How Imaging Finder builds its directory and measures imaging access

By Dr Mark Bekhit, Radiologist · Last reviewed 27 July 2026

Imaging Finder is a public directory designed to help patients and referrers identify plausible diagnostic-imaging clinics to contact. It combines branch-level clinic information from public provider pages, official registers and structured verification work. This page explains how the directory is assembled and reports preliminary results from an independent national geographic-access study built from the same research database.

Important: a directory listing is a starting point, not confirmation that a clinic can perform a particular examination for a particular patient. Availability, referral requirements, Medicare eligibility, fees, preparation and appointment timing must be confirmed directly with the clinic.

Directory scope

The research snapshot used for the current access study contained 1,184 published Australian clinic listings as at 12 July 2026. Each record represents a clinic or hospital imaging department rather than a scanner. Clinic names, addresses, phone numbers, websites, provider groups and listed services can change, so the live directory is continuously maintained and may differ from the frozen research snapshot.

Imaging Finder includes private radiology practices, public and private hospital departments, specialist imaging services and independent clinics where enough public evidence is available to identify a location. Inclusion is not an endorsement, ranking or claim about clinical quality. The directory does not accept payment for higher placement.

How clinic and service information is verified

Identify a clinic. Candidate locations are found from provider location directories, health-service pages, official government registers and structured public business information.
Resolve the branch. Names, addresses and provider groups are normalised so that one clinic is not represented several times under minor spelling or brand variations.
Collect branch-level evidence. Services are added conservatively from the clinic or operator’s own location and service pages, official MRI and PET registers, or another clearly attributable public source.
Record provenance. Source links and evidence notes are retained in the research pipeline so a service claim can be reviewed when a source changes.
Publish cautiously. Ambiguous claims are omitted or held for review. Brand-wide statements are not automatically treated as proof that every branch offers the service.
Recheck and correct. Automated audits look for duplicate records, incomplete locations and stale evidence. Provider corrections are reviewed against current public information before the directory is updated.

A seeded validation audit of the research evidence pipeline checked 31 MRI and 10 PET service claims against operators’ own public information. All 41 sampled claims were supported. The estimated positive predictive value was 100%, with wide 95% confidence intervals because of the sample size: 89.0%–100% for MRI and 72.2%–100% for PET. This supports the conservative evidence approach, but it does not prove every listing is current or complete.

Preliminary national MRI and PET access study

Imaging Finder’s research project is a national cross-sectional geospatial analysis of access to Medicare-eligible MRI and PET. The analysis uses all 420 MRI practices and 139 PET practices on the Commonwealth location registers downloaded on 27 July 2026. It links those sites to 2,462 Statistical Areas Level 2 (SA2s), representing an estimated resident population of 27,613,654 at 30 June 2025.

Access measure	Estimated population within threshold
Medicare-eligible MRI within 30 minutes’ drive	89.5%
Medicare-eligible MRI within 60 minutes’ drive	94.9%
Medicare-eligible PET within 60 minutes’ drive	88.8%
Medicare-eligible PET within 120 minutes’ drive	94.2%

Preliminary results from the July 2026 study snapshot. Drive times use a free-flow car-routing model and do not measure appointment availability, cost or public-transport access.

The national averages conceal a sharp rural gradient. Estimated 30-minute access to Medicare-eligible MRI was 99.6% in Major Cities, 73.1% in Inner Regional Australia, 48.5% in Outer Regional Australia, 26.1% in Remote Australia and 0% in Very Remote Australia. By Modified Monash category it fell from 92.9% in MM3 large rural towns to 35.4% in MM4 medium rural towns and 17.6% in MM5 small rural towns.

Socioeconomic disadvantage was also associated with poorer geographic access after adjustment for remoteness and state. In the study model, each one-decile increase in area advantage was associated with higher adjusted odds of living within the primary threshold: adjusted odds ratio 1.170 for MRI within 30 minutes and 1.386 for PET within 60 minutes. These are area-level associations and must not be interpreted as individual patient effects.

What the study measured

Facility cohort: every row on the Commonwealth Medicare-eligible MRI and PET location registers; no registered site was excluded.
Population geography: 2021 Australian Statistical Geography Standard SA2 boundaries, with population-weighted centroids built from Census mesh-block counts.
Population: Australian Bureau of Statistics estimated resident population at 30 June 2025, allocated to SA2s.
Road travel: the shortest free-flow driving time to candidate sites on the OpenStreetMap road network using OSRM routing.
Equity: results stratified by jurisdiction, Remoteness Area, Modified Monash Model category and the Index of Relative Socio-economic Disadvantage.
Sensitivity testing: national MRI access was recomputed using 60,453 populated SA1 centroids. The 30-minute estimate changed by 0.39 percentage points, supporting the SA2 approach for national estimates.

Private MRI and the Medicare eligibility gap

A secondary, upper-bound analysis added 263 clinics with evidence-backed MRI services that did not match the Commonwealth eligible-practice register. Including all of them increased estimated national 30-minute coverage from 89.5% to 90.4%. The gain was concentrated in Inner and Outer Regional Australia and was zero in Remote and Very Remote Australia. About 249,720 people lived in 25 SA2s where the modelled MRI within 30 minutes was a non-register service.

This secondary cohort should be interpreted cautiously. A clinic may have an MRI service without that examination attracting a Medicare rebate, and the analysis cannot verify non-eligible sites against a statutory list by definition. It estimates a plausible geographic ceiling, not actual scanner capacity, billing or appointment access.

Limitations

Drive times model free-flow car travel. They do not include traffic, public transport, road closures or aeromedical access.
Geographic proximity is not realised access. Waiting time, workforce, scanner capability, referral acceptance, cost, Medicare item rules and patient circumstances all matter.
The analysis uses area centroids and is ecological. It cannot describe every resident’s journey or support individual-level conclusions.
Registers, websites and directory records can lag real-world service changes. The study therefore freezes all sources to stated dates.
Mobile and visiting services were excluded. Site counts are not scanner counts, throughput measures or measures of clinical quality.
The preliminary results have undergone internal review and reproducibility checks but have not yet completed independent peer review.

Governance, conflicts and corrections

Dr Mark Bekhit is a radiologist and the developer and owner of Imaging Finder. The directory is free to use and he derives no revenue from the services analysed. The research uses aggregate public facility and population data and does not contain patient information. Analysis code and the aggregate SA2 dataset are intended for open release with an archived identifier when the manuscript is published.

If a clinic listing or service appears incorrect, please use the contact page and include the clinic name, location and a current source. Corrections are checked before publication so that the public directory and the research provenance remain aligned.