24/7 radiology coverage: the night staffing math
About 1.6 to 2 radiologists per continuous seat: the night staffing math behind 24/7 radiology coverage, plus five questions that expose a thin bench.
"How many radiologists do we need to keep nights covered?" Buyers ask it in those exact words, usually after a quarter of bad overnight turnaround. Here's the flat answer: roughly 1.6 to 2 full-time radiologists per continuous seat, as industry staffing arithmetic goes, once nights off, vacation, CME, and sick leave are priced in. That staffs one awake generalist. Subspecialty match, licensure in the state where your patients are located, and credentialing at your facility each shrink the eligible pool further, which is why the night staffing math behind 24/7 radiology coverage is a panel problem long before it's a technology problem.
This page walks through that math. For the service itself, see nighthawk radiology coverage. If you're still pricing the decision, start with what overnight radiology coverage costs.
Night staffing math is queuing math
Overnight turnaround degrades because emergency studies arrive in unpredictable clusters during exactly the hours when the fewest radiologists sit at a workstation. Daytime volume is heavier but tamer, and a deep bench absorbs it.
Watch one reader's night when a trauma comes in. The activation produces a pan-scan: chest, abdomen and pelvis, plus a head CT and a cervical spine. Read carefully, that's the better part of an hour of one radiologist's full attention. The scanner downstairs keeps running through that hour, so the appendicitis-question CT lands in the queue, then the elderly fall with a head strike, then a chest CT from the floor. Each of them waits, because a single reader signs one study at a time while arrivals ignore the schedule. By the time the trauma is signed, the queue holds ninety minutes of work, and the last study in line blows through any one-hour commitment even though every individual read was fast. A published median never shows this night, because medians are built mostly out of quiet ones. Your emergency physicians remember the loud ones. At 3 a.m., a queue of four studies is a hallway of people waiting on decisions.
The loud night costs more than minutes. Across 18,488 overnight on-call interpretations at a level 1 trauma center, Emergency Radiology found that higher per-shift workload was associated with worse interpretive accuracy among the less experienced readers on the roster, measured in relative value units per shift. That study followed residents, so read it as evidence about workload rather than about any particular bench, and the direction is the one every scheduler expects.
The fix is arithmetic too. A STAT commitment survives the loud night only when spare reading capacity is already sitting idle as the cluster arrives.
Who is allowed to read our overnight studies?
An overnight study can only go to a radiologist who's awake, licensed in the state where your patients are located, credentialed at your facility, and fellowship-trained in what the study shows. Each gate alone is manageable. Stacked, they cut a large panel down to a handful of eligible readers for a given study at a given hour, and no brochure prints that number.
Licensure is the gate most often discovered late. Under 42 CFR 482.22(a)(4)(iii), a distant-site physician reading for a Medicare hospital's patients has to hold a license issued or recognized by the state where that hospital sits, whatever state the reading workstation is in. AstraRad assigns radiologists licensed in the state where your patients are located, and the licensing page explains how that's verified before go-live.
Subspecialty match is the gate most often quietly dropped after dark. Overnight is when a vendor feels the strongest pull to route everything to whoever happens to be awake, and the harder studies are where discrepancies concentrate. A 2025 Academic Radiology meta-analysis pooling 616,918 on-call preliminary reports put the major discrepancy rate at 0.38 percent for ultrasound and 0.64 percent for radiographs, against 1.35 percent for CT and 1.86 percent for MRI. That cohort is residents rather than attendings, but the modality gradient is the point: the cross-sectional studies that arrive at 3 a.m. are the ones least forgiving of a reader outside their training. AstraRad holds the match through the night across neuro, MSK, body, cardiac, breast, pediatric, chest, ER, oncologic, and nuclear medicine: a pediatric MR study goes to a pediatric radiologist at 3 a.m., the same way it does at 3 p.m. The panel behind that is described on our radiologists.
Headroom predicts the bad night
Headroom is the volume a vendor could absorb this month without hiring anyone, and it predicts overnight performance better than any published average. Headcount flatters every vendor, because it counts readers who are asleep, unlicensed in your state, or never credentialed at your facility. Headroom counts what's usable.
A vendor running near full is fragile. One resignation, one sick call, one regional trauma cluster, and your STAT tier slides from under an hour to several hours with the contract unchanged.
AstraRad's published tiers are STAT under 4 hours, with 2-hour and 1-hour STAT on request at a higher multiplier, routine 24 to 48 hours, and stroke protocol under 30 minutes, each measured from last-image arrival to radiologist signature. Measurement definitions for the tiers live on the SLA page.
Headroom is also what coverage borrows on the days the math gets tested: a scanner coming back online and dumping deferred studies on top of normal volume, a four-day holiday weekend read by a reduced local bench, a resignation the remaining panel has to absorb. Weekend and holiday coverage and overflow radiology reads are both, underneath, headroom products.
Put one line in your RFP verbatim: state your current monthly headroom in studies, and state your trailing twelve month compliance rate against published turnaround tiers. A vendor that answers the second and dodges the first has told you something.
An offshore preliminary is a study read twice
The offshore preliminary model means every overnight study gets read twice: a wet read overnight, then a billable overread at your facility the next morning. Medicare does not pay for services furnished outside the United States under 42 CFR 411.9. The Medicare Benefit Policy Manual, Chapter 16 applies that to imaging in as many words: if a radiologist practicing abroad analyzes a study performed on a beneficiary in the United States, neither the radiologist nor the US facility is paid for the work the radiologist did. So the offshore preliminary can never become the final. Your own medical staff carries the overread as morning work, the record shows two opinions, and responsibility for a miss splits between them.
AstraRad's radiologists all work from inside the United States, scheduled in shifts so a fellowship-trained subspecialist is signing at every hour of the night. What lands in your PACS overnight is a final signed report with one signature on it, and nobody at your facility re-reads it in the morning. The difference between read types is its own topic: see STAT, preliminary and final reads, or what teleradiology is for the general model.
Five questions that expose a thin night bench
Five questions, asked before signature, separate a deep night bench from a thin one. Honest answers are specific. Evasive answers all share a shape.
| Question | An honest answer sounds like | An evasive answer sounds like |
|---|---|---|
| What is your current monthly headroom in studies? | A number, in writing | "We scale with demand" |
| How is your panel scheduled across the night hours? | A shift distribution | "We have radiologists nationwide" |
| Is overnight turnaround measured separately from the daily average? | Yes, with the overnight median stated | A single blended average |
| Do overnight studies get subspecialty matched? | Yes, with the subspecialties named | "All our radiologists are highly qualified" |
| What is your compliance rate against published tiers, trailing twelve months? | A percentage with a measurement definition | A best case turnaround with no denominator |
AstraRad publishes what it can answer today: subspecialty matching held through the night, and contractual turnaround tiers with their measurement method on the SLA page. A headroom figure and a compliance rate are not yet published. Quality sits on top of the turnaround tiers: our protocol double-reads one in twenty signed reports independently, with each major discrepancy reviewed at the monthly discrepancy meeting and closed with the radiologist who signed.
Turnaround language belongs in the contract. Writing turnaround time into a teleradiology contract covers how to make these five answers enforceable.
Run the night staffing math before you sign
Three numbers in writing tell you most of what you need: monthly headroom, the night shift distribution, and the trailing compliance rate against published tiers. Any vendor holding a deep bench can produce all three in a day, because they're operating figures, the same ones its own scheduler works from.
One honest caveat. If your in-house group already covers nights deeply and hits its own turnaround targets, stop reading here: external night coverage would add cost without adding capability. This math matters when the bench is thin, or about to be.
The median describes your quiet nights. Headroom decides your loud ones. To check the night staffing math against your own overnight volume, request a volume quote with last month's study counts by hour; per-report pricing with no minimums means running the comparison costs you nothing.
Frequently asked questions
How many radiologists does 24/7 coverage require?
Roughly 1.6 to 2.0 FTE per continuous seat, once nights off, vacation, CME, and sick leave are priced in, and that staffs one awake generalist. Real 24/7 coverage also has to hold subspecialty match, state licensure, and facility credentialing at the same hour, which multiplies the seats. AstraRad covers those hours by scheduling board-certified, fellowship-trained subspecialists in shifts, so the radiologist who signs your 3 a.m. study is qualified for exactly that study.
Why are overnight radiology reads slower than daytime reads?
Arrival variance is highest and available readers are fewest during the same hours. Emergency imaging clusters unpredictably overnight while most vendors thin their bench after dark, so a single trauma or a sick call pushes every study behind it into a queue. The vendors that hold a fast overnight median are the ones carrying spare reading capacity at the moment the cluster arrives.
What is capacity headroom and why should a buyer ask about it?
Headroom is the volume a vendor could absorb today without changing anything: extra studies per month the existing panel can read within the same turnaround tiers. A vendor running near full is one sick call or one trauma cluster away from a four hour STAT, no matter how good its published average looks. Every vendor you evaluate should be asked to state its headroom figure in writing.
Do overnight reads get a subspecialist or a generalist?
That depends entirely on the vendor, and it belongs in the contract. Many overnight models route everything to whichever generalist is awake, which is the configuration where discrepancies on complex studies concentrate. AstraRad matches overnight studies to the relevant fellowship-trained subspecialist across neuro, MSK, body, cardiac, breast, pediatric, chest, ER, oncologic, and nuclear medicine, at the same turnaround tiers used during the day.
How fast can a vendor absorb a sudden volume spike?
Ask for two numbers: monthly headroom and a documented backlog clearance timeline. AstraRad's contract holds the published turnaround tiers of STAT under 4 hours, with 2-hour and 1-hour STAT on request at a higher multiplier, routine 24 to 48 hours, and stroke protocol under 30 minutes, each measured from last-image arrival to radiologist signature. A vendor that can't state either number is asking you to trust that the spike won't happen.
Does a bigger radiologist roster mean faster overnight turnaround?
Not by itself. Headcount is a daytime number, and it says nothing about how many readers are awake, licensed in the state where your patients are located, credentialed at your facility, and matched to the right subspecialty at 3 a.m. Ask for the night shift distribution and the headroom figure instead, then check the overnight median against a compliance rate with a stated measurement window.
Related on AstraRad
- Resources
What overnight radiology coverage costs: 3 models
Overnight and weekend radiology coverage cost by model: a nocturnist runs 1.6 to 2.0 FTE, teleradiology bills per signed report, no coverage bills the ED.
- Resources
Locum radiologist vs teleradiology: fixed vs variable cost
A locum radiologist is a fixed day rate paid whether studies arrive or not. Teleradiology bills per signed report. Compared on cost, lead time and risk.
- Resources
Radiology overflow reading services: how they work
How radiology overflow reading services work: the triggers that route spillover studies, the routing rules that keep control, and the per-read economics.
Put a radiologist's name on your next read.
Tell us your modalities and monthly volume. A complete per-report rate card, with turnaround tiers and SLA terms in writing, lands in your inbox within one business day.