AstraRad teleradiology SLA & QA: commitments in writing

Buyers treat a STAT commitment as a number they can line up across vendors. It compares only once both vendors name the two events their clock runs between. AstraRad names them at the top of the page: the teleradiology SLA clock starts when the last image of your study arrives on AstraRad systems, and it stops when the interpreting radiologist signs the final report. Under that definition, STAT is committed at under 4 hours, with a 2-hour or 1-hour STAT available on request at a higher multiplier, Routine at 24 to 48 hours, and stroke protocol studies inside the STAT tier are committed at under 30 minutes on a dedicated pathway, with compliance audited monthly against system timestamps and reported to you study by study. What follows is the measurement rule, the QA program underneath it, and what happens on a miss.

Each study carries a tier from the moment your staff submits it

The tier is chosen at your facility when the study is sent, and it is attached to the study before the clock starts: STAT for studies that change management the same day, at the standard 4-hour commitment or at a faster 2-hour or 1-hour speed agreed in your contract; Routine for outpatient imaging and backlog work. A stroke activation is flagged as stroke protocol at order, and that flag carries its own commitment: the signed report is due under 30 minutes from last-image arrival, the study is routed ahead of every other STAT read, and the reading neuroradiologist phones a large vessel occlusion or hemorrhage to your designated recipient before sign-off. Escalation stays an operational decision on your side, and the multiplier for each STAT speed is priced in writing on your rate card before you send a single study, so moving a case to STAT never becomes a phone negotiation at 2 a.m.

The compliance percentage in your monthly report is one blended figure covering every tier, stroke protocol studies included, and every one of your reads in the period. Stroke protocol studies are also broken out on their own line, measured against the 30-minute commitment rather than the STAT window. Blending is the part worth pressing any vendor on, including this one. A service that publishes compliance for its STAT tier alone can show a spectacular percentage while routine outpatient work drifts for days behind it, and a service that publishes its strongest month can show a percentage no full year would reproduce. Both practices are legal and both survive most RFP reviews, because the question that would catch them is rarely asked in writing. AstraRad reports a single number covering every study in the window, and an out-of-tier study is never removed from the denominator. Causes outside AstraRad's control do not retire a miss either: a network outage at your facility, a transfer that arrived corrupted and had to be resent, a study that landed without the priors its reader needed. Those causes go in the root cause column of your monthly quality report, and the study still counts against the percentage. Holding the denominator still is the only thing that keeps the figure comparable to itself from one month to the next.

Every miss reaches you the same way it reaches AstraRad's operations team: as a line in the monthly quality report, with study identifier, tier, actual turnaround, and cause. A cause that repeats gets a written corrective action. For the clinical picture behind fast-turnaround work, see STAT radiology reads and STAT CT reads.

How is turnaround time measured?

The clock starts when the last image of the study arrives on AstraRad systems, by DICOM push from your PACS or by portal upload, and stops the moment the interpreting radiologist signs the final report. Everything between those two events counts against the commitment: technical QC, triage, subspecialty assignment, the reading itself, and dictation. There is no separate internal clock that starts later.

Three consequences follow, and each is a question you can put to any vendor:

• The endpoint is a signature on a final report. A preliminary read reaches a different endpoint and travels a shorter clock, which is why turnaround figures built on preliminaries look faster than they perform. STAT vs preliminary vs final reads walks through the distinction in clinical terms.
• A median is a summary, not a guarantee; the audit lives in the distribution. A median resists distortion by a handful of very fast or very slow cases, which makes it honest as a summary and useless as a commitment. Your monthly report breaks turnaround out by tier so you can read the tail yourself.
• Nothing pauses the clock. Coverage is continuous through nights, weekends, and holidays, so a STAT study arriving at 3 a.m. on a Sunday carries exactly the commitment it would carry at noon on a Tuesday.

Compliance is audited monthly against system timestamps. Submission, QC, assignment, signature, and delivery each write an immutable event with an actor and a time, the same logging substrate described in the compliance program documentation. Timestamps are what make an SLA auditable; a percentage assembled by hand is a claim about someone's spreadsheet.

That audit is internal, and the page says so on purpose. AstraRad runs it monthly against its own immutable timestamps, and no third party attests to the result. What stands in place of an attestation is arithmetic you can repeat: every study in your monthly quality report carries its tier and its actual turnaround, so the reported percentage is re-derivable from the log you already receive. Ask any vendor quoting a compliance figure whether their number survives that test.

One in twenty signed reports is read a second time, blind

One signed report in twenty is pulled for a second interpretation by another subspecialist, under a standing QA program modeled on the peer review expectations in the ACR White Paper on Teleradiology Practice. The second reader opens the study with the original report withheld, works from the images and clinical history alone, and dictates independently. Agreement and disagreement are scored case by case.

Classification by clinical consequence. Disagreements are graded by what they would change in the patient's care. Major discrepancies are the ones that could alter management.

Closure with the reader who signed. Each major discrepancy goes to the monthly discrepancy meeting, is reviewed with the radiologist who signed the original, and is closed with a documented outcome. Your quality report carries the discrepancy log alongside turnaround by tier and compliance.

Subspecialty routing as the first control. Every study reaches a radiologist credentialed for that modality and body part: a child's abdominal CT goes to a pediatric radiologist, a coronary CTA to a cardiac reader. That matching happens before any peer review sample is drawn, and it is the first control the double-read program sits behind.

Every commitment quoted elsewhere on this site is defined in this table

These are AstraRad's canonical commitments. Every figure quoted anywhere else on the site, including the facts page, resolves to a row in this table, so every number has exactly one definition. AstraRad does not yet publish historical performance figures such as volume, compliance or discrepancy rates; it publishes what the contract commits to, and reports actual results to each client monthly.

Six questions to put in your RFP

Require written answers to these six questions, from AstraRad and from every other vendor; take none of it on faith.

1. Define the clock. From which event to which event is turnaround measured, and where is that definition written down?
2. Put a compliance percentage in the contract. A tier table states an intention; ask for the measured figure, the window it covers, and the audit method.
3. Ask about misses. Where do out-of-tier studies appear, who reviews them, how long do they stay in the denominator, and are they ever removed?
4. Ask about peer review design. What fraction of signed reports is double-read, does the second reader work blind, and how is the major discrepancy rate measured and reported?
5. Ask who reads your studies. Board certification, fellowship training, subspecialty match to modality and body part, and licensure in the state where your patients are located.
6. Ask how quality reaches you. A monthly report carrying turnaround by tier, compliance, and the discrepancy log should arrive without anyone requesting it.

The longer version of this checklist, with scoring guidance, sits in the guide on how to choose a teleradiology company.

Sign the teleradiology SLA, then audit it monthly

A teleradiology SLA is only as strong as the log that proves it. The tier commitments are a promise and the compliance percentage in your monthly report is a measured outcome over a stated window; the timestamp log is what lets your compliance officer check the second against the first. The monthly quality report is how all three arrive on your desk without a request.

One honest limit: if your sites need onsite procedural coverage, contrast supervision, or preliminary reads ahead of your own morning overread, a remote final-read service is the wrong instrument and AstraRad is the wrong fit. Pricing is per report with no minimums, no subscriptions, and no platform fees. Tier multipliers appear on the rate card so escalation carries a known cost, and our pricing model is published in full. To see the tier multipliers in writing before any commitment, ask for a rate card.