Interrater reliability of auscultation of breath sounds among physical therapists

Background and Purpose. Although auscultation is routinely used in the assessment of respiratory status, the ability of the rater to accurately and consistently identify lung sounds has been questioned. The literature on this issue is sparse and has focused on reliability of auscultation of tape-recorded rather than in vivo lung sounds. The purpose of this study were to determine the interrater reliability of physical therapists in the direct auscultation of lung sounds based on their clinical experience in chest physical therapy and to determine whether the adoption of standardized nomenclature and education on proper technique and interpretation affects reliability. Subjects and Methods. A group of 57 registered physical therapists were stratified by clinical experience into four groups. Sixteen therapists (ie, 4 in each stratum) were randomly chosen using a random number table. The following criteria were developed to delineate clinical experience. Group 1 subjects were senior chest physical therapists with at least 5 years of experience in this area of practice. Group 2 subjects were experienced therapists who had a minimum of 2 years of experience in chest physical therapy and were currently practicing in this area. Group 3 subjects were experienced physical therapists in other areas who were also practicing in chest physical therapy on occasional weekend service. Group 4 subjects were new graduates. Ten patients were evaluated by each group of 4 physical therapists using a teaching stethoscope with one diaphragm/bell and four pairs of earpieces. The education session consisted of discussion of the adoption of standardized nomenclature and education on proper technique and interpretation of auscultation. Interrater reliability was assessed before and after the education session using kappa (kappa) values. Comparisons were made between kappa values before and after the education session to determine the effect of education and between groups to determine the effect of clinical experience. Results. The kappa values before the education session were low, indicating poor reliability in detecting specific abnormal sounds (kappa = -.02-.59). Following the education session, there was a general improvement in reliability (kappa = -.30-.77), especially for group 3 (specialists in other areas). The most improvement was noted for the detection of the quality of breath sounds (kappa = -.08-.50). Conclusion and Discussion. Reliability of auscultation was poor to fair, in general, before the education session. There was a definite improvement in reliability after the education session. There was no clear effect of clinical experience on reliability, and the agreement among observers appeared to depend on the abnormal lung sound present. Limitations of this study and recommendations for future research are discussed.