Objective: To assess reproducibility in data collection and its influence on the calculation of the severity scoring and mortality risk in APACHE II, APACHE III adapted for Spain and SAPS II.
Design: Multicenter, prospective, observational cohort study.
Setting: Nine Spanish Intensive Care Units (ICUs).
Patients: 1,211 consecutive patients admitted during the study period were included. Those patients under 16 years of age, those with a stay in the ICU of less than 24 hours, those admitted for scheduled pacemaker implant and those readmitted to the ICU within the same hospital admission were excluded.
Intervention: None. ENDPOINTS OF INTEREST: The data needed to calculate the severity and mortality risk scores were collected. A total of 10% of the patients were chosen by simple random sampling and the same data were collected by an independent group of intensive care physicians. Finally, the data obtained by the two groups of intensivists were compared.
Results: Significant differences were detected in the acute physiology score (APS) and severity score used for the calculation of APACHE III and SAPS II, and the predicted risk of death calculated for SAPS II. The percentage of agreement on admission diagnosis to the ICU was 50% for both APACHE II and III models. Nonetheless, in most of the patients (76.58% for APACHE II and 79.82% for APACHE III), the difference in the predicted risk of death due to the different assignation of diagnoses on admission to the ICU was less than 10%.
Conclusions: In this study, APS was the most influential factor on the reproducibility of severity scores and risk of death prediction. Admission diagnosis assignment had no significant impact on the reproducibility of the predicted mortality risk.