Clinical diagnosis for Chikungunya infection: An essential aid in a primary care setting where serological confirmation is not available

Abstract

Background Chikungunya virus (CHIKV) diagnosis have become a challenge for primary care physicians in areas where zika virus and/or dengue virus are present. Case definitions for the three arboviral infections are overlapping. Method A cross-sectional analysis was carried out. A bivariate analysis was made using confirmed CHIKV infection as the outcome. Variables with significant statistical association were included in an agreement consensus. Agreed variables were analysed in multiple regression model. The area under the receiver operating characteristic (ROC) curve was calculated to determine a cut-off value and performance. Results 295 patients with confirmed CHIKV infection were included. A screening tool was made using symmetric arthritis (4 points), fatigue (3 points), rash (2 points) and ankle joint pain (1 point). The ROC curve identified a cut-off value and a score ≥ 5.5 was considered positive to identify CHIKV patients with a sensibility of 64.4% and a specificity of 87.4%, positive predictive value of 85.5%, negative predictive value of 67.7%, area under the curve of 0.72, and an accuracy of 75%. Conclusion We developed a screening tool for CHIKV diagnosis using only clinical symptoms as well as proposed an algorithm to aid the primary care physician.