Tjitske van Engelen

173 Classifying study participants in clinical trials When this method would be applied in the entire OPTIMACT study group of 2,418 patients, we estimate to save a total of 703 hours of work by medical specialists. This has to be set against the (less expensive) hours of students and residents, which would approximate 900 hours and 286 hours, respectively. Calculations can be found in (Supplementary material 2). Validity Validation by the expert panel of 60 randomly selected cases resulted in agreement between students (30 cases) or students and resident (30 cases) versus the expert panel for 50 of the 60 cases (83% concordance; 95% CI 74–93%). In 46 cases (77%) there was total agreement, in 10 cases (17%) partial agreement, and in 4 cases (7%) disagreement on the classification. Further evaluation of the 14 partial and disagreement cases revealed that 4 cases were due to discordance on labels from the additional diagnostic categories only or procedural errors, leaving 10 cases as “true” disagreement. Of the 30 cases classified by students only there was total agreement with the expert panel in 27 cases and partial agreement in 3 (Table 1): one case where the students overdiagnosed an upper respiratory tract infection (URTI), one case of demand ischemia which the students labelled as an acute coronary syndrome and one case where the students deemed a new finding of atrial fibrillation a chance finding, not related to the reason for ED presentation. Of the 30 participants classified during the consensus meeting between students and a resident, there was total agreement with the expert panel in 19 cases, partial agreement in 7 cases and disagreement in 4 (Table 2). Amongst the partial agreement cases (7/30) there was one case where the students and the resident diagnosed a community-acquired pneumonia (CAP) in addition to a pulmonary embolism, whereas the expert panel deemed the infiltrate on the chest x-ray an infarction due to the pulmonary embolism. Furthermore, there were two cases where the students overdiagnosed an URTI and one case where the students deemed a history of lung transplantation not relevant for the current ED presentation. There was one case with discordance due to a procedural error, one case with discordance on the students’ additional diagnostic label thoracic pain of unknown origin and one case with discordance on the students’ additional diagnostic label extrathoracic pathology. In the four cases of disagreement, there was one case concerning a patient with a pulmonary infection in addition to a urinary tract infection. The chest x-ray described a bronchiolitis and the students classified this as a lower respiratory tract infection (LRTI), other than CAP. The expert panel reassessed the chest x-ray, found arguments that contradicted the chest x-ray report, and assigned a diagnosis of CAP. There were two cases where the students overdiagnosed an URTI and one case with discordance on the additional diagnostic label thoracic pain of unknown origin. 8

RkJQdWJsaXNoZXIy MTk4NDMw