Russia · Moscow · See the Russia profile · See the Moscow profile
Evidence: Observational / pre–post Top 6% 87/100 · Ask Evidence Copilot about this practice
Moscow's public health department centrally benchmarks and deploys a marketplace of 50+ vetted computer-vision diagnostic tools as a mandatory "second reader" for radiologists across 150 city facilities; a 3,409-scan prospective study found AI-assisted radiologists beat AI runnin
Moscow's public health system began evaluating AI diagnostic-imaging tools in 2019 through the 'Moscow Experiment,' a prospective programme run by the city's Research and Practical Clinical Center for Diagnostics and Telemedicine Technologies, which benchmarked computer-vision services against real clinical cases from a large imaging archive.
The city aims to centrally vet AI diagnostic tools before deployment and make them available as a mandatory 'second reader' for radiologists — assisting rather than replacing clinical judgement — across its public health network.
The evaluation fed into MosMedAI, a city-run marketplace that went live in early 2024, giving radiologists in Moscow's public polyclinics and hospitals access to 50+ vetted AI services for CT, mammography and other imaging. The deployment now covers 150 medical facilities, including children's hospitals, and since early 2023 AI-assisted mammography reading has been funded through Russia's compulsory health insurance system.
The strongest evidence is a prospective multicentre diagnostic-accuracy study run across 67 Moscow health organisations between April 2022 and December 2024, covering 3,409 brain CT studies including 1,101 confirmed intracranial haemorrhage cases. Radiologists working with AI assistance reached 98.91% sensitivity, against 95.91% for the AI model working alone, with the assisted workflow's specificity at 99.83%. The health department also reports that AI-assisted mammography reading has cut interpretation time more than eight-fold.
Independent coverage of the study frames its central finding candidly: AI alone is not yet a substitute for a human radiologist, and the value comes from combining the two — a conclusion this record reports as published, without inflating it into a stronger claim than the study itself makes.
National / regional programmes
Indicative funding routes for practices of this type — always check each programme's current calls and eligibility rules.
Do you run this practice? Claim it — verified implementers get a public contact pathway and can propose corrections.
Where this practice's information was retrieved from, and when.
Kenya
Jacaranda Health's PROMPTS service uses a Swahili-language AI model to answer 10,000-12,000 daily SMS questions from nearly 3 million enrolled …
Congo (the Democratic Republic of the)
DR Congo's malaria programme (PNLP) uses Fio's 'Deki Reader' devices, applying computer-vision analysis to read malaria rapid diagnostic tests; a …
South Korea
South Korea's welfare ministry links 47 categories of government data — health, finance, utilities, employment — to flag at-risk households …
Papua New Guinea
A truck-mounted digital X-ray unit using Delft Imaging's CAD4TB AI software screened 43% of Daru Island's population for tuberculosis, finding …
Open full copilot Grounded in cited practices — always check the sources.