Teleradiology pricing per read: what sets the rate in 2026

Two teleradiology quotes are on your desk, $14 apart on CT, and nothing on either page explains the gap. Teleradiology pricing per read looks like a single number, but the number is the output of a cost stack, and understanding the stack is what lets you tell a genuinely lean rate from a stripped one that grows surcharges after signature.

This page takes the per-read rate apart: what it funds, why it varies about eightfold across modalities, what moves it between vendors, and where per-read structure beats its alternatives. Published market figures by modality live on the companion page teleradiology cost per read; here the subject is the mechanics underneath those figures.

One disclosure before the mechanics, because this page is written by a vendor. AstraRad sells exactly this product: final signed reports from board-certified, fellowship-trained subspecialists, billed per report with no minimums, subscriptions, or platform fees, and its per-read amounts appear on a written rate card rather than on this page. The analysis below applies to any bidder's quote, including ours.

What does a per-read rate actually pay for?

Five inputs, in descending order of size.

Radiologist time. The dominant input. A fellowship-trained radiologist's interpretation time scales with study complexity, and the cleanest public proxy for that scaling is the Medicare work RVU assigned to each CPT code in the CMS Physician Fee Schedule relative value files. A two-view chest X-ray (CPT 71046) carries a work RVU of 0.22; a CT head without contrast (70450) carries 0.85; an MRI brain without contrast (70551) carries 1.48. That is roughly a sevenfold spread in valued physician work between the cheapest and the most expensive common study, and it is why no honest vendor prices all modalities at one rate.

Subspecialty scarcity. A generalist can be scheduled; a pediatric neuroradiologist at 2 a.m. must be planned for. Vendors that route by fellowship carry a wider, more expensive panel, and the rate reflects it. Which studies actually benefit from the match is mapped in subspecialty reads by study type.

Licensing and credentialing overhead. Every reader needs an active license in every state they read for, renewed on that state's cycle, plus credentialing at every client facility. A 40-state panel is a standing administrative operation, and it is funded out of the per-read rate.

Quality infrastructure. Double-read programs, discrepancy review, peer review reporting. This input varies most between vendors: at some it is a real cost center inside the rate, at others it is absent, which is one reason their rate is lower.

Platform, insurance, and margin. PACS gateways, HL7 interfaces, malpractice premiums for the panel, and the vendor's margin. Malpractice for remote reads prices per reader per state and is invisible to the buyer until it lapses, which is why credential files belong in procurement.

The practical use of this list: when a quote undercuts the market band, ask which of the five inputs it removed. Sometimes the answer is efficient operations at scale. Sometimes it is a generalist panel, no QA program, or a preliminary read masquerading as a final. The question costs one email.

What drives teleradiology pricing per read up or down?

Holding the vendor constant, four variables move your rate. Each belongs on the rate card as a printed number, and the difference between a printed number and a discretionary one is most of what contract review is for.

Variable	Typical market effect	What to demand in writing
Priority tier	STAT adds roughly 20 to 50 percent over routine	One fixed multiplier, printed, applied automatically
Committed volume	Steps the rate down a tier ladder	The ladder itself, and what happens when you miss a tier
Study complexity	Multi-region and advanced studies price above base modality	Every study type as its own line, all on one page
After-hours timing	Separate surcharge at some vendors	Either no surcharge, or its exact amount and hours

Volume tiering deserves one concrete illustration because it is the variable buyers most underestimate. RADIEWCARE, a European provider and one of the few anywhere publishing a full tier table, lists single-region CT at 23 euros for facilities under 100 monthly studies, stepping to 10.50 euros above 5,000, checked September 3, 2026. That is a 54 percent spread on the identical study, driven entirely by volume band. US vendors quoting privately run the same mechanics without publishing them, which means a small facility comparing its quote to a big-system anecdote is comparing different tiers, and the anecdote is not evidence of overcharging.

The timing surcharge is the variable to treat most skeptically. Reading at 3 a.m. genuinely costs more to staff well, but a vendor already charging a STAT multiplier and then stacking an after-hours fee on the same study is charging twice for one scarcity. AstraRad's position, for calibration: one printed multiplier for STAT under 1 hour and urgent under 4 hours, no nights, weekend, or holiday surcharge, and no surge premium when volume spikes, because the panel is staffed on scheduled US shifts around the clock. The full commercial structure is on the pricing page, and a written rate card with every study type on one page reaches you within one business day of a request.

Why do published rates cluster where they do?

The few public US price points anchor the market's floor. NDX Imaging publishes starting rates of $12 for X-ray, $28 for ultrasound, $32 for mammography, $40 for CT, $60 for MRI, $38 for nuclear medicine, and $99 for PET-CT, checked September 3, 2026, with the note that rates vary by monthly volume and expected turnaround. Set those against the work RVU spread above and the structure is visible: the published X-ray-to-MRI ratio (5x) tracks the work RVU ratio (about 7x) with the fixed per-study costs of QA, connectivity, and administration compressing the low end, because a $12 X-ray and a $60 MRI cost the vendor the same amount to route, deliver, and archive.

Three cautions when using any published number. A "starting at" figure generally assumes committed volume and a routine tier, so treat it as the bottom of a band. A rate that low may describe a generalist read at some vendors, and the subspecialty version sits higher. And five or so study types on a typical card, the multi-region CTs, coronary CTA, cardiac MRI, and advanced MRI, carry no published market figure at all, which makes them exactly the lines to demand in writing before you sign; the modality-by-modality detail is on teleradiology cost per read.

Where a quote sits far below the band, the missing money usually reappears as a platform fee, a minimum, or an addendum charge. The catalog of those recovery mechanisms is in teleradiology hidden fees, and it is worth reading before any rate on any card looks attractive.

How do you benchmark a per-read quote against Medicare?

The one public, national, study-level price reference in US radiology is the Medicare Physician Fee Schedule, and it makes a usable floor for any per-read quote if you benchmark against the right component.

The method takes about an hour for a full modality mix. Pull your ten highest-volume CPT codes from last quarter's billing. Look each one up in the CMS Physician Fee Schedule search tool and record the professional component, the modifier 26 amount, which prices the interpretation alone. The global amount includes the technical component, the scanner and technologist costs that stay with your facility, so benchmarking a teleradiology quote against the global figure makes every vendor look cheap and tells you nothing. Set each vendor line against its CPT professional component and you have a defensible spread for every study type you actually send.

Read the spread with two facts in mind. First, a teleradiology rate near or below the professional component is buying interpretation as a commodity; the Medicare figure funds no overnight availability, no contractual turnaround tier, no subspecialty routing, and no QA reporting, and a vendor matching it must be cutting one of those or making it up elsewhere on the invoice. Second, a rate meaningfully above the professional component is not automatically padded: a 2 a.m. final read from a neuroradiologist inside a one-hour SLA is a different product from a next-week interpretation, and the difference is what the premium prices. The benchmark's job is to size the premium so you can ask what it buys, line by line, with the vendor's answers in writing.

One caution on drift: the fee schedule updates every January, and its conversion factor has moved down in multiple recent years, so a benchmark built in one year overstates the floor the next. Date the spreadsheet and rebuild it annually alongside the contract review.

Teleradiology pricing per read versus the alternatives

Per-read is one of four structures a bidder might quote, and each has a volume profile where it wins.

Per read. You pay for signed reports and nothing else. Wins at variable, seasonal, or growing volume, because a slow month simply invoices less. Its weakness is budgeting precision: finance teams that want one fixed number dislike a line that moves with clinical demand.

Flat monthly fee. One number for a capacity band. Wins on budget predictability and can win on unit cost when your volume sits reliably near the band's ceiling. Every study you fall short converts directly into a higher effective rate; a 900-study fee spread over 600 studies costs half again as much per study as the same fee at capacity.

Per work RVU. The rate tracks study complexity automatically, which is fair in both directions, and reprices silently when your mix shifts toward cross-sectional imaging. Requires the buyer to compute a blended work RVU per study to compare against anything else.

Radiologist FTE equivalent. Buying scheduled reader hours. Wins only at volumes high and steady enough to fill the schedule; below that it is idle capacity billed monthly.

The full worked comparison, with the same month priced under three structures, is in per-report versus subscription pricing. The one-line summary: reduce every quote to expected cost per signed study at your real volume, then rerun it at 80 percent of that volume, and let the slow-month number decide.

Reading a per-read quote in ten minutes

A short procedure, in order.

1. Count the study types on the card. Twelve or more lines mean the vendor prices complexity honestly; four lines mean multi-region and advanced studies will be priced by exception later, after you have signed.
2. Find the STAT multiplier. It should be one printed factor. If STAT pricing is "quoted case by case," your emergency volume has no contract.
3. Ask whether the rate buys a final signed report on every tier. The difference between prelim and final pricing is a second interpretation someone must fund, explained in STAT versus preliminary versus final reads.
4. Itemize the six inclusions: final report, subspecialty routing, critical-results calls, QA participation, free addenda, connectivity. Price any that sit outside the rate.
5. Ask for the volume ladder and the consequence of missing a tier.
6. Compute cost per signed study at your volume and at 80 percent of it.

A vendor that survives all six steps with printed answers is quoting a price. One that survives three is quoting an estimate, and the gap between the two becomes your problem on the first invoice, not theirs. Keep the six answers from every bidder in one spreadsheet; when the finalists are two rates a few dollars apart, the spreadsheet usually shows they are not the same product at all, and the decision stops being about the rate.

AstraRad's answers to the six are already in writing: per signed report by study type, one printed priority multiplier, final reads at every tier from board-certified fellowship-trained subspecialists, all six inclusions inside the rate, no minimums or platform fees, and 99.4 percent of reports inside their SLA tier over the trailing 12 months, measured from last-image arrival to signature and published on the SLA page. The rate card itself takes one business day from a request.