PBM Reality 2026: Cross-Referencing 53 Drug Interactions, 55 Patient Assistance Programs, and 50 Insurer Formularies

Methodology

This audit cross-references three first-party OmniRx datasets compiled and maintained at omnirx.org:

• Interaction database: 53 drug-drug and drug-substance interaction records with severity grade, pharmacological mechanism, clinical recommendation, and primary source. Severity is graded on a four-step scale (Severe, Moderate, Mild, None) following DailyMed, FDA labeling, and peer-reviewed pharmacology literature accessed through PubMed.
• Patient assistance program (PAP) database: 55 programs, segmented by sponsor type (manufacturer, nonprofit, federal, discount), eligibility income floor, covered-drug list, and application route. Eligibility is normalized to percent of the HHS Federal Poverty Guidelines for an individual filer.
• Formulary database: 50 records, comprising 15 insurer-level formularies (UnitedHealthcare, BCBS, Aetna, Cigna, Humana, Kaiser, Anthem, Ambetter, Molina, Medicare Part D, TRICARE, Medicaid, VA, Marketplace, Medicare Advantage), 25 drug-level coverage explainers (Ozempic, Wegovy, Humira, Eliquis, Jardiance, and 20 others), and 10 cross-cutting guides.

For PBM industry context we draw on the Federal Trade Commission's July 2024 Interim Staff Report on Pharmacy Benefit Managers and the January 2025 second interim report. The FTC reports document that the three largest PBMs (CVS Caremark, Express Scripts, OptumRx) processed roughly 79% of U.S. prescription claims in 2023 and that vertical integration with insurers and specialty pharmacies has compressed independent pharmacy reimbursements while preserving rebate retention by the PBM. Our audit attaches consumer-facing data (interactions, PAPs, formulary explainers) to the regulatory frame the FTC has established.

Cross-tabulation. The cross-tab metric is the count of drugs that appear in BOTH the interaction database and the drug-coverage formulary records, using case-insensitive substring matching against drug names and slugs. This identifies medications where clinical risk and PBM pricing variance compound on the same patient.

Exclusions and limitations. Interaction count of 53 is below the 100+ threshold needed for tight category-level confidence intervals on severity statistics. Insurer formulary records describe tier structure rather than enumerating per-drug tier placement at full plan-by-plan scale; readers seeking specific tier placement for a specific drug on a specific plan should consult their plan's formulary document directly. PAP enrollment and approval rates are not publicly disclosed by manufacturers and are not in this dataset. PBM rebate retention data is summarized at the industry level, not the per-drug level. GAO's 2019 report on PBMs (GAO-19-498) and Drug Channels Institute coverage by Adam Fein supply the broader industry context.

Public dataset and pre-computed aggregates are available at data.json under a CC-BY 4.0 license. Source records live in OmniRx's interaction, assistance, and formulary sections.

Finding 1: 49.1% of cataloged drug interactions are Severe

Headline: 26 of 53 interactions Severe, 16 Moderate, 7 Mild, 4 None

Nearly half of the interactions in the OmniRx database (26 of 53, 49.1%) are graded Severe, meaning the combination is associated with clinically significant adverse outcomes documented in FDA labeling, FAERS reports, or peer-reviewed literature accessible through PubMed. Another 30.2% are graded Moderate, bringing the Severe-or-Moderate share to 79.2% (42 of 53). The skew toward higher severity is partly editorial selection: the OmniRx interaction catalog prioritizes the combinations a typical U.S. patient is most likely to encounter, weighted toward documented adverse outcomes rather than theoretical pharmacokinetic concerns. The FDA Adverse Event Reporting System consistently shows that polypharmacy-related adverse events disproportionately involve a small number of high-prevalence drug pairs; our catalog reflects that concentration.

Finding 2: Ibuprofen is the single most-flagged drug, appearing in 10 of 53 interactions

Headline: Ibuprofen 10, Alcohol 8, Lisinopril 5, Warfarin 5, Metformin 4

Ibuprofen leads the OmniRx interaction catalog with 10 appearances, more than any other single drug or substance. The next four spots are Alcohol (8), Lisinopril (5), Warfarin (5), and Metformin (4). The pattern is consistent with the broader pharmacology literature: NSAIDs interact pharmacokinetically and pharmacodynamically with anticoagulants, ACE inhibitors, antiplatelets, and several diuretic and renal-clearance pathways, so any catalog that aims at clinically relevant interactions concentrates NSAID rows. The FDA's 2015 strengthened NSAID warning and subsequent FAERS surveillance have repeatedly flagged NSAID interactions as a high-volume safety signal. Warfarin's high count is structural: its narrow therapeutic index and shared CYP450 metabolism create a long list of clinically meaningful interactions, documented across PubMed-indexed pharmacology reviews.

Finding 3: 43.6% of patient assistance programs cap eligibility at 400% of the Federal Poverty Level

Headline: Modal PAP income floor is 400% FPL (24 of 55); 25.5% extend to 500% FPL

The modal income floor across the 55 OmniRx-tracked patient assistance programs is 400% of the Federal Poverty Level, used by 24 programs (43.6%). Another 14 programs (25.5%) extend eligibility to 500% FPL, and 5 (9.1%) reach 600% FPL or higher. Four programs (7.3%) have no income test, typically discount cards or copay programs sponsored by the manufacturer for commercially insured patients. The 400% FPL line corresponds to roughly $60,240 in annual income for an individual in 2025 (per HHS poverty guidelines), the same threshold used for ACA premium subsidies before the American Rescue Plan extension. The clustering at this line is not coincidence: PAP designers anchor on visible regulatory thresholds because they reduce administrative friction at intake. KFF analysis of out-of-pocket burden in Medicare Part D underscores why this matters: patients above 400% FPL but below comfortable-private-pay thresholds fall into a coverage gap that PAPs only partially fill.

Finding 4: 65.5% of patient assistance programs are manufacturer-run, not independent

Headline: 36 manufacturer, 13 nonprofit, 4 federal, 2 discount-card

Manufacturer-run programs make up 65.5% of the OmniRx PAP catalog (36 of 55). Nonprofits account for 23.6% (13 of 55), federal programs 7.3% (4 of 55), and commercial discount cards 3.6% (2 of 55). The dominance of manufacturer programs is structurally important: a manufacturer PAP is a marketing-spend line item, not a charity, and is typically restricted to the manufacturer's own brand portfolio (Pfizer RxPathways covers Pfizer drugs; Lilly Cares covers Lilly drugs). Nonprofits like the PAN Foundation and Partnership for Prescription Assistance cover broader portfolios but exhaust funding cycles for many disease states. The federal layer (Medicare Extra Help, Medicaid, VA) catches the lowest-income patients but leaves the 200-400% FPL band to a patchwork of manufacturer programs that the patient or pharmacist must individually navigate. The structural critique embedded in the FTC PBM report applies here too: opacity benefits the intermediary, not the patient.

Finding 5: U.S. insurers run formularies on three different tier counts, plus closed formularies

Headline: 4-tier (Kaiser), 5-tier (Aetna, Cigna, BCBS, Humana, Part D), 6-tier (UHC, Anthem), closed (VA, TRICARE)

The 15 insurer-level formularies in the OmniRx database run on three different tier counts: Kaiser uses a 4-tier system, Aetna, Cigna, most BCBS plans, Humana, and standard Medicare Part D use 5 tiers, and UnitedHealthcare and Anthem use 6 tiers. VA and TRICARE run closed formularies with non-comparable structures. This means a patient comparing coverage across two plans cannot map "Tier 3 on Aetna" to "Tier 3 on UHC" without translating through the tier-count difference. The structural variance is the mechanism behind the patient-pays-more-for-the-same-drug pattern that KFF Health System Tracker and The Commonwealth Fund consistently document. CMS publishes Medicare Part D formulary files in a standardized format precisely because cross-plan comparability is not native to the commercial market. Drug Channels Institute coverage by Adam Fein has tracked the slow drift toward 6-tier systems as PBMs add specialty and preferred-specialty splits to capture rebate-eligible placement separately from cost-share placement.

Finding 6: 12 drugs sit at the intersection of clinical risk and PBM pricing variance

Headline: 48% of OmniRx drug-coverage entries (12 of 25) overlap the interaction catalog

Cross-tabulating the interaction database against the drug-coverage formulary records returns 12 drugs that appear in both: Ozempic, insulin, Adderall, metformin, atorvastatin, levothyroxine, gabapentin, lisinopril, omeprazole, sertraline, amoxicillin, and birth control. These drugs are simultaneously high-prevalence (so a patient is likely to be taking them), clinically interactive (so polypharmacy compounds risk), and pricing-variable (so insurer choice affects out-of-pocket cost). The overlap is the operational definition of where PBM opacity costs patients the most. The FTC's January 2025 second interim report singles out insulin specifically, finding that the three largest PBMs marked up several specialty generic drugs by hundreds and even thousands of percent above acquisition cost.

Finding 7: GLP-1 and weight-loss drugs occupy 20% of OmniRx's drug-coverage records

Headline: 5 of 25 drug-coverage entries are GLP-1 or weight-loss (Ozempic, Wegovy, Mounjaro, Zepbound, +1)

Five of the 25 OmniRx drug-coverage entries (20%) cover GLP-1 receptor agonists or weight-loss-indicated drugs: Ozempic, Wegovy, Mounjaro, Zepbound, plus diabetes-coded indications that appear on multiple GLP-1s. The concentration reflects the 2024-2026 demand surge that has reshaped formulary politics. KFF analysis documents that GLP-1s alone now account for a meaningful share of total Medicare Part D drug spending growth, and CMS has signaled that step-therapy and prior-authorization protocols for GLP-1s will be reviewed in 2026 plan-year rule-making. For patients, the practical effect is that the same active ingredient (semaglutide) is covered for one indication and not another (Ozempic for diabetes is covered, off-label for weight loss is not), and the brand-vs-brand routing through a PBM determines whether Wegovy, Zepbound, or compounded alternatives are accessible.

Finding 8: Eliquis and Jardiance are each covered by 4 separate patient assistance programs

Headline: PAP overlap on cardiometabolic blockbusters; 4 each for Eliquis and Jardiance

Eliquis and Jardiance each appear on 4 different patient assistance programs in the OmniRx database, the highest overlap rate in the catalog. Eleven other high-prevalence drugs appear on 3 PAPs each: Mounjaro, Trulicity, Humalog, Farxiga, Symbicort, Imbruvica, Xarelto, Ozempic, NovoLog, Tresiba, Humira, Rinvoq, and Opdivo. The overlap is intentional manufacturer marketing: a high-rebate, brand-tier drug attracts both the manufacturer's own PAP (Bristol Myers Squibb Patient Assistance Foundation for Eliquis, Lilly Cares for Jardiance) and downstream nonprofit programs that backstop the manufacturer program when income limits or insurance status disqualify the patient. The pattern is documented in Drug Channels Institute coverage of co-pay accumulator and maximizer programs, and the consumer-facing implication is that a patient on Eliquis has more PAP options than a patient on warfarin even though warfarin is generic and far cheaper.

Finding 9: Severe + Moderate combine for 79.2% of catalog interactions

Headline: 42 of 53 interactions are clinically actionable

Severe (49.1%) and Moderate (30.2%) interactions together account for 79.2% of the OmniRx catalog (42 of 53). The remaining 20.8% split between Mild (13.2%) and None or theoretical (7.5%). For a typical U.S. patient on three or more chronic medications (a population that CDC NCHS estimates at roughly 24% of adults), the practical implication is that an interaction flag in the OmniRx catalog is more likely than not to warrant a conversation with a pharmacist. We do not present this as medical advice; we present it as a calibration on how to read the catalog.

Finding 10: The same drug at the same wholesale cost can route through wildly different patient-out-of-pocket paths

Headline: Tier system + prior-auth + step-therapy + PAP availability = patient cost variance

This is the BLUF re-stated as a pricing thesis. A drug's wholesale acquisition cost is a small fraction of what determines a patient's out-of-pocket cost. The chain runs: PBM negotiates a list price and a rebate (rebate retained partly by PBM, partly passed to plan sponsor), insurer assigns a tier on its formulary (4-, 5-, or 6-tier system), prior authorization or step therapy may be required, the patient's deductible and coinsurance phase determines the cost-share at point of sale, and a PAP may or may not exist to absorb residual cost. Mark Cuban Cost Plus Drugs is the operational counter-example: a transparent cost-plus-15%-plus-pharmacy-fee model that bypasses the PBM-rebate chain entirely for a curated drug list. The Cost Plus model is also documented in coverage by KFF Health News and STAT as the most credible existing test of the hypothesis that PBM rebate retention, not wholesale cost, is the dominant driver of patient-paid prices for the maintenance medications most patients take.

What this means for patients

Republish this chart

<iframe src="https://omnirx.org/research/pbm-spread-pricing-2026/embed/severity-distribution.html" width="100%" height="440" frameborder="0" loading="lazy" title="Drug interaction severity distribution across 53 OmniRx records, 2026"></iframe>

Press kit and downloads

Full press kit at /press/, machine-readable aggregates at data.json. Media inquiries via the contact on the press page. We are particularly interested in inquiries from KFF Health News, STAT, ProPublica, and other PBM-focused investigative desks.