Applicability of Winthrop Score for the Diagnosis of Influenza A in the Emergency Department of Hospital Pablo Arturo Suárez, January to March of 2018
Abstract
Introduction: In 2010, the Department of Infectious Diseases at Winthrop University Hospital designed a score system for the diagnosis of Legionella pneumonia. In this study, we applied the score to patients with acute respiratory symptoms suspected of having type A influenza. The identification of patients at medium to high risk of Influenza A allows for early initiation of treatment.
Objective: To study the applicability of the Winthrop score for the diagnosis of Influenza A.
Methodology: A prospective cohort study was performed in 2018 at Hospital Pablo Arturo Suárez, in Quito, Ecuador. Patients 0 to 100 years old presenting to the emergency department with influenza-like illness in January-March of 2018 were included in the study. Winthrop score results were then compared with the result of the reverse transcription polymerase chain reaction (RT-PCR) for influenza A, the gold standard for diagnosis. Sensitivity, specificity, positive and negative predictive values, and likelihood ratios were used to establish the diagnostic performance of this point system for influenza A within the sample at large and in subgroup analyses by age (<5 years, 5-65 years, and >65 years) and comorbidities.
Results: 149 patients were enrolled in the study period. The study population included 81 males (54.4%) and the majority of patients were less than 5 years of age (N=85, 57.0%). Furthermore, almost one-third of the patients were less than one year old (N=38, 25.5%). According to the Winthrop point system, 68.5% of the cases had a low probability of having influenza (n = 102), 8.7% of cases had a medium probability (n = 13) and 22.8 % of cases had a high probability (n = 34). The RT-PCR test for influenza was positive for 26.2% of patients (n = 39). The Winthrop point system had a sensitivity of 97.4%, specificity of 91.8%, positive predictive value of 80.8%, negative predictive value of 99.0%, positive likelihood ratio of 11.9, and negative likelihood ratio of 35.8 in the total study population. For children under 5 years, a sensitivity of 100%, specificity of 96.3%, positive predictive value of 77.7%, negative predictive value of 100%, positive likelihood ratio of 27, and negative likelihood ratio of 0. In patients older than 6 years, a sensitivity of 96.9%, specificity of 89%, positive predictive value of 84.21%, negative predictive value of 98%, positive likelihood ratio of 8.8, and negative likelihood ratio of 29.4. Testing in patients over 65 years had a sensitivity of 100%, specificity of 90%, positive predictive value of 87.5%, negative predictive value of 100%, positive likelihood ratio of 10 and negative likelihood ratio of 0. Finally, patients with comorbidities had a sensitivity of 90%, specificity of 88.24%, positive predictive value of 81.82%, negative predictive value of 93.75%, positive likelihood ratio of 7.65, and negative likelihood ratio of 8.82.
Conclusions: The Winthrop score performed well in predicting Influenza A in patients with acute respiratory symptoms. This score may be useful in settings were Influenza A PCR testing is unavailable.