Objective: To determine the agreement of the diagnosis 'acute respiratory distress syndrome' (ARDS) established by intensive care (IC) specialists, with that according to internationally accepted objective criteria.
Design: Descriptive inventory.
Method: All 119 hospitals in the Netherlands with an IC department were asked to participate; 34 did (29%). On 3 consecutive days, IC specialists completed a case-record form concerning the respiratory status and additional treatment of all of their patients. In the case of mechanical ventilation, the specialist could indicate whether ARDS or another condition was the cause of the respiratory insufficiency. In addition to this, objective data were requested so that the investigators could establish whether there was ARDS on the basis of the North American-European Consensus Conference (NAECC) criteria and the less generally accepted 'Lung injury score' (LIS > or = 2.5).
Results: Of the 266 patients about whom a case record form was returned, 151 were mechanically ventilated. ARDS was diagnosed in 36 of these patients according to the IC specialists, in 24 according to the NAECC and in 20 according to the LIS criteria (p < 0.05 versus IC specialists). The chance-corrected agreement (kappa) between the diagnoses by IC specialists and the NAECC criteria was 0.34 (p < 0.001) and between the IC specialists and the LIS 0.44 (p < 0.001). The kappa between the NAECC and the LIS diagnoses was 0.42 (p < 0.001). Using NAECC and/or LIS criteria as the golden standard, a correct ARDS diagnosis was made by the IC specialists in the case of 20 patients, a false-positive diagnosis in the case of 16 patients and a false-negative diagnosis in the case of 13 patients.
Conclusion: In this investigation, the diagnosis of ARDS was more frequently established than would occur according to the NAECC criteria.