Budesonide should be boring. It is a synthetic corticosteroid that has been on the market for decades, used for conditions spanning pulmonary medicine and gastroenterology — COPD, asthma, Crohn’s disease, ulcerative colitis, eosinophilic esophagitis. It is available as a dry powder inhaler, a nebulization suspension, an oral capsule, a rectal foam, and an extended-release tablet. It is off-patent, multi-source, and thoroughly generic. There is nothing exotic about the molecule, the pharmacology, or the clinical use. What makes budesonide interesting is what happens when you try to determine its price.
The answer, it turns out, depends on which database you consult, which formulation you specify, and which unit convention the data publisher applies — and the results are not merely different. They are irreconcilable without substantial analytical work that no federal system performs automatically.
Start with the NADAC
entry for budesonide inhalation suspension 0.5 mg/2 mL, 30 ampules. The unit price is published per milliliter, reflecting the NCPDP billing convention for liquid formulations. Now look at the WAC for the same product. The unit price may be published per ampule, per package, or per gram of active ingredient, depending on the pricing compendium. Pull the ASP for budesonide from CMS’s quarterly files, and the unit convention shifts again — CMS publishes ASP using its own unit definitions, which may correspond to a single billing unit, a per-milligram price, or a per-dose figure depending on the HCPCS code.
The resulting per-unit prices across these three datasets are not comparable without normalization. The numbers differ not because the underlying drug costs differ — they may or may not — but because the denominator changes. One database reports cost per milliliter. Another reports cost per ampule. A third reports cost per billing unit that may or may not correspond to either of the other two. The prices occupy different numerical scales, and comparing them directly produces results that are arithmetically meaningless.
This is the problem that Daily Remedy encountered during a routine internal audit of its analytics platform — a discrepancy flagged precisely because the team was cross-referencing WAC, NADAC, and ASP systematically. Most analytics platforms do not attempt this reconciliation for every product. They accept the unit conventions published by each dataset and compare within datasets rather than across them. The cross-dataset comparison is where the structural incompatibilities become visible.
The instinct is to normalize — to convert all unit prices to a common denominator and then compare. But which denominator? Converting to NCPDP billing units (EA, ML, or GM) standardizes the physical quantity but does not address therapeutic non-equivalence. A milliliter of nebulization suspension and a milliliter of rectal foam deliver budesonide through different routes, at different bioavailabilities, for different clinical purposes. Normalizing their prices per milliliter makes the numbers comparable while making the comparison clinically meaningless.
Stoichiometric normalization — converting to cost per milligram of active pharmaceutical ingredient — offers a molecule-level comparison but sacrifices the clinical context entirely. The delivery mechanism, the excipients, the formulation technology, and the route of administration all contribute to therapeutic effect and manufacturing cost. A price per milligram of budesonide that ignores whether the milligram is delivered via a metered-dose inhaler or an enteric-coated capsule is a number without clinical meaning.
The course-of-treatment approach, which SSR Health and others have adopted, addresses the clinical dimension by rolling unit prices into a cost per treatment course based on labeled dosing. But budesonide’s multi-indication profile fractures this approach. An eight-week induction course for eosinophilic esophagitis costs a different amount than a three-month course for Crohn’s disease, which costs a different amount than indefinite maintenance therapy for COPD. Each is a legitimate course of treatment. Averaging across them produces a composite that does not describe any single patient’s experience. Reporting them separately produces multiple prices for the same molecule that differ by indication — accurate, but unwieldy as a comparative metric.
What budesonide reveals is not a flaw specific to budesonide pricing but a structural characteristic of the American drug pricing data ecosystem. Every multi-formulation, multi-indication drug encounters the same set of problems to varying degrees. Budesonide simply makes them visible because its formulations span both retail and physician-administered channels, because its unit conventions diverge across all three major pricing databases, and because its clinical use varies enough to expose the limitations of standardized dosing assumptions.
The question posed in the email thread between Daily Remedy and Charm Economics — whether normalizing to NCPDP billing units is a sound first step, or whether it merely pushes the problem upstream — does not have a clean answer. It is a sound first step in the sense that NCPDP units provide a physical standard that eliminates packaging-level ambiguity. It is insufficient in the sense that physical standardization does not solve the clinical and economic non-equivalence that different formulations represent.
The deeper question is whether the American drug pricing system should have a reconciliation layer — a federally maintained crosswalk that maps every NDC across every pricing dataset to a common unit, with explicit handling of multi-formulation products and clear documentation of conversion assumptions. The technical requirements for such a system are well understood. The institutional will to build it is absent. CMS, which publishes multiple pricing datasets using different unit conventions, has not prioritized unit harmonization. The FDA maintains the NDC directory but not the pricing crosswalks. Manufacturers report to different programs using whatever units each program requires.
The result is that every organization attempting cross-dataset drug pricing analysis — payers, PBMs, analytics vendors, researchers, journalists — must build its own reconciliation infrastructure from scratch. Each solution is bespoke, undocumented, and potentially inconsistent with every other solution. The industry has chosen, implicitly, to distribute the cost of reconciliation across thousands of individual users rather than centralizing it in a shared standard. The aggregate cost of this distributed approach — in duplicated effort, in undetected errors, in analytical conclusions that diverge because underlying unit assumptions diverge — is impossible to quantify and almost certainly enormous.
Budesonide does not need a better price. It needs a better infrastructure for comparing the prices it already has.
