Skip to content

Accuracy of laypeople, symptom checkers, and large language models with patient-generated case vignettes

Accuracy of laypeople, symptom checkers, and large language models with patient-generated case vignettes

Status
Active, not recruiting
Phases
Unknown
Study type
Observational
Source
DRKS
Registry ID
DRKS00032895
Enrollment
200
Registered
2023-10-18
Start date
2023-10-18
Completion date
Unknown
Last updated
2025-04-07

For informational purposes only — not medical advice. Sourced from public registries and may not reflect the latest updates. Terms

Conditions

Various diseases are shown in the form of case vignettes. No disease/health problem is examined directly.

Interventions

Group 1: Medical laypeople are presented with a selection of 45 different case vignettes (sampled from "ask-the-doctor" platforms) that they are asked to evaluate Group 2: Symptom checkers are tested

Sponsors

Fachgebiet Arbeitswissenschaft, Technische Universität Berlin
Lead Sponsor

Eligibility

Sex/Gender
All
Age
18 Years to No maximum

Inclusion criteria

Inclusion criteria: US residents

Exclusion criteria

Exclusion criteria: - Participation in prior studies of the chair/institute - Professional medical training

Design outcomes

Primary

MeasureTime frame
Accuracy when assessing vignettes. For each case vignette, participants decide whether it is (a) an emergency, (b) urgent symptoms (seeking care within 1 day), (c) a non-emergency where care is sufficient within a week, or (d) symptoms that can be self-treated.

Secondary

MeasureTime frame
- Accuracy differentiated by triage level - Safety of advice - Comprehensiveness (for symptom checkers only) - Inclination to overtriage - Item Difficulty - Item-Total Correlation - Capability Comparison Score

Countries

United States

Contacts

Public ContactMarvin Kopka

Fachgebiet Arbeitswissenschaft, Technische Universität Berlin

marvin.kopka@tu-berlin.de+49 30 314-70806

Outcome results

None listed

Source: DRKS (via WHO ICTRP) · Data processed: Feb 4, 2026