Cuevas SA, Fernández RF

Language: Spanish
References: 0
Page: 27-36
PDF size: 164.64 Kb.

ABSTRACT

Introduction: predictive models are an important tool in intensive care. The SEGRAV 23 scale has proven to be useful to establish a prognosis in pediatric patients.
Objective: determine SEGRAV 23 higher risk interventions.
Methods: an observational analytical retrospective cohort study based on application of the SEGRAV 23 model was conducted with community-acquired pneumonia patients at the Pediatric Intensive Care Unit of "Dr. Luis Díaz Soto" Central Military Hospital during five years (2007-2008, 2012-2014). The sample consisted of 356 patients. Adjustment chi square was estimated, and independence verified by Pearson's chi-squared test and Fisher's exact test, with a statistical significance level of p< 0.05. Nonpoint variables and those with a lesser influence on the deceased category (p> 0.05) were discarded. Odds ratio (OR) was estimated with a confidence interval (CI) of 95 % to determine risk.
Results: mortality was 2.53 % (356 admissions/9 deaths), with a predominance of non-critical patients (222/62.4 %; p< 0.05). All the deaths were among very critical (12/3.4 %) and critical (9/2.5 %) patients. The most frequent procedures were one central venous route (71/19.9 %) and treatment for severe hydroelectrolitic disorders (63/17.7 %). Transcranial Doppler, total parenteral nutrition, and the treatment for disseminated intravascular coagulation did not score. Tomography, one central venous route, surgery and pleurotomy exhibited a not very significant relationship to mortality (p> 0.05). Cardiopulmonary resuscitation (OR= 1 380; CI 95 % [113.198- 16 823.63]) and the use of FiO2≥ 60 % (OR= 454.67; IC 95 % [48.89-4 228.57]) displayed higher risk.
Conclusions: of the 23 SEGRAV 23 diagnostic and therapeutic interventions, 15 were found to be associated with a higher risk of mortality.