Cost of Incomplete Studies Reaching the Reader

The bill arrives after the study already left Pending

A shoulder MRI clears Pending at 10:40. At 1:15 a radiologist opens it and finds that the comparison study named in the order was never retrieved from the outside facility. The read stops there. A request goes out, the prior arrives the next morning, and the study is read at 9:20, roughly a day after the images were available.

Nothing in that sequence was a diagnostic question, and nothing in it gets logged as a cost. The reader lost the thread of a list already in progress. A PACS admin absorbed an interruption that arrived as a chat message instead of as queued work. Someone called the outside facility. The report landed a day later than the exam time suggests it should have. The study itself looks clean in every report that counts it.

Three costs that never share a column

Ops teams usually want one number that proves the problem. Incomplete studies don't produce one. The cost lands in three places that rarely sit on the same report:

• Turnaround drag. Exam-complete to report-signed stretches while someone hunts a prior, fixes laterality, or waits on a callback. A blended TAT number mixes that waiting with real reading load, which is why turnaround work keeps pointing upstream of the dictation desk.
• Interrupt and callback rework. A reader stops, opens a chat or a ticket, and restarts later. The same study may return to Pending and then re-enter the reading path after remediation. That second pass is real work even when the final report looks clean.
• Downstream cleanup. Thin reasons for exam, wrong laterality, and mismatched descriptors resurface weeks later in denials, addenda, and "can you clarify" threads. Fixing the record before reading costs less than arguing about it after.

None of that requires a national ROI percentage. It requires noticing that the same gaps that keep a study from being read-ready still create work when the gate is soft. Those gaps are ordinary rather than exotic. When one hospital graded 762 of its inpatient imaging referrals against a published completeness scale, only 8% landed in the top two grades, and the diagnostic question was missing from 63% of requests. That is a single center's inpatient mix, not a benchmark for your practice, which is rather the point.

Why the average TAT number hides it

A practice can post an acceptable average turnaround while a thin slice of studies burns disproportionate time. Most studies move. A minority stall on priors, paperwork, or classification, and those stalls pull readers into administrative loops and PACS admins into after-the-fact remediation.

Routing doesn't change that shape. Reordering a queue of unfinished records only decides which unfinished record gets expensive first. That's the split in Pending QC vs worklist smart routing: readiness decides whether a study belongs in the reading path, routing decides order among the studies that already do.

Measure your own baseline before anyone sells you a savings figure

If leadership wants a number, build it from your queue rather than from a pitch deck. One week of aging and returned studies is usually enough to make the pattern visible:

1. Count studies that reached a reader and were returned, parked, or interrupted for missing priors, blank or unusable reasons for exam, laterality or body-part conflicts, open holds, or missing paperwork.
2. For each, note where the clock was: still in Pending, already on a worklist, or mid-dictation.
3. Tag the next owner of the fix, whether that is a PACS admin, a tech, the referring office, or the outside-prior desk, and note whether that owner was named before the reader noticed.
4. Split turnaround into exam-complete to read-ready and read-ready to report start when you can. Incomplete studies inflate the first interval, and treating that as a reader-capacity problem sends the fix to the wrong team.

A perfect cost model isn't the goal. A local rate, a short list of failure modes, and a named owner for the gate are enough to decide whether the expensive path is rare enough to live with or frequent enough to staff.

The cheaper checkpoint is still Pending

The operational answer is unglamorous. Hold unfinished studies in Pending with issue context and an owned next step, and release only what meets local readiness criteria. That work belongs with PACS admins, because the tools that clear it (tags, statuses, tasks, source corrections) sit in ops access rather than on the reading roster.

A soft expectation that "readers will catch it" isn't a gate. It is a staffing decision nobody wrote down, funded out of the reading desk's capacity. A hard gate keeps an incomplete study in Pending until remediation closes, and only then makes it eligible to read.

Where Corena fits

Corena is operational QC for studies waiting in a PACS Pending queue. It helps admins see whether a study is complete, correctly described, and ready for the reading workflow, with evidence-backed findings and proposed actions a human reviews before anything is written back. It does not interpret images, generate clinical reports, or assign the radiologist.

If incomplete studies are already reaching your readers, the most useful first conversation is usually a walkthrough of your current Pending criteria against real returned cases. Request a demo to map that gate onto your queue without turning the reading desk into the QC desk.