OBJECTIVES. To evaluate the pattern of unplanned readmissions to the intensive care unit and identify patients at risk of readmission.

DESIGN. Nested case-referent study.

SETTING. Tertiary hospital, Hong Kong.

PATIENTS. A total of 146 patients with unplanned intensive care unit readmission were compared with 292 control patients who were discharged from the intensive care unit alive and never readmitted. Cases and controls were matched for age, gender, and disease severity.

MAIN OUTCOME MEASURES. Patient demographics, initial and pre-discharge clinical parameters, reasons for readmission, and outcomes were studied.

RESULTS. During the 30-month study period, the readmission rate was 5.1%. Readmitted patients had significantly higher mortality and longer mean hospital lengths of stay (both P<0.001). Most patients in this cohort (36.3%) were readmitted for a respiratory cause. Based on classification tree analysis, postoperative patients with sepsis (adjusted P=0.043), non-operative septic patients with fluid gain 24 hours pre-discharge (adjusted P=0.013), and non-septic patients with increased sputum quantity on discharge (adjusted P=0.006) were significantly associated with intensive care unit readmission.

CONCLUSION. Incomplete resolution of respiratory conditions remained an important reason for potentially preventable intensive care unit readmission. Attention to fluid balance and sputum quantity before intensive care unit discharge might prevent unplanned intensive care unit readmission.

weblink here