The Fax Machine Is the Bottleneck: Turning Requisitions Into Booked, Protocolled Studies

The queue at a community imaging clinic does not begin at the scanner. It begins in a fax tray, where a handwritten requisition sits waiting for someone to decipher the clinical indication, key it into the RIS, work out whether there is a prior study worth pulling, send it for protocolling, and then phone a patient to book. The scanner is idle for none of that, and the patient’s wait includes all of it.

This is the transformation we would run for that clinic. It is presented as a concept, built on the same method we use in client engagements, so the before, the after, and the path between them are all visible.

What does the wait actually cost?

The Canadian Association of Radiologists, with Deloitte, reports that Canadian patients wait an average of 84 days for an MRI and 66 days for a CT, and estimates that chronic underinvestment in diagnostic imaging costs Canada $64 billion a year in lost GDP. In 2023, more than two million Canadians lost an average estimated $1,017 in annual wages while waiting for imaging. The CAR published its national maximum wait time access targets for MRI and CT in a policy statement in March 2026.

Almost all of that is capacity: scanners, radiologists, operating hours, and no software fixes any of them. This study is about the part that is not capacity. Between a referrer signing a requisition and a patient standing in front of a machine, there is a stretch of purely administrative time during which the scanner is fully available and the patient is waiting on a fax tray. That segment is small next to 84 days, and it is the segment a clinic actually controls.

Before and after: the requisition

Drag the handle. The before is the fax tray and four handoffs; the after is the structured requisition designed in its place.

The radiologist still assigns the protocol and a person still confirms the priors. What changes is that the requisition arrives as data instead of as an image of handwriting.

What we built

The concept above is the output of the same engagement steps we run in client work, and several are decisions a competent peer could argue with.

• We show unread fields as unread rather than filling them with a best guess. Handwritten requisitions are genuinely hard, and a system that confidently transcribes an ambiguous clinical indication is worse than one that admits it cannot. This makes the demo less impressive on purpose.
• We made prior-study attachment a confirmation rather than a match. Automatic matching above a threshold would be simpler and would eventually attach the wrong patient’s prior, which in imaging is a serious harm. Common names are not an edge case, they are Tuesday.
• We have the system check for a missing clinical indication at intake rather than at protocolling. That single check moves a whole category of delay from the radiologist’s queue to the moment the fax arrives, when the referrer’s office is still reachable.
• We kept protocol assignment as a radiologist decision against written rules, with the evidence shown. The protocol determines what the study can answer, and a wrong one costs a patient another wait cycle.
• We wire evals from day one against the clinic’s own historical requisitions, scored separately on field extraction and on whether the proposed protocol matched what the radiologist chose, because those fail differently.

How it works

The RIS and PACS stay the sources of truth. An incoming requisition is structured on arrival, with anything unreadable flagged rather than guessed and a missing clinical indication raised while the referrer is still reachable. Candidate prior studies are surfaced from PACS for a person to confirm. A radiologist confirms or changes the proposed protocol against the rules the lead radiologist wrote. Only then does booking happen.

Because requisitions, reports and images stay in Canadian infrastructure and processing runs with zero data retention, the design lines up with what PHIPA expects: a privacy impact assessment before deployment, an agreement naming the processor as an agent under section 17, and an audit log on every access to personal health information.

Why this pays back

The return is administrative hours and a shorter administrative segment of the patient’s wait. Break-even for the first slice is about a hundred and twenty requisitions a week at the clinic’s own loaded administrative cost.

The honest limit, and it is a large one: this does not shorten an 84-day MRI wait in any way a patient would notice. The wait is scanner hours and radiologist hours, and this touches neither. It removes days from the administrative segment, it stops requisitions stalling on a missing indication, and it gives booking staff their afternoons back. A clinic buying this expecting to move the number in the CAR report will be disappointed, and should hear that before signing rather than at the first quarterly review.