Accuracy of a composite score using daily SAPS II and LOD scores for predicting hospital mortality in ICU patients hospitalized for more than 72 h.

摘要: In most databases used to build general severity scores the median duration of intensive care unit (ICU) stay is less than 3 days. Consequently, these are not appropriate tools for measuring prognosis in studies dealing with ICU patients hospitalized more 72 h. Purpose: To develop a new prognostic model based on score (SAPS II), an organ dysfunction (LOD) and evolution both during first 3 days stay. Design: Prospective multicenter study. Setting: Twenty-eight units (ICUs) France. Patients: A training data-set was created four ICUs 18-month period (893 patients). Seventy percent were medical (628) aged 66 years. The SAPS II 38. hospital mortality rates 22.7% 30%, respectively. Forty-seven (420 patients) transferred from wards. this population, calibration (Hosmer-Lemeshow chi-square: 37.4, P=0.001) discrimination [area under ROC curves: 0.744 (95% CI: 0.714–0.773)] original relatively poor. validation data set random panel 24 French March 1999 (312 Measurements and main results: LOD calculated (SAPS1, LOD1), second (SAPS2, LOD2), third (SAPS3, LOD3) calendar days. alterations assigned value "1" when increased time "0" otherwise. multivariable logistic regression select variables measured three days, independently associated death. Selected were: at admission [OR: 1.04 1.027–1.053) per point], 1.16 1.085–1.253) transfer ward 1.74 1.25–2.42)], as well SAPS3-SAPS2 1.516 1.04–2.22)], LOD3-LOD2 2.00 1.29–3.11)]. final has good properties curve: 0.794 0.766–0.820), Hosmer-Lemeshow C statistic: 5.56, P=0.7]. set, maintained accuracy 0.826 0.780–0.867), 7.14, P=0.5]. Conclusions: using their days properties. We propose its use benchmarking evaluating over-risk death ICU-acquired nosocomial infections.