Teleradiology in 2026: how it actually works

Your facility produces imaging around the clock, and radiologist hours do not stretch to match it. Teleradiology closes that gap: studies leave your PACS, a licensed radiologist elsewhere reads them, and a report returns into your systems, on a clock you can hold the vendor to.

This guide covers the entire operation as it runs in 2026: the path a study travels, the read tiers, the turnaround clocks, the licensing and credentialing work that actually sets the timeline, the quality machinery, and the costs. It is written for the person who signs the contract, and it names the trade-offs a sales deck will not.

One disclosure before the detail. AstraRad is a subspecialist teleradiology practice: board-certified, fellowship-trained radiologists reading final signed reports, billed per report with no minimums and no platform fees. The operational claims below about how this should work are the standards we hold ourselves to, and a written rate card reaches you within one business day of a request. Everything else in this guide applies to any vendor you evaluate, including us.

What is teleradiology and who buys it?

Teleradiology is the interpretation of imaging studies by a radiologist who is not at the facility where the images were acquired. The images move; the physician does not. The practice is decades old, but three things changed its role from a night-shift patch to core infrastructure: PACS became universal, bandwidth stopped being a constraint, and the radiologist supply stopped keeping pace with imaging volume. The Association of American Medical Colleges projects a US physician shortage of up to 86,000 by 2036, and radiology sits squarely inside it; imaging volume keeps growing while residency slots barely move. Grand View Research puts the global teleradiology market at about 19.2 billion dollars in 2025 with double-digit annual growth, per its teleradiology market report, checked September 3, 2026.

The buyers fall into four groups, and their problems differ:

Buyer	The gap teleradiology fills
Hospitals and health systems	Nights, weekends, subspecialty depth, and surge volume the employed group cannot absorb
Outpatient imaging centers	Full reading coverage or overflow, without employing a radiologist for variable volume
Radiology groups	Overnight final reads so partners are not re-reading prelims at 7 a.m., plus vacation and leave cover
Urgent care and orthopedic clinics	Same-day X-ray and ultrasound over-reads at plain-film volumes no local group wants

If you are deciding whether to outsource at all, start with the arithmetic in in-house vs outsourced radiology. The rest of this guide assumes you have a gap to cover and want to know how the machine works.

The path a study takes, from scanner to signed report

Every teleradiology read follows the same seven steps. What separates vendors is how fast and how visibly each step runs.

1. Acquisition. Your technologist completes the study exactly as before. Nothing about image capture changes, which is why technologist workflow disruption is the wrong fear; the right questions are about steps four and seven.

2. Transmission. The site's PACS or the modality itself pushes the study by DICOM to the vendor, over site-to-site VPN or TLS. Priors and the order information travel with it, because a read without priors is a worse read: the American College of Radiology's guidance on teleradiology practice is explicit that interpretation should happen with relevant prior studies and clinical information available. Low-volume sites often skip interface work at first and upload through a secure portal.

3. Validation. The receiving platform confirms the study arrived complete: series counts, image counts, and a match against the order. Studies that fail validation generate a callback to your technologist. Ask any vendor how incomplete studies are handled, because a study sitting unread for want of a missing series is the most common silent turnaround failure.

4. Routing. Software assigns the study to a radiologist worklist. In a generalist operation, routing is availability-based: the next free reader takes it. In a subspecialty operation, routing keys on modality and body part first, so a brain MRI reaches a neuroradiologist and a child's abdominal CT reaches a pediatric radiologist. Priority sits on top: STAT studies jump every queue. At AstraRad, routing is subspecialty-first across a panel of 240 fellowship-trained radiologists in ten subspecialties, and the priority tier is set by your ordering workflow rather than negotiated per study.

5. Interpretation. The radiologist reads with priors, dictates, and signs. Two details are worth contract language. First, whether the signer is the reader: ghost reading, where one physician interprets and another signs, is explicitly unacceptable under ACR guidance. Second, critical findings: a suspected critical result should generate a phone call to your care team within minutes, documented on the report with time and recipient.

6. Delivery. The signed report returns as an HL7 ORU message into your RIS and EMR, with PDF delivery wherever you want a copy. FHIR-based delivery is increasingly common alongside HL7 v2, though in 2026 almost every production results interface in US radiology still speaks ORU. The interface mechanics are covered in PACS integration for teleradiology.

7. Follow-through. Addenda, clarification requests, and discrepancy handling. A referring physician who questions a finding needs a path back to the interpreting radiologist, and that path should not be a ticket queue. Ask who answers a clinical question about a signed report, and how fast.

Prelim or final: which teleradiology product are you buying?

The single most consequential line in a teleradiology contract is whether the read is preliminary or final. A preliminary read answers the acute clinical question overnight and leaves the billable final interpretation to your own radiologists in the morning. That model exists because it is cheap for the vendor, and it made sense when overnight teleradiology was staffed by moonlighters. Its cost lands on you: every overnight study is handled twice, morning worklists start the day behind, and discrepancies between prelim and final create their own quality workload.

A final read carries the interpreting radiologist's signature and is done when it is delivered. It requires the reader to be licensed in your state, credentialed at your facility if you are a hospital, and enrolled where billing requires it, which is why final-read services carry more administrative setup and generally higher per-study prices. The full comparison, including the discrepancy data between prelim and final interpretation, is in STAT vs preliminary vs final reads.

The market has moved firmly toward finals over the past decade, and buyer expectations moved with it. AstraRad only delivers final signed reports; there is no preliminary tier on the rate card at any hour. If a vendor quotes you a number that looks 30 percent below market, the first thing to check is whether that number buys a prelim.

How fast should teleradiology turnaround be?

Turnaround is a contract term first and a performance number second, and the two should be read separately. The contract sets a ceiling per tier. The measurement tells you where studies actually land. A vendor that publishes only one of the two is asking you to guess the other.

Typical US contract tiers in 2026:

Tier	Common contractual ceiling	What clinically depends on it
STAT	30 to 60 minutes	Stroke, trauma, acute chest and abdomen, any ED disposition decision
Urgent	2 to 4 hours	Inpatient workups, urgent care dispositions, same-visit decisions
Routine	24 to 48 hours	Outpatient and screening volume

Three questions separate a real commitment from a brochure number. What event starts the clock: last-image arrival is the honest trigger, and any definition that starts on "study assigned" or "study opened" hides queue time you will experience but the vendor will not report. What event stops it: radiologist signature, and never "report available." And what fraction of studies land inside the tier: ask for the compliance percentage over a trailing period, and ask for the median as well as the ceiling, because the median is what your 2 a.m. actually feels like.

AstraRad's tiers are STAT under 1 hour, urgent under 4 hours, and routine under 24 hours, measured from last-image arrival to radiologist signature. The measured median STAT turnaround is 30 minutes, remeasured monthly, and 99.4 percent of reports came back inside their tier over the trailing 12 months. The definitions and the current numbers live on the SLA page, which is the canonical dated source.

Licensing and credentialing set your timeline

Technology connects in days. Paperwork takes weeks, and it is the part of a teleradiology launch most often underestimated. Two separate regimes apply.

State licensure. The interpreting radiologist must hold a license issued or recognized by the state where the patient is located at the time of the study. This is per reader, per state, with no national teleradiology license. The Interstate Medical Licensure Compact has compressed the timeline for physicians who qualify: as of 2026 the compact covers 44 states plus Washington DC and Guam, per the IMLC Commission, and an expedited compact license typically issues in weeks where a traditional application can take months. What the compact has not changed is the verification burden on you: a vendor claiming 50-state coverage should be able to show you, per reader assigned to your account, an active license in your state, and you can confirm each one against the state board's public lookup.

Hospital credentialing and privileging. A hospital must credential and privilege every physician who interprets its studies, including remote ones. Doing that individually for a 40-reader panel would sink the project, which is why CMS allows credentialing by proxy: under 42 CFR 482.22, your governing body may rely on the distant-site telemedicine entity's credentialing decisions through a written agreement that meets the regulation's conditions. Most hospital teleradiology contracts are built on it. The mechanics, and the residual work your medical staff office still owns, are in credentialing by proxy for teleradiology. Imaging centers without medical staff bylaws skip privileging, which is one reason outpatient launches run faster than hospital launches.

Plan two to six weeks from countersignature to first clinical read for an imaging center, longer for a hospital where committee calendars set the pace. AstraRad commits to a first signed report within 10 business days of countersignature where licensure is already in place.

How is quality controlled when the radiologist is remote?

Distance changes nothing about the standard of care, and the correct answer to "how do you assure quality" is always a described process with numbers attached. Four mechanisms make up a credible program.

Peer review. A defined fraction of signed reports is re-read by a second radiologist and scored, on a program modeled on ACR peer review standards. Ask for the sampling rate and whether the second read is blind; a reviewer who can see the first report tends to agree with it. AstraRad double-reads 1 in 20 reports, blind, by a second subspecialist.

A published discrepancy rate. Discrepancies happen in every radiology practice on earth; the question is whether the vendor counts them and shows you the count. Major discrepancy rates in well-run practices sit under 1 percent of signed reports; AstraRad's runs under 0.3 percent, with each case reviewed at a monthly discrepancy meeting and closed with the reader. Treat a claimed rate of zero as a measurement failure. The scoring methodology behind numbers like these is explained in RADPEER scoring.

Critical results communication. Documented phone communication of suspected critical findings, with time and recipient on the report. This is a patient-safety mechanism and, for hospitals, a survey item.

A quality report you receive. Turnaround by tier, SLA compliance, the discrepancy log, and what changed because of it, monthly, to you. A program whose outputs stay inside the vendor is a program you cannot audit.

One honest caveat: peer review percentages and discrepancy rates are self-reported across this entire industry, ours included. The verification move is to ask for the methodology in writing and check that the numbers are consistent everywhere the vendor publishes them.

What does teleradiology cost in 2026?

Almost every US teleradiology contract is priced per study, and almost nobody publishes rates. The main exception is NDX Imaging, which lists starting prices on a public prices page: X-ray from $12, ultrasound from $28, CT from $40, MRI from $60, and PET-CT from $99, checked September 3, 2026. Assembled from the few public rate sheets, market surveys, and buyer-reported contract terms in circulation, industry-typical bands run roughly $10 to $15 per X-ray, $30 to $50 per single-region CT, and $45 to $75 per single-part MRI, with STAT priority adding 20 to 50 percent. All of those are market figures from third parties; none is an AstraRad price.

The base rate is the least of it. What moves the invoice is structure: monthly minimums that bill for volume you did not send, platform and PACS access fees that add fixed cost, surge premiums on spikes, and separate charges for subspecialty routing or critical results calls. A facility committed to 1,200 studies that sends 900 pays for 1,200, which puts the real per-study price a third above the negotiated rate. The full modality-by-modality breakdown is in teleradiology cost per read, and the structural comparison in per-report vs subscription pricing.

AstraRad's commercial model is one price per signed report, set by study type on a written rate card, with a fixed printed multiplier for STAT and urgent priority. No minimums, no subscriptions, no platform fees, no surge premium. The pricing page sets out the model, and the rate card itself arrives within one business day of a request.

Where teleradiology is the wrong answer

An honest guide names the cases where the model loses.

Interventional and procedural work. A remote radiologist cannot do a biopsy, a drainage, or a fluoroscopy-guided procedure. Facilities with procedural volume keep on-site radiologists regardless of how reads are covered, and the realistic design is hybrid.

High-touch clinical consultation. A radiologist down the hall who joins tumor board, walks to the ED, and knows the referring physicians by name delivers something a worklist does not. Teleradiology practices manage this with phone access to the interpreting radiologist, and the good ones make that access fast, but a call is still a call.

Steady volume with a narrow mix. If your volume is predictable and concentrated in a few study types, an employed or contracted local radiologist is usually cheaper per read than outsourcing the same volume. Teleradiology's economics win on variability, nights, and subspecialty breadth; they do not automatically win on a flat, full weekday worklist.

Sites that cannot feed the read. Interpretation quality depends on clinical history and priors arriving with the study. A facility whose order data is chronically thin will get thinner reports from any remote reader, and fixing the order workflow has to come first.

How do you evaluate a teleradiology vendor?

Run the evaluation as procurement, with every claim reduced to a document or a number. The short checklist:

1. Read tier per service line, prelim or final, in the contract language itself.
2. Turnaround definitions: clock start, clock stop, compliance percentage, and the median, all in writing.
3. Licensure evidence per reader for your state, and the credentialing-by-proxy agreement if you are a hospital.
4. Subspecialty routing rules: what actually determines which radiologist opens your study.
5. The quality program: sampling rate, blindness, discrepancy rate, and who receives the monthly report.
6. The complete price structure: base rates for your full study mix, the priority multiplier, and every recurring fee, on one page.
7. References from facilities your size, on your study mix, called by you.

The longer version of this framework, including the questions that expose a padded claim, is in how to choose a teleradiology company, with a companion piece on verifying vendor claims.

The evaluation ends with paper. Ask each finalist for a written rate card covering your entire study mix and a copy of their SLA definitions. AstraRad sends both on request: the complete per-report rate card lands in your inbox within one business day, through a request on the contact page, and the first signed report follows within 10 business days of countersignature.