Impact of spectrograms on the classification of wheezes and crackles in an educational setting. An interrater study
AbstractBackgroundChest auscultation is a widely used method in the diagnosis of lung diseases. However, the interpretation of lung sounds is a subjective task and disagreements arise. New technological developments like the use of visual representation of sounds through spectrograms could improve the agreement when classifying lung sounds, but this is not yet known.AimsTo test if the use of spectrograms improves the agreement when classifying wheezes and crackles.MethodsWe used 30 lung sounds recordings. The sample contained 15 normal recordings and 15 with wheezes or crackles. We produced spectrograms of the recordings. Twenty-three third to fifth-year medical students at UiT the Arctic University of Norway classified the recordings using an online questionnaire. We first showed the students examples of how wheezes and crackles looked in the spectrogram. Then, we played the recordings in a random order two times, first without the spectrogram, then with live spectrograms displayed. We asked them to classify the sounds for the presence of wheezes and crackles. We calculated kappa values for the agreement between each student and the expert classification with and without display of spectrograms and tested for significant improvement. We also calculated Fleiss kappa for the 23 observers with and without the spectrogram.ResultsWhen classifying wheezes 13/23 (1 with p<.05) students had a positive change in k, and 16/23 (2 with p<.05). All the statistically significant changes were in the direction of improved kappa values (.52 - .75). Fleiss kappa values were k=.51 and k=.56 (p=.63) for wheezes without and with spectrograms. For crackles, these values were k=.22 and k=.40 (p=<0.01) in the same order.ConclusionsThe use of spectrograms had a positive impact on the inter-rater agreement and the agreement with experts. We observed a higher improvement in the classification of crackles compared to wheezes.