Dataset

Specialty Drug Rates

What each insurer pays for every specialty and infusion drug, and for administering it.

What do insurers pay for this drug, and for giving it?

Negotiated rates for every specialty and infusion drug code, plus the administration codes, measured against the federal benchmark.

What you getRate distributions by insurer, state, code and setting

What it coversSpecialty and infusion drugs, plus administration codes. National.

What it is joined toRates, with Medicare, ASP+6% and acquisition cost attached

Buy the slice you need

You do not have to take the whole grid. Every slice is the same columns, the same quality gates and the same quarterly refresh over less of it — and slices stack, so nothing is ever bought twice.

One molecule, every insurer

A single drug across all 37 insurers and every state. The floor price, and the usual way a launch team starts.

$6,000a year

Get this

One therapeutic area

Every drug in a class — the immunology set, the bone agents, the anti-emetics — across all 37 insurers.

$6,000a year

Get this

One state, every drug

The whole specialty set in a single state. For a provider who only bills in one place.

$8,000a year

Get this

All specialty and infusion drugs

The whole drug set with the administration codes, all 37 insurers, all 50 states and DC. This is the standard buy.

$16,000a year

Get this

What is in it

Every specialty and infusion drug code, plus the five administration codes (96413, 96415, 96365, 96360, 96372)
37 insurers across all 50 states and DC
Split by setting — doctor’s office, hospital outpatient, home
Each rate next to Medicare, ASP+6% and what the drug costs to buy
A new build every quarter, in Parquet and CSV

Questions it answers

What does this insurer pay for zoledronic acid in Texas, and how does that compare with Medicare?
Which of our top ten drugs are paid below the federal benchmark, and in which states?
What is a practice actually paid for the first hour in the chair, separately from the drug?

What it does not do

It will not tell you what one named practice is paid.

These are market-level ranges by insurer, state and code. Provider-level detail is a different resolution and is in development.

It will not tell you what an insurer requires before paying.

That is Payer Coverage Requirements.

Who buys this

Infusion centers and specialty practices pricing their next contract
Field reimbursement and market access teams at drug manufacturers
Investors underwriting an infusion or oncology platform

Why it is worth the money

The administration codes are included

Most rate data covers the drug and stops there. This includes CPT 96413, 96415, 96365, 96360 and 96372, so you can see what a practice is paid for chair time as well as for the drug. No other published dataset benchmarks these.

Every drug rate sits next to the federal benchmark

Each row shows what commercial pays as a percentage of ASP+6%. You can tell in one column whether an insurer pays above or below what Medicare would.

You get the range, not just an average

Every row carries the 10th through 90th percentile, how wide the spread is, and how many providers are behind it. You can see straight away whether a number rests on three rates or three thousand.

Drug cost is attached

Acquisition cost sits in the same row as the rate, so you can work out margin without buying a second dataset.

What a row looks like

Real rows from the current build, in the shape this dataset ships.

Real rows · 2026-Q2-06 rate_market, the production table · as of 21 August 2026

Insurer	State	Rates	p10	Median	p90	Quality	Flagged
uhc	NY	6,221	$23.36	$33.23	$54.72	72.5	no
uhc	CA	6,092	$25.52	$34.83	$64.87	72.5	no
uhc	FL	5,923	$24.03	$32.99	$58.86	72.5	no
aetna	FL	5,611	$30.03	$1,591.69	$4,227.98	65.0	yes
uhc	TX	5,472	$24.41	$33.22	$57.82	72.5	no
aetna	CA	5,329	$31.64	$1,654.20	$4,295.29	65.0	yes

Denosumab, professional claims, six real rows out of the table you would license. Watch rows four and six: Aetna files this drug per vial where UnitedHealthcare files it per milligram, so the same drug reads as $1,591 instead of $33. The pipeline caught it, flagged the row and recorded how it was corrected — those rows ship flagged rather than deleted, so you can see what we would have excluded.

The columns

Published in full, before you talk to anyone. Whether this fits your model is a technical question and you should be able to answer it yourself.

`rate_intel`

Specialty and infusion rates at tax-ID level, with the ghost count, the confidence interval and the federal benchmark already on every row. This is the production table, column for column.

Key vintage × payer × state × billing_code × billing_code_modifier × billing_class × tin

Column	Type	What it is
`vintage` key	TEXT	Which quarterly build the row belongs to. Current is 2026-Q2-06.
`payer` key	TEXT	The insurer the rate is attributed to.
`src_payer`	TEXT	The file it came out of. Kept because one insurer’s file can carry rates for several.
`state` key	TEXT	Two-letter state.
`billing_code` key	TEXT	HCPCS or CPT.
`billing_code_modifier` key	TEXT	Part of the key. Averaging across modifiers gives a wrong number that looks right.
`service_code`	TEXT	Place-of-service grouping as the insurer filed it.
`billing_class` key	TEXT	professional or institutional.
`negotiated_type`	TEXT	negotiated, fee schedule, derived, or blank.
`tin` key	TEXT	Tax ID. This is what makes the table entity-level rather than market-level.
`n_obs`	BIGINT	Observations behind the row after filtering.
`n_raw`	BIGINT	Observations before filtering.
`n_ghost`	BIGINT	How many were ghost rates — codes that provider would never bill. Counted, not silently dropped.
`n_unit_uncertain`	BIGINT	Rows whose billing unit could not be resolved. Per-vial against per-milligram lives here.
`rate_raw`	DOUBLE	Before unit normalization.
`rate_final`	DOUBLE	After. This is the number to use.
`sd_used`	DOUBLE	Standard deviation behind the interval.
`ci_low`	DOUBLE	Lower bound of the confidence interval.
`ci_high`	DOUBLE	Upper bound. A wide interval is telling you something.
`veracity`	TEXT	measured where the rate is observed, otherwise how it was derived.
`proxy_source`	TEXT	What stood in when the rate was not directly observed.
`confidence`	DOUBLE	0 to 1. Set your own floor; ours is high.
`medicare_bench`	DOUBLE	Medicare allowed for the same code and locality.
`asp_per_unit`	DOUBLE	CMS average sales price. Drugs only, null for procedures.
`asp_plus6`	DOUBLE	ASP plus 6 percent, the federal drug benchmark.
`ratio_vs_medicare`	DOUBLE	The rate over the Medicare allowed amount.
`ratio_vs_asp6`	DOUBLE	The rate over ASP+6%. Under 1 means commercial pays less than the federal benchmark.
`anchor_source`	TEXT	What the row was anchored against, usually medicare.
`triangulation`	TEXT	Whether independent sources agreed.

Traps in this data, and how we handle them

These catch people out whoever they buy from. They are worth knowing before you model anything with rate or coverage data, including ours.

Most rows in the raw files are rates nobody would ever bill

Why it bitesPublished research on Transparency in Coverage data found the median insurer’s file was 95.7% ghost rates — provider and code pairs like a podiatrist priced for open-heart surgery. Across 61 insurers, 95.4% of pairs were ghosts. An average taken straight off a raw file is mostly noise.

How we handle itEvery row carries a distinct provider count and a 0 to 100 quality score, and we flag rather than delete, so you can set your own floor and see what you excluded. Our own published work uses 70 and above.

Some rates are a percentage, not a dollar amount

Why it bitesInsurers must publish a dollar figure wherever one can be calculated in advance. Where the contract is genuinely a percentage of billed charges, they publish the percentage instead. A 45 sitting next to a 4,500 is not a cheap version of the same thing.

How we handle itThe rate type travels with the value in every row. We never silently convert a percentage into dollars, and percentage rows are kept out of dollar comparisons rather than quietly averaged in.

A contracted rate is not what got paid

Why it bitesThese files describe what an insurer agreed to pay. They say nothing about deductibles, coinsurance, denials, retroactive adjustments or what actually landed on the remittance.

How we handle itWe say so on every page rather than in a footnote. If you need what was paid rather than what was agreed, this is the wrong dataset and we will tell you so before you buy it.

One insurer publishes many networks

Why it bitesA large insurer files rates for PPO, HMO, exchange, rental and administrative-services-only networks. Blending them produces a single rate that no provider is actually paid.

How we handle itNetworks stay distinct. Where you want one number we tell you which network it came from rather than averaging across products that are not comparable.

Averaging across modifiers gives a wrong number that looks right

Why it bitesJW and JZ change how drug wastage is billed. 26 and TC split a professional read from the equipment. Collapse them and the distribution shifts without any obvious sign.

How we handle itModifier is part of the key. The schema says so above the columns, and the sample is built so you can see the effect yourself.

Insurer names on multi-state files are not always what they seem

Why it bitesA file downloaded into one folder can carry a different company name inside it, which matters for plans operating across several states.

How we handle itWe are re-deriving names from file contents. Until that is finished every row carries a confidence column, so you can keep only the attribution you trust.

What to know before you buy

We are re-checking insurer names on files from plans that operate in several states. Every row carries a confidence column so you can keep only the attribution you trust.
Quarter-to-quarter change columns arrive with the next build. Today the file is a single point in time.
Six insurer groups are not in this build. They are named on the coverage page.

Load it against your own model

Three states, the production columns, real rows, downloaded directly with no form in front of it. It is the fastest way to answer the fit question.

Download the free sample