Using Mixture Models to Diagnose Dengue Virus Infections
Author Information
Author(s): Hamins-Puértolas Marco, Buddhari Darunee, Salje Henrik, Huang Angkana T., Hunsawong Taweewun, Cummings Derek A.T., Fernandez Stefan, Farmer Aaron, Kaewhiran Surachai, Khampaen Direk, Srikiatkhachorn Anon, Iamsirithaworn Sopon, Waickman Adam, Thomas Stephen J., Endy Timothy, Rothman Alan L., Anderson Kathryn B., Rodriguez-Barraquer Isabel
Primary Institution: University of California, San Francisco
Hypothesis
Can a Bayesian mixture model improve the accuracy of diagnosing dengue virus infections from paired serological assays?
Conclusion
The study found that using multiple serological assays improves the accuracy of dengue infection detection.
Supporting Evidence
- The model was tested on 677 pairs of serum samples.
- It achieved F1 scores of 87-90% for detecting infections.
- The model outperformed traditional methods that used single assays.
Takeaway
This study shows that using different blood tests together can help doctors figure out if someone has dengue fever more accurately.
Methodology
A Bayesian mixture model was developed and tested using simulated data and applied to real-world data from two longitudinal cohort studies.
Potential Biases
Potential biases due to varying immune responses and sample timing across populations.
Limitations
The model may not fully capture recent infections and requires further validation in different populations.
Participant Demographics
Participants included children and adults from two cohort studies in Kamphaeng Phet, Thailand.
Statistical Information
P-Value
p<0.05
Statistical Significance
p<0.05
Digital Object Identifier (DOI)
Want to read the original?
Access the complete publication on the publisher's website