The accuracy of a clinical parameters-based scoring system to predict spontaneous intracranial hemorrhage in children under one year old

Harris Alfan (Departments of Child Health, Sriwijaya University Medical School/Moh Hoesin Hospital, Palembang, Indonesia.)
Rita Dewi Arifin (Departments of Child Health, Sriwijaya University Medical School/Moh Hoesin Hospital, Palembang, Indonesia.)
Erial Bahar (Departments of Biostatistics, Sriwijaya University Medical School/Moh Hoesin Hospital, Palembang, Indonesia.)
Syarif Darwin Ansori (Departments of Child Health, Sriwijaya University Medical School/Moh Hoesin Hospital, Palembang, Indonesia.)

Paediatrica Indonesiana
Vol 55 No 3 (2015): May 2015

Background Previous studies show that most children aged less than 1 year had intracranial hemorrhage without any history of trauma. The sign and symptoms of spontaneous intracranial hemorrhage (SIH) in children varies. To minimize morbidity and mortality, early detection and accurate diagnosis are required. Head CT scans area widely used for diagnosing SIH. Unfortunately, not all health facilities in Indonesia have CT scans.Objective To determine the accuracy of a clinical parameters-based scoring system in predicting spontaneous intracranial hemorrhage (SIH) in children under one year old.Methods This diagnostic study included children aged under one year who were admitted to Mohammad Hoesin Hospital, Palembang. Patients who showed any signs of increased intracranial pressure were recruited. Data were collected from medical records from January 2007 to September 2013. Through the use of logistic regression analysis, clinical parameters showing significant relationships with computerized tomography (CT)-scan confirmed SIH were selected as predictors. Each predictor was given a score based on an adjusted ratio. The cut-off point of the total scores from all patients was determined using a receiver operating curve (ROC) analysis. The accuracy of the total scores was calculated using a 2x2 validity test.Results Of the 186 children included in this study, 98 (52.7%) had SIH and 93 (94.8%) were under 3 month-old. The predictors for SIH used included age (>3 months: score 0; 1-3 months: score 3), gender (female: score 0; male: score 1), pallor (no: score 0; yes: score 1), bulging fontanel (no: score 0; yes: score 1), pupil (isocoria: score 0; anisocoria: score 2) and history of shaken baby (no: score 0; yes: score 3). The ROC analysis showed that the area under the curve (AUC) was 95.3% with a cut-off point of 4.5, had a sensitivity of 88.7% and a specificity of 93.1%Conclusion This scoring system based on clinical parameters had good accuracy for predicting SIH in children under 1 year of age who exhibited signs of increased intracranial pressure.