Accuracy of 22 Symptom Checker Apps Tested on 45 Clinical Vignettes in 2020

Two data sets present the triage and diagnostic accuracy of 22 commonly used symptom checker apps sampled in 2020. The apps were tested on 45 fictitious patient case vignettes, yielding 796 unique triage evaluations and 520 unique diagnostic evaluations. The data was collected by Anna Dames for an MSc degree at Technische Universität Berlin and supplements a 2022 paper.

Use Cases

• Benchmarking triage accuracy of symptom checker apps based on the 796 unique evaluations.
• Comparing diagnostic advice performance across apps based on the 520 unique diagnostic evaluations.
• Analyzing the consistency of app performance across different clinical vignettes described in the study.

Strengths

• Contains evaluations from 22 commonly used symptom checker apps.
• Based on testing with 45 standardized clinical case vignettes.
• Provides a substantial sample size with 796 triage and 520 diagnostic evaluations.

Limitations

• Column-level documentation is absent; field semantics must be inferred after download.
• Row count is unknown, which may limit suitability assessment.
• Data may reflect temporal bias inherent to the 2020 sample and specific vignette adaptations.

Provenance

Source

Anna Dames, Technische Universität Berlin; paper authored by Malte L Schmieding.

Collection Method

Apps were tested on a set of 45 clinical case vignettes, which were originally compiled by Semigran et al. in 2015 and later adapted.

Time Range

2020