Quality you can show,
not assert.
Checks run before signature, not after. The evidence is a byproduct of the work: your radiologists get audited less and you get more to show.
Every recommendation gets an answer.
- Captured with a due date as it is dictated, and put in front of the ordering provider as something they answer.
- Escalates by clinical significance, not by date, so a three-week-overdue nodule outranks a two-year-overdue shoulder film.
- It lands in the chart, not in a silo. Recommendations ride the same results feed your EMR already receives — no second interface.
What the loop is worth — measured, not promised
Brigham and Women's (Harvard) published the ledger. They hired a safety-net team — chart review, tracking, phone outreach — to chase 13,670 follow-up recommendations, and the chase generated $6.1 million in revenue from completed examinations, $980,628 of it attributable to the program: about $350,000 in net revenue for every person they hired, with early revenue inside six months (Jhala et al., JACR 2024). Here, the part they hired people for is what the software does. The recommendation is captured structured at dictation, tracked, deadlined and escalated on its own — you don't staff a program, your existing schedulers and navigators get a worked list. Your completion number will still be your own — published rates range from roughly a third to three quarters — so we don't predict it: the registry measures it, issued through recovered, valued by your finance office at your own rates. We publish no adherence or revenue figure of our own.
They hired a safety-net team to find, track and chase these. Here, that part is the product — captured at dictation, tracked, deadlined and escalated on its own. Nobody new gets hired; your existing schedulers and navigators get a worked list.
Critical results close the loop, with timestamps.
An advanced critical-results communication system, built for the ease of both the radiologist and the ordering provider. Critical findings are flagged automatically from the report text — an acute intracranial hemorrhage, a pulmonary embolism — and offered for communication before the report is signed. The call reaches the person who ordered it, they acknowledge it, and the whole exchange is recorded rather than remembered as phone tag.
Measured against difficulty, not just the clock.
Raw turnaround starts arguments because it punishes whoever took the hard case. The objection is true — so the number is adjusted against your own casemix.
Fully customizable QA metrics
Every metric on this page is a definition your group and your facility own — what is measured, which clock starts it, the target it is held to, and who sees it. Definitions are versioned and editable in the product, and a metric you need that isn't here is usually a short conversation rather than a roadmap item.
The caveat we state ourselves, before a radiologist raises it
wRVU is a proxy for complexity, not a measure of reading difficulty. A hard normal and an easy positive can carry the same wRVU. We say that in the packet rather than let it be discovered.
The reader can say “that one doesn't count.”
- A radiologist can flag any study whose clock isn't representative — a PACS or network outage, images arriving late, an interruption for higher-acuity work, waiting on comparison priors, a priority mis-assigned upstream.
- The evidence is collected for them. A flag arrives with its own timestamp trail, so the footnote is defensible rather than an assertion.
- Flagging annotates — it never deletes. The study stays in the metric with its footnote, and raw and adjusted are both always computed. Excluding flagged studies is a policy your facility co-signs, never a default we ship.
- One of the reasons is our own failure. Reporting-platform downtime is its own code in the taxonomy — we count our outages against ourselves, on the same record.
Why this decides whether the numbers survive contact with a reading day
A turnaround programme that cannot be corrected by the person who was there gets worked around, and then the department has no data at all. Giving the reader the footnote — with evidence attached and no power to erase the study — is what makes the measurement something they will defend rather than dodge.
wRVU, counted as you sign.
Not a month-end spreadsheet. The day builds on screen — per radiologist, refreshed as reports are signed.
And the rest of it
Everything else a department tends to ask for, without a project to produce it.
The question your radiologists will ask.
They will want to know whether this is a measurement tool pointed at them. The honest answer, and the design position, is that the switch is not ours and it is not yours.
Who sees what
The radiologist always sees their own numbers first, and that cannot be switched off by anyone. The group decides, in writing and at install, whether individual-level data is visible to its own leadership. The facility sees what the group publishes: aggregate and reader-blinded by default, anything further being a negotiated term of the agreement rather than a product default. We see nothing.
Agreeing the governance first is what makes the evidence actually arrive — the alternative is radiologists who quietly work around the tool, and then nobody has any data at all.
Want it to work some other way?
Configurability is the whole design, and it does not stop at the settings that happen to exist today. If your group or your facility wants a behaviour changed, a metric defined differently, a template shaped another way or something added outright — ask. It is very often possible, and you will be talking to the radiologist who builds it rather than to a roadmap.