Published in:
Open Access 01-12-2023 | Zika Virus | Research article
Accuracy of Dengue, Chikungunya, and Zika diagnoses by primary healthcare physicians in Tegucigalpa, Honduras
Published in: BMC Infectious Diseases | Issue 1/2023
Login to get accessAbstract
Background
Dengue, Chikungunya, and Zika are co-endemic in Honduras and are often misdiagnosed due to similar clinical and epidemiological behavior. Most arboviral infections reported in primary care are based on clinical diagnoses without laboratory confirmation. Therefore, the accuracy of physicians’ diagnoses and the factors that affect them needs to be evaluated.
Methods
A cross-sectional study with convenience sampling at primary healthcare centers was conducted from June to September 2016 and 2017. Clinical data and dried blood spots on Whatman 903 filter paper from 415 arboviral cases and 248 non-arboviral febrile cases were collected. Viral RNA was extracted from a 6-mm DBS paper disc and confirmed by RT-qPCR and sequencing.
Results
Only 30.84% of diagnostic accuracy was observed in physicians in primary care when comparing arboviral clinical diagnosis with RT-qPCR detection. Moreover, in Dengue and Zika clinical cases, only 8.23% and 27.08% were RT-qPCR confirmed, respectively. No Chikungunya cases were confirmed. In 2017, 20.96% of febrile cases were RT-qPCR confirmed arboviral infections. The symptoms of 45.5% of arboviral cases can fit more than one case definition for arboviruses. The “symptom compliance” and “patient with suspected close contact” were the criteria most utilized by physicians for arboviral diagnosis. The pattern of the epidemiological curves of the arboviral clinical cases didn’t match the one of the RT-qPCR confirmed cases.
Conclusions
Low diagnostic accuracy for overall and individual arboviral infections was observed in physicians. Unspecific symptomatology, overlapping case definitions, and reported close contact to an arboviral patient might contribute to misdiagnosis. Without laboratory confirmation, surveillance data may not reflect the real behavior of these diseases and could impact health interventions.