Might a syndromic approach for children with signs of severe illness usefully guide treatment?
Our gold standard was the final diagnosis. Disease progression during an inpatient stay might, through misclassification, tend to weaken the apparent performance of syndromic diagnosis. On the other hand, the use of the same clinical signs to define severity of final diagnosis and syndrome diagnosis will tend to improve sensitivity. Notwithstanding these considerations, positive predictive values for three of the severe syndromes (pneumonia, diarrhoea, and malaria) ranged from 46% to 70%, with sensitivities for identifying the gold standard diagnosis of at least 96%; this suggests that a syndromic approach may at least provide a clear rationale for how to target treatment. It increasingly is recognized that assignment of a single diagnosis is problematic (4, 6, 8, 11). The syndrome-based approach provides a structured format for prescribing multiple treatments to severely ill children a situation that, in our clinical experience, is often appropriate. It remains unclear, however, how t
