Federal Enforcement Signals Expanding Payer-Behavior Scrutiny

UnitedHealthcare MA issues a SNF PA denial (CPT 99304–99306, MS-DRG 945–952 qualifying stays) citing that nH Predict projects fewer days than requested or that the patient does not meet predicted functional recovery thresholds. UR reviewer rubber-stamps the AI output without documented individualized clinical review, asserting the patient 'does not meet medical necessity criteria for skilled nursing level of care' — effectively substituting algorithmic LOS prediction for CMS-compliant clinical assessment. (inferred) Plans may argue the AI recommendation constitutes a clinical decision support tool used by a human reviewer, satisfying individualized review requirements.

(1) Immediately upon SNF PA denial from UHC MA, request in writing the specific AI model name, version, and clinical inputs used in the denial decision — cite the PSI report findings and CMS 2024 MA final rule AI-disclosure expectations. (2) Ensure the treating physician generates a standalone EHR attestation documenting skilled care necessity, prognosis, and why lower-acuity settings are clinically inappropriate — this note must predate or accompany the PA request, not be created reactively. (3) UR team should log denial rates by payer and service line monthly; UHC post-acute denial rates above 10–15% should trigger automatic physician advisor escalation. (4) File a CMS complaint citing the PSI report's 22.7% UHC post-acute denial rate (2022) as evidence of systemic improper denial patterns.

'I am requesting this peer-to-peer review because the denial does not reflect an individualized clinical assessment of this patient. The Senate PSI report documented that nH Predict-driven denials systematically deviated from traditional Medicare medical necessity standards. My patient's clinical presentation — [specific findings: wound care needs, IV medication requirements, complex nursing needs per 42 CFR 409.33] — independently satisfies skilled care criteria under CMS SNF coverage policy. I am asking your reviewing clinician to confirm that a qualified physician conducted an individualized review of this patient's record, not solely an algorithmic output, consistent with CMS MA regulations at 42 CFR 422.101(b).'

Humana MA denies LTACH authorization (MS-DRG 947–949, HCPCS S9125 in transitional contexts) following an internal Post-Acute Analytics model recommendation, asserting the patient can be managed at SNF or acute rehabilitation. The denial letter cites 'failure to meet LTACH medical necessity criteria' without specifying which clinical criteria were not met. (inferred) Humana's internal training materials — as documented in the PSI report — may have oriented UR staff toward denial-maximizing strategies for LTACH cases, making this denial pattern a watch item where Humana's LTACH denial rate grew 54% between 2020 and 2022.

(1) For all Humana MA LTACH referrals, require a standardized physician attestation template addressing LTACH qualifying criteria, documented failure of lower-acuity alternatives, projected LOS rationale, and specific clinical complexity drivers (ventilator dependence, daily IV antibiotics, complex wound care >2×/day). (2) Obtain the formal denial letter within 24 hours and confirm it names a physician reviewer and cites specific clinical criteria — if not, flag as a potential Post-Acute Analytics algorithmic denial and escalate to physician advisor. (3) Track Humana MA LTACH denial rates quarterly; rates above 15% should trigger UR committee review and CMS complaint filing, citing PSI LTACH denial growth data. (4) Prepare a standardized appeal packet referencing PSI findings on Humana's 54% LTACH denial-rate increase.

'This patient meets LTACH medical necessity criteria under CMS guidelines at 42 CFR 412.23(e) based on the following individualized clinical findings: [ventilator weaning protocol day X, complex wound stage IV requiring surgical nursing, IV vasopressor titration]. The PSI congressional investigation documented that Humana's LTACH denial rate grew 54% between 2020 and 2022 coinciding with deployment of Post-Acute Analytics. I am asking your reviewing clinician to confirm this denial reflects an individualized clinical assessment — not a financial-savings model projection — and to specify which clinical criteria this patient does not meet so we can address them directly in the appeal record.'

CVS/Aetna MA denies or curtails home health authorization (HCPCS G0179, G0180, G0181, G0182) asserting the patient no longer meets homebound status or skilled care criteria, based on Post-Acute Analytics model output projecting cost savings from denial. The denial may be issued as a concurrent-review curtailment — approving initial HHA but denying continuation after a predetermined algorithmic visit threshold — rather than an upfront PA denial, making it harder to identify as AI-driven. (inferred) CVS Post-Acute Analytics was projected to generate $77.3M in savings over three years, creating a documented financial incentive misalignment that is a watch item for all HHA concurrent review denials.

(1) Require all HHA transitions for CVS/Aetna MA patients to include a physician certification narrative (not checkbox only) completed before discharge that explicitly addresses homebound status barriers and skilled care necessity — documented in the EHR and transmitted with the HHA referral. (2) For any concurrent review curtailment from CVS/Aetna MA after April 2021, request in writing the specific algorithmic tool used in the determination — building an institutional database of AI-assisted payer denials supporting OIG complaints, CMS audit referrals, and litigation support. (3) For your own AI tools used in clinical documentation or authorization support, label all outputs as 'decision support' with mandatory clinician attestation, and ensure annual model bias testing for post-acute and high-acuity populations is conducted and documented — failure creates parallel liability to what the PSI report documented against UHC, CVS, and Humana.

'The PSI report documented that CVS/Aetna's Post-Acute Analytics tool was projected to save $77.3M over three years — a financial incentive structure that creates a duty of heightened individualized review on every HHA authorization curtailment. My patient meets homebound status under 42 CFR 424.22 and requires skilled care under 42 CFR 409.42 based on the following clinical findings: [specific deficits and skilled needs]. I am asking your reviewing clinician to confirm this curtailment reflects an individualized determination on this patient's clinical trajectory and not the application of an algorithmic visit threshold.'