The signature gate
No report leaves the clinical service without a licensed radiologist’s signature. The rule is enforced in the platform’s code, not in a policy document.
One pathway carries a film from a clinic's phone to a signed report — built from parts Afghan clinics already have, spending the scarce resource, a radiologist's attention, only where it is needed.
1
What the clinic already has. No PACS, no DICOM feed.
2
The thread the clinic already uses, on the phone in their hand.
3
Orientation, crop, exposure. Identifying detail minimized at intake.
4
The sickest patient first — not the most recent upload.
5
A licensed radiologist, by name and license number.
6
A Dari or Pashto PDF carrying a verification code.
Software sits alongside steps 3 and 4 — normalizing the photograph, ordering the queue, and pre-drafting text for the radiologist to correct. It does not read, it does not clear, and it does not sign.
Most Afghan clinics have no PACS and never will. The pathway starts from what they do have: a film on a lightbox and a phone. The platform accepts a WhatsApp photograph of the film and normalizes it — orientation, crop, exposure — into something readable.
Identifying details are minimized at intake and the study is ordered by urgency, so a radiologist’s scarce attention lands on the sickest patient first, not the most recent upload.
A licensed radiologist reads the study in a purpose-built cockpit and signs the report. The signature is not a flourish: the reader’s name and license number are stamped onto the report, and nothing is delivered without them.
The signed report returns as a Dari or Pashto PDF in the same WhatsApp thread, typically about an hour after the photo. Every report carries a verification code any recipient can check — so a paper printout in a district clinic can still be trusted.
On the general pathway, software triages the queue and pre-drafts reports for the radiologist to correct. It does not read, it does not clear, and it does not sign. There is no WHO-recommended algorithm that rules out a wrist fracture or a pneumonia — those reads are human work.
The one place the evidence supports a larger role for software is tuberculosis screening, where the WHO has conditionally recommended computer-aided detection for triage in people aged 15 and over. That architecture — and its limits — is laid out in full on the TB screening program page.
No report leaves the clinical service without a licensed radiologist’s signature. The rule is enforced in the platform’s code, not in a policy document.
Every delivered report resolves at a public verification page — the Mohr-e Asil seal. A forged or altered report fails the check.
Every step of every study is written to an audit log that cannot be edited or deleted — the raw material of both accountability and research.