Wheezes, crackles and rhonchi: simplifying description of lung sounds increases the agreement on their classification: a study of 12 physicians' classification of lung sounds from video recordings

Abstract

BackgroundThe European Respiratory Society (ERS) lung sounds repository contains 20 audiovisual recordings of children and adults. The present study aimed at determining the interobserver variation in the classification of sounds into detailed and broader categories of crackles and wheezes.MethodsRecordings from 10 children and 10 adults were classified into 10 predefined sounds by 12 observers, 6 paediatricians and 6 doctors for adult patients. Multirater kappa (Fleiss' κ) was calculated for each of the 10 adventitious sounds and for combined categories of sounds.ResultsThe majority of observers agreed on the presence of at least one adventitious sound in 17 cases. Poor to fair agreement (κ<0.40) was usually found for the detailed descriptions of the adventitious sounds, whereas moderate to good agreement was reached for the combined categories of crackles (κ=0.62) and wheezes (κ=0.59). The paediatricians did not reach better agreement on the child cases than the family physicians and specialists in adult medicine.ConclusionsDescriptions of auscultation findings in broader terms were more reliably shared between observers compared to more detailed descriptions.

Subjective and objective parameters in paediatric respiratory conditions: cultural adaptation to Portuguese population

“…However, in a recent publication, the ERS reported that a poor to fair agreement (κ < 0.40) was usually found for the detailed descriptions of the adventitious sounds, whereas moderate to good agreement was reached for the combined categories of crackles (κ = 0.62) and wheezes (κ = 0.59) (31). Also the existing studies about pulmonary auscultation in children have reported a poor to fair reliability (31,32). Other reasons may be related to these indings, such as the dif iculty in performing the technique of pulmonary auscultation in children, since they have small thoraxes, high respiratory rates, irregular breathing patterns and sometimes crying is present (33).…”

“…This is also a priority for the European Respiratory Society (ERS), who created the ERS Task Force on lung sounds, which includes experts from several countries, aiming the establishment of a consensus' nomenclature for pulmonary auscultation (30). However, in a recent publication, the ERS reported that a poor to fair agreement (κ < 0.40) was usually found for the detailed descriptions of the adventitious sounds, whereas moderate to good agreement was reached for the combined categories of crackles (κ = 0.62) and wheezes (κ = 0.59) (31). Also the existing studies about pulmonary auscultation in children have reported a poor to fair reliability (31,32).…”

Explaining Snowstorms

“…However, the latter technique of auscultation has improved revolutionary after the invention of the stethoscope by Laënnec [1]. However, despite being of tremendous value over decades, auscultation is only moderately reliable in the diagnosis and medical decision making even when used by experienced physicians [2]. Nevertheless, despite its limitations students are taught auscultation as the primary method for diagnosing chest diseases, owing largely to the low low costs and immediate available results associated with this method, which have contributed to preserving auscultation as a primary method of chest diagnostics, both in pulmonology and cardiology cases.…”

“…In many situations it has a reliability comparable with that of a CT-scan of the lung, and better compared to that of the chest X-ray, as well as a better performance than standard auscultation [6,7]. Further, it also reduced interobserver variability for the interpretation of lung sounds [2]. A major advantage of lung ultrasonography is that some lung abnormalities can be easily detected, sometimes within seconds, at the patient's bedside.…”

Auscultation of respiratory sounds: how to practise, how to teach?

“…They were asked to state whether the findings occurred during inspiration or expiration and whether the wheezes were high pitched or low pitched. The kappa values between the doctors were poor to fair, varying from 0 to 0.43 (9). When the fine and coarse crackles that were observed in either inspiration or expiration were combined as a single crackles category, and when the wheezes and rhonchi in either inspiration or expiration were combined as the single category of wheezes, the kappa values rose to 0.60, which still only meant modest agreement (9).…”