Objectives: To determine whether children who experience longer intensive care unit (ICU) stays after open heart surgery may be identified at admission by clinical criteria. To identify factors associated with longer ICU stays that are potential targets for quality improvement.
Setting: Tertiary pediatric cardiac surgical center.
Design: A retrospective review was performed of pre-, intra-, and postoperative factors for children undergoing open heart surgery. All factors were evaluated for strength of association with length of ICU stay (LOS) using a negative binomial model. After multiple analysis, factors were deemed significant if associated with a LOS with p < .02.
Patients: A total of 355 pediatric patients who had cardiac surgery with cardiopulmonary bypass in a 1-yr period from April 1999 until March 2000.
Measurements and Main Results: Children who fell above the 95th percentile for LOS in our institution occupied 30% of bed days and had a three-fold greater mortality. Of all clinical factors considered, those significantly associated with LOS were as follows:preoperative-mechanical ventilation, neonatal status, medical problems, and transfer from abroad;intraoperative-higher operative complexity, increased cardiopulmonary bypass time or ischemic time, and circulatory arrest; and postoperative-delayed sternal closure, sepsis, renal failure, pulmonary hypertension, chylothorax, diaphragm paresis, and arrhythmia. A model combining all factors identified preoperative mechanical ventilation, neonatal status, major medical problems, operative complexity, cardiopulmonary bypass time, and a postoperative complication score as independently associated with LOS (p < .01).
Conclusions: At the time of ICU admission after open heart surgery, clinical criteria are evident that highlight a child's risk of longer ICU stay. These pre- and intraoperative factors relate to LOS independent of subsequent postoperative events. Those postoperative complications that are most strongly associated with increased LOS are identified and, therefore, made accessible to quality control.