To determine the clinical and laboratory features of patients who are readmitted to the ICU after discharge, and to use these to develop a score to predict risk of readmission.
In this observational cohort study in a mixed, 31-bed, medico-surgical intensive care unit, 226 consecutive adult ICU admissions were included during a 13-week period (excluding patients admitted for uncomplicated postoperative surveillance).Variables related to epidemiology, past history, clinical and lab data were recorded. Data were analyzed by univariate regression analysis. Continuous variables were transformed in categorical variables using the threshold obtained by a Lowess smoothing function. The categorical variables were entered into a multiple logistic regression model. The coefficients of the regression were used to compute the score for each significant item. A p < 0.05 was considered significant.
Of 226 admissions during the study period, 34 (17.4%) patients died in the ICU. The mean length of ICU stay was 6.4 days. A total of 42 patients (22%) were readmitted. The score developed from the statistical analysis of the 272 variables includes 8 simple variables with a maximum score of 41 (Table). The area under the receiving operating characteristic (ROC) curve was 0.74.
This simple score, based on routinely collected variables, can help to evaluate the risk of readmission to the ICU after discharge.
ICU readmission is associated with worse outcomes. Identifying patients at risk of readmission may prevent premature ICU discharge, may promote the use of increased surveillance such as intermediate care unit or telemetry on the floor, and may help prepare the patient, the relatives, and the heath care professionals for the possibility of readmission.
VariableScorePresence of decubitus ulcer12Hyperleukocytosis or inflammatory syndrome (WBC > 17,000/mm3 or CRP > 20 mg/dL)6Maximal sequential organ failure assessment (SOFA) score > to 96Paroxysmal atrial fibrillation or need for anti-arrhythmic medication4Chronic obstructive pulmonary disease (COPD) or asthma requiring chronic treatment4Morbid obesity or parenteral nutrition or decompensated diabetes4Hypoalbuminemia (serum albumin <3.5 g/dL)3Alteration of mental status, polytrauma or notable physical dependence2
Michaël De Cubber, None.