lines, and three large MCOs (including UnitedHealthcare, which adjudicates over a million outpatient-hospital lines a quarter) report almost none. Either ABA members and providers are concentrated to a degree enrollment shares would not predict, or the MCOs classify ABA claims into different rows. The report cannot distinguish these, and the second reading would mean the per-MCO ABA rates are not measuring the same thing. That is my inference, not a published statement.

For the family venture conversation this changes the wording, not the direction. The question to a multi-site ABA operator stays "what is your denial rate by payer, and how far is your worst from your best," but the public benchmark to hold against it is "about 2x between large payers, with small-book payers running 25-45% on thin volume," not 7x. The small-book observation may be the sharper wedge: Humana's ABA denial rate is 23% to 48% in six of seven quarters, and UnitedHealthcare's climbed from 4% to 36% over three quarters. For a provider, the low-volume payer is exactly where nobody on staff knows the rules, and the data is at least consistent with that. It is also consistent with noise at n=29, so it is a hypothesis to test on remittance data, not a claim to publish.

One caution on reuse. The older Act 710 series is the only long Medicaid behavioral-health denial series located, and it says BH was denied slightly more than non-BH for five years, with rehab spiking to 20%+ three separate times. Anyone citing Louisiana for "behavioral health is not denied more" is citing ABA and residential in the newer format only. The clean statement is service-specific: ABA is denied less; specialized outpatient behavioral health is denied about the same or slightly more.

Why this is in the vault

It is the n=7 check the family RCM/EOB venture needed before quoting the Louisiana ABA rate to Michael Holzum or any operator, and it replaces the 7.2x payer-spread figure in [[2026-10-17-state-medicaid-behavioral-health-denial-rates-by-category]] with a defensible ~2x among volume-carrying payers, which resizes the payer-benchmarking pitch.

Open follow-ups

Related

Sources

Vault

Primary, downloaded and parsed by me (all live on ldh.la.gov on 2026-11-17; file list recovered from the Wayback CDX index)

Act 233 format, Table 1.8 "Outpatient and Professional Claims Adjudicated" (ABA era):

Act 710 format, Appendix A "Claim Denial Rates ... For All MCEs Combined, By Quarter" (19 reports, CY2019 Q1 through CY2023 Q3, each carrying a rolling four quarters, so 2018 Q2 onward), all under https://ldh.la.gov/assets/docs/LegisReports/ACT710/ unless noted:

Research budget used: QMD 0 of 5 (the parent brief supplied the vault priors). WebSearch 1 of 3. WebFetch 0 of 3; the 26 PDFs were pulled by curl and one Wayback CDX query located the filenames.

Not reached / limits: No report later than 2025 Q2 was located (filename-pattern probes for 2025 Q3 through 2026 Q3 returned 404; LDH has no stable naming, so a later report may exist under a name I did not guess). Per-MCO behavioral-health denial rates in the Act 710 era are published as charts without numeric tables in the text layer, so there is no per-MCO BH spread before 2023 Q4. The initial Act 710 report (covering CY2017) was downloaded but not parsed. The Act 710 non-BH comparator is my own basket and is not the same category set as the Act 233 non-BH rows.