Suspected Pediatric Influenza Risk-Stratification Algorithm: A Clinical Decision Tool.

ABSTRACT

BACKGROUND AND OBJECTIVES: Influenza causes significant annual burden among children. Current guidelines recommend empiric treatment for a broadly defined group of children at high risk for influenza complications, resulting in overtreatment or costly viral testing. This study creates an algorithm for clinicians to risk stratify children with influenza-like illness (ILI) according to likelihood of influenza infection. METHODS: A retrospective analysis was performed on 818 children seen in the emergency department from November 2012 to April 2013 for ILI. We reviewed medical records for symptoms, influenza risk factors, and viral assay results. Classification and regression tree analyses were performed separately for children older and younger than 2 years. RESULTS: In children younger than 2 years, populations likely to test positive were those with an influenza-positive contact, unimmunized children, and those presenting in high-incidence influenza periods. In this subgroup, immunized patients in low-incidence seasons and those with absence of cough are low risk for influenza infection. For children 2 years and older, high-risk populations were unimmunized children, those presenting in high-incidence influenza periods and those with myalgia or absence of diarrhea. CONCLUSIONS: These risk-stratification analyses were summarized into Suspected Pediatric Influenza Risk-Stratification Algorithm (SPIRA). For those in whom influenza infection is likely, clinicians may consider empiric treatment. Conversely, patients whom SPIRA identifies as unlikely to be infected with influenza are candidates for viral testing and targeted treatment. In assessing children with ILI, SPIRA aids clinicians in determining who to test versus treat empirically, saving children from costly viral testing or unnecessary antiviral exposure.