AUTOMATED SEVERITY SCORING IN LEVEL-1 TRAUMA PATIENTS

The authors evaluated the sensitivity and specificity of a computerize d Simplified Acute Physiology Score (SAPS) for outcome prediction in L evel I trauma patients admitted to a Surgical ICU (SICU). SAPS was com pared with the combined Trauma Score (TS) and Injury Severity Score (I SS). 1434 consecutive trauma patients admitted to the SICU over a 3-ye ar period were studied. All patients had the SAPS automatically calcul ated on the first SICU day. Patient data was extracted from an electro nic flowsheet, and the most abnormal values for the previous 24 hours were used to calculate the SAPS. TS and ISS were calculated by a train ed nurse. The relationship among the severity scores, ICU length of st ay (LOS), and survival was evaluated. A logistic regression equation w as calculated for SAPS alone and for TS combined with ISS. The predict ive power of the severity methods was compared using Receiver Operatin g Characteristic (ROC) curve analysis. Scores for survivors and non-su rvivors were compared with Student's t-tests. 1085 patients had comple te data available. There were 995 survivors and 90 non-survivors. The mean (+/- standard error of the mean) ISS was 12.7 (+/- 3.2) with 36 p er cent of the patients having an ISS greater-than-or-equal-to 15. The mean SAPS was 8.1 (+/- 2.5). Survivors had a significantly lower SAPS than non-survivors, 7.0 versus 20.2 (P < 0.0005) and a shorter LOS, 2 .5 versus 4.9 days (P < 0.002). ROC curve analysis revealed no statist ically significant difference in the areas under the two curves, indic ating that the SAPS was equivalent to TS combined with ISS in outcome prediction (P > 0.70). Computerized SAPS is equivalent to the combinat ion of TS and ISS for prediction of patient outcome after SICU admissi on following major trauma. Since SAPS measurements can be automated, i t is the easier system to use.