Admission Risk Score to Predict Inpatient Pediatric Mortality at Four Public Hospitals in Uganda.

摘要: Mortality rates among hospitalized children in many government hospitals sub-Saharan Africa are high. Pediatric emergency services these often sub-optimal. Timely recognition of critically ill on arrival is key to improving service delivery. We present a simple risk score predict inpatient mortality children. Between April 2010 and June 2011, the Uganda Malaria Surveillance Project (UMSP), collaboration with National Control Program (NMCP), set up an enhanced sentinel site malaria surveillance program for at four public different districts: Tororo, Apac, Jinja Mubende. Clinical data collected through March 2013, representing 50249 admissions were used develop (derivation set). One year subsequently from same hospitals, 20406 admissions, prospectively validate performance (validation Using backward selection approach, 13 out 25 clinical parameters recognizable initial presentation, selected inclusion final logistic regression prediction model. The presence individual was awarded either 1 or 2 based coefficients. For each patient, composite generated. further categorized into three categories; low, medium, Patient characteristics comparable both sets. Measures included receiver operating curves area under curve (AUC), demonstrating good ability deathusing derivation (AUC =0.76) validation dataset =0.74). datasets, category as follows: low (0.8% vs. 0.7%), moderate (3.5% 3.2%), high (16.5% 12.6%), respectively. Our analysis resulted development that ably predicted While studies needed, this approach could be improve existing triage systems.