True versus Apparent Malaria Infection Prevalence: The Contribution of a Bayesian Approach
2011

Estimating True Malaria Prevalence Using Bayesian Methods

Sample size: 504 publication 10 minutes Evidence: moderate

Author Information

Author(s): Speybroeck Niko, Praet Nicolas, Claes Filip, Van Hong Nguyen, Torres Kathy, Mao Sokny, Van den Eede Peter, Thi Thinh Ta, Gamboa Dioni, Sochantha Tho, Thang Ngo Duc, Coosemans Marc, Büscher Philippe, D'Alessandro Umberto, Berkvens Dirk, Erhart Annette

Primary Institution: Research Institute Health and Society (IRSS), Université Catholique de Louvain la Neuve, Brussels, Belgium

Hypothesis

Can a Bayesian approach provide a more accurate estimate of malaria prevalence without a gold standard diagnostic test?

Conclusion

Bayesian methods can effectively analyze malaria prevalence when no gold standard diagnostic test is available.

Supporting Evidence

  • Bayesian methods provide a framework for combining multiple diagnostic tests.
  • Results showed that ELISA had the highest estimates of malaria prevalence.
  • Microscopy was closer to the true estimate in Vietnam and Peru compared to other tests.
  • Significant differences in test sensitivity and specificity were observed across different countries.
  • The study included data from three countries: Vietnam, Cambodia, and Peru.

Takeaway

This study shows that using a special math method called Bayesian can help us understand how many people really have malaria, even when we don't have the best tests.

Methodology

Bayesian models were developed to estimate malaria prevalence using various diagnostic tests without a gold standard.

Potential Biases

Potential bias from using expert opinions and the variability of diagnostic test performance across different settings.

Limitations

The study relies on expert opinions for estimating test characteristics, which may introduce bias.

Participant Demographics

Participants were aged 2 to 60 years from Vietnam, Cambodia, and Peru.

Statistical Information

P-Value

p<0.05

Confidence Interval

95% CI

Statistical Significance

p<0.05

Digital Object Identifier (DOI)

10.1371/journal.pone.0016705

Want to read the original?

Access the complete publication on the publisher's website

View Original Publication