Objectives : to assess the operational characteristics of the physical examination in the diagnosis of pneumonia and to evaluate the inter-observer agreement. Framework : the studies which evaluated the physical examination as a diagnostic tool in pneumonia are methodologically deficient. Design : blind, cross-sectional study for diagnostic evaluation. Patients : adults who consult the emergency department and hospitalization of the FCI by acute respiratory symptoms or exacerbation of the same. Measurements : physical examination by two independent observers, X-ray was taken and readed by radiologist expert. Data were taken aiming to calculate the pneumonia severity scale SP.Results : of 198 patients, 85(42%) had pneumonia radiographically. The operational features of the examiner1 were: sensitivity:63.2%, specificity:54.1%, LR(+)=1.36, LR(-)=0.68. For examiner2: Sensitivity:34.3%, specificity:71 7%, LR(+)1.17, LR(-)0.92. The correlation between the clinical diagnosis for pleural effusion was k=0.052, notsignificant (p = 0.445); and for pneumonia was k=0.25 and was significant (p =0.022). When measuring by pneumonia severity index score, there was a increased sensitivity according to the severity (II:Sensitivity:40%; III:Sensitivity:57%; IV:Sensitivity:75%,V:Sensitivity:80%). Conclusions : physical examination is neither sensitive nor specific in the diagnosis of pneumonia. There is a weak agreement rate in the chest physical examination for the diagnosis of pleural effusion and pneumonia. Pneumonia is more likely to be diagnosed on clinical grounds when severity is higher.
