Skip to main content
lab-notes.aiPublic-service AI playbooks
Opportunity 03/17HealthHealth and care

Diagnostic Imaging Support

Explore what it would take to support a reporting clinician deciding which scan to read next, without creating an irresponsible stand-in dataset.

Published sources
2
Starter data
None — by design
Domain build partner
build-diagnostic-imaging-support
Last reviewed

From the draft strategy

Opportunity

The draft strategy names AI-based analysis of medical images as a potential public-service application: support for spotting fractures, cancers, and strokes in scans, with no clinical dataset or safety work attached to the suggestion.

Where the draft says it

Table 2 — potential public-service applications

Read the draft strategy source (opens in a new tab)

Source register

Research already done

Real published source

We reviewed these published sources so you do not have to start from zero. Each source shows what it covers, how it can be accessed, and why it matters.

2 sources

  1. Department of Health

    Open

    Diagnostic waiting times

    What it covers
    How many people are waiting for a diagnostic test in Northern Ireland, how long they have waited, and how long reporting turnaround takes at each health and social care trust.
    Why it matters here
    It is the closest published figure to the pressure this example is meant to relieve, and it counts people and weeks rather than describing a single image.

    Open the source (opens in a new tab)

  2. Department of Health

    Open

    Hospital waiting times statistics

    What it covers
    Quarterly outpatient, inpatient, diagnostic, and cancer waiting figures broken down by trust and by length of wait.
    Why it matters here
    It shows what is publicly available around imaging services and, by omission, that no image and no report ever leaves the clinical record.

    Open the source (opens in a new tab)

Starter data

Starter dataset

No synthetic dataset — by design

Why we stopped

A scan cannot be honestly stood in for by a file of invented numbers, and every tabular stand-in we sketched for this task drifted towards describing individual patients rather than groups.

What responsible work needs instead

Anyone taking this further needs a partner radiology service and access to a real imaging archive under formal clinical-research governance, with a named sponsor, ethics approval, and reporting clinicians involved from the start.

Clone and build

Domain build partner

Build with a domain-aware coding agent

Clone the repository, then ask your coding agent to load the skill that ships with it. The skill brings the opportunity, sources, starter data, known unknowns and constraints into the conversation before anything is proposed. It follows the open Agent Skills standard, so it works in Claude Code, Codex and any other agent that reads skills.

01 / Clone

git clone https://github.com/Hypership-Software/lab-notes.ai.git

02 / Invoke

Claude Code
/build-diagnostic-imaging-support
Codex
$build-diagnostic-imaging-support

Conditions, not a checklist

Before you build

These are not footnotes. They are conditions that any responsible prototype would need to address.

  • Image errors change patient care

    Missing a finding, or marking one that is not there, changes what happens to a patient, so nothing on this page should be read as a claim that image analysis is safe to use in care.

  • Published collections may not transfer

    Openly published imaging collections come from other populations, other scanners, and other working practices, so how a system behaved on them says little about how it would behave here.

  • Visual support can misdirect attention

    A confident outline drawn on a scan can pull a clinician's eye to the wrong place, which makes the support itself a source of harm rather than only a missed benefit.