# Comparative Analysis of Molecular Detection Algorithms for Enhanced Dengue Surveillance: A Field Case Evaluation

## TL;DR

This study evaluates the BioFire Global Fever Panel's accuracy for detecting dengue virus in real-world settings, comparing it with qRT-PCR and finding it reliable but costly.

## Contribution

The study provides real-world performance data of the BioFire Global Fever Panel for dengue detection using field surveillance data and advanced statistical modeling.

## Key findings

- The BioFire Global Fever Panel showed high sensitivity (94.5%) and moderate specificity (75.7%) when compared to qRT-PCR.
- A sequential testing algorithm combining qRT-PCR and the BioFire Panel achieved near-perfect sensitivity (99.3–100%).
- The BioFire Panel offers faster turnaround time but is more expensive than qRT-PCR.

## Abstract

The BioFire® Global Fever Panel (GFP; BioFire Diagnostics, Salt Lake City, UT) is an assay designed for detecting multiple pathogens, including dengue virus (DENV; Orthoflavivirus denguei). Although its performance has been primarily assessed using composite reference standards (CRS) or through direct comparisons with other tests, its real-world accuracy remains uncertain. The performance of the GFP in combination with real-time reverse transcriptase polymerase chain reaction (qRT-PCR) testing is evaluated in the present study using dengue surveillance data from 224 individuals with acute fever (≤5 days) in a dengue-endemic area of the Peruvian Amazon. Dengue virus RNA was detected in serum and whole blood using qRT-PCR testing and the GFP, respectively. Two reference standards were developed: one based on CRS and another based on latent class analysis. The best latent class model (LCM) included a viral load surrogate. Using qRT-PCR testing or the LCM as a reference, the GFP exhibited a sensitivity of 94.5% (95% CI: 88.4–98.0%) and a specificity of 75.7% (95% CI: 66.8–83.2%). With CRS, the GFP displayed a sensitivity of 95.6% (95% CI: 90.7–98.4%) and 100% specificity. The simultaneous and sequential two-test algorithms for negative yielded comparable performance and exhibited high sensitivity compared with other two-test sequential algorithms. The sequential testing algorithm, which used qRT-PCR testing followed by the GFP for negative results, offered the best balance in cost, time, and performance (sensitivity: 99.3–100%; specificity: 75.7–100%). Although the GFP is more expensive than qRT-PCR testing, it offers a shorter turnaround time per sample. Overall, the findings suggest that the GFP is a reliable and high-performing tool for DENV testing, but its cost may restrict its use in resource-limited settings.