The Fukushima research has examined data form a cohort study of 10,000 Japanese children under 18 years old with influenza during three months to demonstrate that the relative risk of A-type abnormal behavior of patients with oseltamivir was 30 times greater than without oseltamivir. By contrast, our research group found that patients who had been administered no neuraminidase inhibitors (NI) or those administered peramivir had higher risk of abnormal behavior than those administered oseltamivir, zanamivir, or laninamivir. A plausible explanation for this gap is that the two studies specifically examined different criteria to report abnormal behavior. In actually, some A-type abnormal behavior might not be life-threatening. Our definition of severe abnormal behavior is better matched to public health concerns and comparison among incidents according to the administered drug is more appropriate as an analytical procedure.