Lung Sound Classification Data from a Digital Stethoscope Clinical Trial

A prospective, single-center clinical trial involving 64 participants evaluated a novel digital stethoscope, Skeeper SM-300®. The dataset contains classifications of lung sounds (normal, crackle, wheezing, decreased sound) from both conventional and digital auscultation, assessed by three specialist raters. It was created by Sue In Choi and last updated in May 2026.

Use Cases

• Benchmarking digital stethoscope performance based on classifications from three specialist raters.
• Training machine learning models for lung sound classification based on categories like crackle and wheezing.
• Analyzing inter-rater agreement in medical audio assessment based on reported kappa statistics.
• Studying patient comfort and acceptability of remote digital health devices based on survey results mentioned in the description.

Strengths

• Data is from a structured, investigator-blinded clinical trial involving 64 participants.
• Includes ground truth classifications from conventional stethoscopy and three specialist raters, enabling validation.
• Reported statistical agreement (kappa values between 0.8 and 0.83) suggests high inter-rater reliability for the digital method.

Limitations

• Description metadata is limited; actual data quality requires manual inspection after download.
• Column-level documentation is absent; field semantics must be inferred after download.
• Row count is unknown, which may limit suitability assessment.

Provenance

Source

figshare

Collection Method

Single center, prospective, single-arm, investigator-blinded clinical investigation.

Freshness

Last updated 2026-05-13 17:40:27; freshness should be verified.