Outcome of pediatric intensive care at six centers in Mexico and Ecuador

Objective: To improve understanding of the causes of morbidity and mortality among critically ill children in the countries studied. Design: Survey of hospital records between 1992 and 1994. Setting: Six pediatric intensive care units (ICUs) (four ICUs in Mexico City and two ICUs in Ecuador). Patients: Consecutive patients (n = 1,061) admitted to the units studied. Interventions: None. Measurements and Main Results: The mortality rate for low-risk patients (Pediatric Risk of Mortality [PRISM] score of less than or equal to 10, n = 701) was more than four times the rate predicted by the PRISM score (8.1% vs. 1.8%, p <.001), with an additional 11.3% of this group incurring major morbidity. The mortality rate for moderate-risk patients (PRISM scores of 11 to 20, n = 232) was more than twice predicted (28% vs. 12%, p<.001). For low-risk patients, death was significantly associated with tracheal intubation, central venous cannulation, pneumonia, age of <2 months, use of more than two antibiotics, and nonsurgical diagnosis (after controlling for PRISM score). Central venous cannulation and tracheal intubation in the lower-risk groups were performed more commonly in units in Mexico than in one comparison unit in the United States (p <.001). Conclusions: For six pediatric ICUs in Mexico and Ecuador, mortality was significantly higher than predicted among lower-risk patients. Tracheal intubation, central catheters, pneumonia, sepsis, and nonsurgical status were associated with poor outcome for low-risk groups. We speculate that reducing the use of invasive central catheters and endotracheal intubation for lower-risk patients, coupled with improved infection control, could lower mortality rates in the population studied.