Skip to content
Partaw Research
Menu

Mission

The films exist.
The readers don’t.

Across Afghanistan, clinics own working X-ray machines and have no radiologist to read the films — for a child’s fractured arm as much as for a tuberculosis screen. Partaw Research works to close that reading gap as a public good, and to build the missing evidence base while doing it.

About 100

radiologists per million people in high-income countries

Fewer than 2

radiologists per million people in low-income countries

Radiologists per million people, one mark each. The gap is not a shortage of machines or of films; it is a shortage of the people qualified to read them. The low-income figure is an upper bound — fewer than two per million — so it is drawn down to the one whole mark that bound admits.1
Two-thirds

67% of people, without basic radiology

of the world’s population lacks access to basic diagnostic radiology. 1
Dari or Pashto
every report is delivered in the clinic’s own language, over the WhatsApp thread it already uses.
One signature
no report reaches a clinic without a licensed radiologist’s name and license number on it. That rule lives in the platform’s code, not in a policy document.

Not a machine gap.
A reader gap.

An X-ray is only as useful as its reading. Decades of aid and procurement have put working machines into district clinics and provincial hospitals; the training pipelines that produce people qualified to read the output have not kept pace, and the few specialists there are concentrate in the largest cities. So films wait, travel with relatives to Kabul, or go unread — and the fracture is set blind, the pneumonia is treated on guesswork, the tuberculosis walks back out the door.

This is what we mean by radiology access as a public good: the reading, not the machine, is the scarce resource. It can be delivered over infrastructure clinics already hold in their hands — a smartphone and a WhatsApp thread — by a licensed radiologist who may be anywhere, with a signature and a verification code that make the report worth trusting.

What we do

A service worth studying. A study worth serving.

The reading, the evidence, and the corpus are one operation, run three ways at once.

Get films read
A licensed radiologist’s signed report, in Dari or Pashto, returned over the channel clinics already use — WhatsApp. Access and turnaround, where before there was no reader at all.
Build the evidence
Teleradiology in this setting is barely described in the literature. We run programs as studies — protocols, outcomes, and publications with academic partners — so the model can be judged, improved, and reused.
Grow a governed dataset
Consented, de-identified reads accumulate into a research corpus of rural Afghan radiology — a population nearly absent from the datasets modern imaging is built on — and, once counted weekly by province and finding, into a disease signal for the national surveillance system.

Outbreaks are counted.
From a sample.

In the first half of 2026, Afghanistan’s national surveillance counted suspected Crimean-Congo haemorrhagic fever rising twelvefold and acute watery diarrhoea more than threefold, while respiratory infection and pneumonia — the largest line in the table — ran above a hundred thousand cases a month through the winter3.

Every one of those counts comes from the sentinel sites: one public facility in seven, reporting weekly. Private clinics — where much of the country’s X-ray capacity sits — report to no one. The films read on this platform come from those clinics, and a chest film read and signed says whether a pneumonia is present, where, and when. That is the surveillance program.

Jan 2026: 41 suspected cases Feb 2026: 38 suspected cases Mar 2026: 59 suspected cases Apr 2026: 152 suspected cases May 2026: 176 suspected cases Jun 2026: 483 suspected cases Jan Jun

Crimean-Congo haemorrhagic fever × 12

41 → 483 suspected, a month

Jan 2026: 5,299 with dehydration cases Feb 2026: 5,968 with dehydration cases Mar 2026: 6,308 with dehydration cases Apr 2026: 10,703 with dehydration cases May 2026: 13,645 with dehydration cases Jun 2026: 18,232 with dehydration cases Jan Jun

Acute watery diarrhoea × 3.4

5,299 → 18,232 with dehydration, a month

Jan 2026: 128 suspected cases Feb 2026: 173 suspected cases Mar 2026: 115 suspected cases Apr 2026: 283 suspected cases May 2026: 224 suspected cases Jun 2026: 255 suspected cases Jan Jun

Dengue fever +99%

128 → 255 suspected, a month

Jan 2026: 1,923 suspected cases Feb 2026: 2,603 suspected cases Mar 2026: 2,395 suspected cases Apr 2026: 3,181 suspected cases May 2026: 2,318 suspected cases Jun 2026: 2,414 suspected cases Jan Jun

Measles +26%

1,923 → 2,414 suspected, a month

Jan 2026: 123 confirmed cases Feb 2026: 93 confirmed cases Mar 2026: 81 confirmed cases Apr 2026: 87 confirmed cases May 2026: 46 confirmed cases Jun 2026: 87 confirmed cases Jan Jun

COVID-19 −29%

123 → 87 confirmed, a month

Jan 2026: 162,472 cases Feb 2026: 157,046 cases Mar 2026: 105,904 cases Apr 2026: 131,697 cases May 2026: 80,330 cases Jun 2026: 65,673 cases Jan Jun

Respiratory infection & pneumonia −60%

162,472 → 65,673 cases a month

The numbers, as a table
New cases, 2026 JanFebMarAprMayJun
Crimean-Congo haemorrhagic fever (suspected) 413859152176483
Acute watery diarrhoea (with dehydration) 5,2995,9686,30810,70313,64518,232
Dengue fever (suspected) 128173115283224255
Measles (suspected) 1,9232,6032,3953,1812,3182,414
COVID-19 (confirmed) 1239381874687
Respiratory infection & pneumonia (ARI) 162,472157,046105,904131,69780,33065,673
New cases reported each month by the national surveillance system, January to June 2026, one panel per disease. Each panel is drawn on its own scale from zero; the panels are ordered by how far the count moved, so the outbreaks come first. Transcribed from WHO Afghanistan’s monthly situation reports and reconciled against their cumulative totals.3

The corpus

A corpus that exists nowhere else

Rural and peri-urban Afghan radiology, photographed from real films on real phones in a conflict-affected, high-burden setting — a population and an image modality nearly absent from the corpora that imaging research and imaging AI are built on. It accumulates read by read.

Research & partnerships
Consented at intake, before a film enters any research use.
De-identified at the platform boundary, before research use.
Shared under data-governance agreements. Never sold.

Co-author it. Advise it. Validate on it.

We are looking for academic radiology and global-health programs, imaging-AI teams, TB-ecosystem NGOs and funders, and the surveillance programmes and implementing entities who need a private-clinic signal. What is on offer — authorship, advisory seats, a governed validation site, and a surveillance feed — is laid out plainly on the research page.


References

  1. 1 Radiological Society of North America (2022), “The Global Radiologist Shortage — an Escalating Crisis”: fewer than two radiologists per million inhabitants in low-income countries versus roughly one hundred per million in high-income countries; an estimated two-thirds of the world’s population has little or no access to basic radiology services. www.rsna.org/news/2022/may/global-radiologist-shortage
  2. 2 World Health Organization, Global Tuberculosis Report — Afghanistan country profile. Afghanistan is a high-TB-burden country in the WHO Eastern Mediterranean Region. www.who.int/teams/global-tuberculosis-programme/tb-reports
  3. 3 World Health Organization, Afghanistan Emergency Situation Reports Nos. 60–65 (January–June 2026), “Summary of infectious disease cases in Afghanistan” annex: new monthly cases by indicator from the national indicator-based surveillance system. The sentinel-site basis of that system (519 of 3,750 facilities; no private-facility reporting mechanism) is described in Ahmadi et al. (2024), PMC11059675. www.emro.who.int/images/stories/afghanistan/emergency-situation-report-65-June-2026.pdf