Medical-AI trial: a large language model to help diagnose rare diseases — can it shorten the diagnostic odyssey? (cluster-randomized) (NCT07650799)
A multicenter, prospective, cluster-randomized, parallel-controlled real-world trial of whether a rare-disease diagnostic large language model (LLM) can improve diagnostic quality, efficiency, and health-economic outcomes for physicians managing patients with suspected rare or diagnostically unresolved disease. Primary outcomes are top-3 candidate diagnostic accuracy and rare genetic-disease diagnostic yield. 1,055 patients.
Trial overview (primary data)
- StatusNot yet recruiting
- ConditionsRare Disorders, Rare Diseases
- InterventionsOTHER: AI system
- SponsorPeking Union Medical College Hospital
- Target enrollment1,056 participants
- Period2026-08-01 〜 2027-12-01
Key points
- A real-world test of whether a rare-disease diagnostic LLM improves diagnostic quality, efficiency, and health economics for physicians managing suspected rare or unresolved disease.
- Multicenter, prospective, cluster-randomized, parallel-controlled effectiveness study. 1,055 patients.
- Primary outcomes are top-3 candidate diagnostic accuracy and rare genetic-disease diagnostic yield.
- Rare diseases number in the thousands; the years-long diagnostic odyssey burdens patients and families.
- It measures real-world outcomes by randomization, not only accuracy; an evaluation stage, not an establishment of diagnosis replacement.
- Rare diseases number in the thousands and present non-specifically, so diagnosis can take several physicians and years.
1Thousands of rare diseases, delayed diagnoses
Rare diseases each affect few patients, but there are thousands of them, and together they are far from rare. Many present with non-specific symptoms, and reaching the right diagnosis can mean moving from doctor to doctor over years. This is called the diagnostic odyssey, a heavy burden on patients and families.
A large language model (LLM) that can handle vast medical knowledge raises the hope of listing candidate diseases from symptoms and shortening that odyssey.
2Testing an LLM in real clinical practice
This study tests such a rare-disease diagnostic LLM in real practice. Per the registry summary, it evaluates, as a multicenter, prospective, cluster-randomized, parallel-controlled real-world effectiveness study, whether the LLM can improve diagnostic quality, efficiency, and health-economic outcomes for physicians managing patients with suspected rare or diagnostically unresolved disease.
The primary outcomes are the diagnostic accuracy of the top-3 candidates the LLM lists and the rare genetic-disease diagnostic yield (the proportion reaching a diagnosis). The size is 1,055, randomized by cluster (groups of sites or physicians).
3Why the diagnostic odyssey runs long
Rare diseases have few patients each while numbering in the thousands. The length of the path to a correct diagnosis has a reason at each stage.
- 1Symptoms are non-specificMost rare diseases present without distinguishing features
- 2Patients move from doctor to doctorSeveral physicians are seen before the right diagnosis
- 3Years passCalled the diagnostic odyssey, it burdens patients and families
- 4Candidates shorten itAn LLM proposes candidate diseases, with accuracy of its top three measured
This is a multicentre, prospective, cluster-randomised, parallel-controlled real-world effectiveness study, with primary endpoints of top-three diagnostic accuracy and the rate at which rare genetic diagnoses are reached, across 1,055 cases. The risk of plausible error means the tool's strength in supporting physician judgement has to be measured carefully.
Why it matters
It confirms in a randomized comparison not only accuracy on paper but real-world outcomes: whether it aids physicians, shortens the odyssey, and helps even accounting for cost. Rare diseases, with scattered information and few specialists, may be an especially good fit for an LLM (this study is an evaluation stage, not an establishment of diagnosis replacement).
FAQ
What is the diagnostic odyssey?
Does the AI confirm a rare-disease diagnosis?
Sources (primary)
Source: ClinicalTrials.gov (U.S. NIH/NLM, public domain). This site does not provide medical advice. Verify the latest and exact details with the official source. This site is not endorsed or certified by the NIH/NLM.
- ClinicalTrials.gov (study record, original)
- NCT ID: NCT07650799