Accuracy of laypeople, symptom checkers, and large language models with patient-generated case vignettes
- Disease/Condition
- None ;Various diseases are shown in the form of case vignettes. No disease/health problem is examined directly.
- Study type
- Observational
- Primary sponsor
- Fachgebiet Arbeitswissenschaft, Technische Universität Berlin10587 BerlinGermany
- Funder
- Fachgebiet Arbeitswissenschaft, Technische Universität Berlin10587 BerlinGermany
- Location
- United States
- Phase
- N/A
- Start date
- N/A
- End date
- N/A
In this study, we will determine the accuracy and test-theoretic metrics of laypeople, symptom checkers, and large language models. In order to use more externally valid cases than previous case vignette-based studies, we use cases phrased by patients as they experience these symptoms. These descriptions are sampled from freely and publicly available "ask-the-doctor" platforms. In addition to accuracy, we will also determine how test-theoretic measures affect the selection of an appropriate vignette subset for laypeople, symptom checkers, and large language models and whether this subset differs between these three actors.
- Registry
- Deutsches Register Klinischer Studien
- Trial ID
- DRKS00032895
- Type
- Non-Device Trial
Access comprehensive clinical trial information for DRKS00032895 through Pure Global AI's free database. This phase not specified trial is sponsored by Fachgebiet Arbeitswissenschaft, Technische Universität Berlin10587 BerlinGermany and is currently Not yet recruiting. The study focuses on None ;Various diseases are shown in the form of case vignettes. No disease/health problem is examined directly..
This page provides complete trial specifications, intervention details, outcomes, and location information. Pure Global AI offers free access to Deutsches Register Klinischer Studien data, helping medical device and pharmaceutical companies navigate clinical research efficiently.