Computerized scoring system for the diagnosis of foreign body aspiration in children.

OBJECTIVES: Foreign body aspiration (FBA) is a life-threatening event in children. The gold standard for diagnosis is bronchoscopy, but there is no consensus regarding indications for the procedure. The aim of this study was to formulate a predictive model for assessing the probability of FBA in suspected cases as an aid in the decision to perform diagnostic bronchoscopy.

METHODS: The files of 150 patients who underwent bronchoscopy for suspected FBA at our center between 1996 and 2004 were reviewed for medical history, physical examination, and radiologic studies. The findings were analyzed by logistic regression.

RESULTS: Using the file data, we formulated a predictive model wherein each parameter received a numeric coefficient representing its significance in evaluating suspected FBA. The most significant parameters were age 10 to 24 months, foreign body in the child's mouth and severe respiratory complaints during the choking episode, hypoxemia, dyspnea or stridor following the acute event, unilateral signs on lung auscultation, abnormal tracheal radiogram, unilateral infiltrate or atelectasis, and local hyperinflation or obstructive emphysema on chest radiogram.

CONCLUSIONS: In our predictive model, every case of suspected FBA can be assigned a score based on the specific parameters present, which is then entered into a probability formula to determine the likelihood of a positive diagnosis. This model may serve as a useful tool for deciding on the use of bronchoscopy in all children with suspected FBA.