A SCORING SYSTEM TO QUANTIFY ILLNESS IN BABIES UNDER 6 MONTHS OF AGE

Ordinal logistic regression analysis was used on a sample of 1007 babies under 6 months of age seen at home and in hospital to identify the symptoms and signs of serious illness. A model with seven symptoms and 12 signs was used as the basis for a scoring system where each symptom and sign was assigned an integer score proportional to its regression coefficient. A baby's illness score was then the sum of scores for any signs or symptoms that were present, asymptomatic babies scoring 0. The scoring system gave a sensitivity (score 13 or more) of 92.1%, specificity (score less than 8) of 98.3% and positive predictive value (score 20 or more) of 98.0% when applied to a notional community sample. The system has been made into a score card called Baby Check, designed for use by parents and health professionals to quantify the severity of a baby's illness. Baby Check indicates when the baby's condition warrants further assessment or treatment.