Résumé : OBJECTIVE. Existing scoring systems for the diagnosis of group A streptococcus pharyngitis are insensitive or inapplicable in low-resources settings. Bacterial cultures and rapid tests can allow for antibiotic prescription abstention in high-income regions. These techniques are not feasible in many low-resources settings, and antibiotics often are prescribed for any pharyngitis episode. However, judicious antibiotics prescription in the community also is of concern in low-income countries. The objective of this study was to develop a clinical decision rule that allows for the reduction of empirical antibiotic therapy for children with pharyngitis in low-resources settings by identifying non-group A streptococcus pharyngitis. PATIENTS AND METHODS. We prospectively included children with pharyngitis in 3 public hospitals of Brazil during 9 months in 2004. We filled out clinical questionnaires and performed throat swabs. Bilateral χ2 (2-tailed test) and multivariate analysis were used to determine score categories. The outcome measures were sensitivity, specificity, positive likelihood ratio, and posttest probability of non-group A streptococcus infection with the clinical approach as compared with throat culture. RESULTS. A total of 163 of the 220 children had non-group A streptococcus pharyngitis (negative culture). We established a 3-questions decision rule (age and viral and bacterial signs) with 3 possible answers. The use of this score would prevent 41% to 55% of unnecessary antimicrobial prescriptions. The specificity of the score for non-group A streptococcus pharyngitis was >84%. CONCLUSION. Such a clinical decision rule could be helpful to reduce significantly unnecessary antibiotic prescriptions for pharyngitis in children in low-resources settings. Copyright © 2006 by the American Academy of Pediatrics.