The hidden cost of the prelim-plus-overread model

Seven forty on a weekday morning, and the first case on your day radiologist's worklist is a chest CT from eleven hours ago. It's already been read once. Now it has to be opened again, checked against the preliminary text, and signed by a physician who wasn't in the building overnight, and behind it sits the rest of the night's studies waiting for the same second pass. That queue is the hidden cost of the prelim-plus-overread model: you pay for two radiologist touches on every overnight study and bill one report, because the professional component belongs to the physician who signs the final. The Medicare Claims Processing Manual pays the professional component of a radiology service to the physician who performs the interpretation, and states that the interpretation of a diagnostic procedure includes a written report. Vendor pricing pages quote the per-study fee. This page prices the workflow around it.

A preliminary read commits your own radiologists to finish it

Buying overnight preliminary coverage means buying half an interpretation, with your own day roster committed to supplying the other half. A preliminary read moves care forward during the night and is superseded in the morning by a signed final report from a different physician. Medicare conditions payment for an interpretation on a written report prepared for inclusion in the patient's medical record, and in this model that report is the one your day radiologist signs, not the one your vendor sent overnight. If the vocabulary needs untangling at your facility, our explainer on STAT, preliminary, and final reads separates the three cleanly.

That second half shows up in four places, and none of them appear in the contract. Every overnight study returns to a day radiologist who opens it, reviews the images against the preliminary text, and signs. Some fraction of those overreads changes the finding, which triggers notification, documentation, and occasionally a clinical intervention. Every changed finding becomes an amended report, a records event that touches HIM, the ordering clinician, and sometimes risk management. And the whole queue lands at the start of the day shift, where it competes with the outpatient schedule and the morning ED surge. All four costs are paid in radiologist salary, which is exactly why they feel free.

Costing the prelim-plus-overread model to one number

The prelim-plus-overread model can only be compared with final-read pricing on one basis: fully loaded cost per billable signed report. Everything else is a unit price argument, and unit prices are where this model looks best.

Here's why this comparison almost never gets made, and why it's worth an afternoon of your time. The inputs live in three different systems. The prelim fee arrives as a vendor invoice, the overread hours dissolve into salaried time, and the discrepancy work gets logged, when it gets logged at all, as a quality activity. No dashboard adds the three together, so nobody ever sees a single number. Building one takes three measurements. Time at least 20 overreads with a stopwatch, splitting plain film from cross-sectional; the mental model is a quick glance, while the task includes loading priors, reading the preliminary text, and re-dictating in the house template. Then pull your last 90 days of overreads and count how many changed the impression in a way that altered management; that measured figure is your operating discrepancy rate, and it beats any literature value for this purpose. Last, compute loaded cost per radiologist minute: total compensation plus benefits plus malpractice, divided by minutes spent reading, which run well below scheduled hours. Now the model prices itself: prelim fee, plus overread minutes at your loaded rate, plus reconciliation time at your measured discrepancy rate, plus amendment administration. Set that sum beside a per-report final-read price and the spreadsheet advantage either survives or it doesn't.

For unit price context on the final-read side, our breakdown of teleradiology cost per read by modality covers industry-typical ranges. This page stays on the total cost of the model.

The fourth cost resists a formula and often outweighs the rest. When the overread queue eats the first 90 minutes of the day shift, those minutes come straight out of the outpatient backlog, and you can value them at the marginal revenue of the displaced studies or at the cost of the locum hired to absorb them. Either valuation belongs in the comparison.

Discrepancies surface after the patient has gone home

A two-step interpretation can disagree with itself, and the disagreement arrives hours after care has moved on. An emergency department sending 100 overnight studies will see a handful of clinically meaningful changes at the morning overread, whatever your exact discrepancy rate turns out to be, and the patients involved are frequently already discharged. A meta-analysis of 616,918 on-call resident preliminary reports put pooled major discrepancy rates at 0.38 percent for ultrasound, 0.64 percent for radiographs, 1.35 percent for CT and 1.86 percent for MRI, so the count you should expect depends heavily on your overnight modality mix.

Each change starts a reconciliation. The day radiologist confirms it, spending time beyond the routine overread. Someone locates and phones the treating clinician, often across a shift change and often on the second or third attempt. The closed-loop communication gets documented under your critical results policy, which the ACR Practice Parameter for Communication of Diagnostic Imaging Findings asks be done as soon as possible for any significant variation between the preliminary and final interpretations, with documentation of that communication incorporated into the final report. An amended report then moves through dictation and HIM, and the ED or referring practice arranges follow-up, sometimes a recall visit or repeat imaging. All of it is unavoidable clinical work once the discrepancy exists, and it exists because the interpretation was split into two passes separated by hours.

There's also a dimension the cost model can't hold but a risk manager will. Both the preliminary and the final live in the record, and when they disagree, the interval between them is the documented period in which a patient was managed on information the record later contradicted. That interval is discoverable, and its length is a choice you make when you decide how the first interpretation gets signed.

One more effect deserves plain words. A physician who knows a colleague will re-read every case in the morning writes a looser first report, and radiologists say so openly about preliminary contracts. The incentive sits in the workflow itself, and professionalism alone never fully cancels it.

Six hours a night: the arithmetic at 100 overnight studies

At 100 overnight studies, on assumptions you should replace with your own measurements, none of them an AstraRad figure, the prelim-plus-overread model absorbs roughly six hours of day-radiologist time per night. Assume three minutes of overread per study, and the queue consumes 300 radiologist minutes before anyone opens a new case. Assume two or three reconciliations per hundred overreads, a working assumption until you've measured your own rate, at 25 minutes each, and that adds 50 to 75 minutes more. Call it six hours a night, recurring, before amendment administration and before the outpatient studies displaced from the morning schedule. Over a month, that's roughly a full reading week absorbed by work the contract never priced.

The prelim fee has to beat that gap to win. Sometimes it does. Often nobody checks, because the overread hours are salary and the prelim fees are an invoice, and the two never meet in the same report. If you're weighing the broader staffing question, our comparison of in-house versus outsourced radiology runs the same arithmetic across the whole coverage model.

When does the prelim-plus-overread model still make sense?

Preliminary workflows earn their keep in teaching programs, in groups that want their own signatures on every report, and in hyperacute cases where a phone impression precedes a fast final. In an academic program, resident preliminaries finalized by attendings are how radiologists learn to read; the overread cycle is the educational product, not overhead. Some practices deliberately keep every signature in-house for continuity, credentialing, or referral relationships, and for them the second touch is the known price of that continuity. And a verbal impression given within minutes of a hyperacute study, followed shortly by the signed final, is a sequencing choice inside a single interpretation; nothing sits provisional overnight.

If one of those describes your facility, a final-read service like AstraRad is the wrong fit, and preliminary coverage is the honest choice. Outside them, the model is usually inherited: it was the only overnight option when the contract was written, and it has renewed on autopilot since.

Signed finals dissolve the morning queue

When the overnight study arrives with a signed final report attached, there's nothing left to finalize, and the standing morning queue stops existing. The effect shows up at the far end of the workflow as well. A single-centre study of overnight attending final reporting recorded imaging-related emergency department recalls of 54 and 61 in two prior years, against 7 after the switch. That is a decrease of at least 90 percent of expected recalls. Day radiologists open the current day's work at the start of the shift. Overnight discrepancy reconciliation stops being a standing category, amendments fall to the residual rate any practice carries, and one physician is accountable per study, which simplifies peer review, discrepancy tracking, and credentialing files.

AstraRad delivers signed final reports at every priority tier, 24/7/365. We route each overnight study to a fellowship-trained subspecialist licensed in the state where your patients are located, so a 2 a.m. head CT is interpreted by a neuroradiologist and a 3 a.m. wrist series by an MSK reader, on the same tiers as daytime work. The contract commits to STAT studies inside 4 hours, with 2-hour and 1-hour STAT on request at a higher multiplier, and routine inside 24 to 48 hours, each measured from last-image arrival to radiologist signature, and every one arrives signed. Quality runs as a sampling program outside the care timeline: 1 in 20 signed reports gets a blinded independent double read, and the methodology is published on our SLA and quality page. Studies travel by DICOM from your PACS or by portal upload, and results return over HL7 or FHIR on a HIPAA aligned platform covered by a signed BAA. Pricing is per signed report with no minimums, which matters for this exercise: the per-report price is the entire cost, with no third column to estimate. If you're replacing an overnight preliminary contract specifically, our nighthawk radiology coverage page walks through how the transition runs.

Run your own numbers before you talk to any vendor, including us. Time twenty overreads, pull ninety days of discrepancies, and price a radiologist minute honestly, then set the loaded total beside a single per-report price. That total is what the prelim-plus-overread model costs you. Ask us for a rate card when you have the figure in hand. The prelim fee is the half of the bill that gets printed. The morning queue is the half that never does.