Improving statistical inference on pathogen densities estimated by quantitative molecular methods: malaria...

Area of Science:

• Molecular biology
• Biostatistics
• Infectious disease epidemiology

Background:

• Quantitative molecular methods (QMMs) like q-PCR and QT-NASBA are vital for estimating pathogen density but are often semi-quantitative with poorly reported reliability.
• Assessing the reliability and diagnostic sensitivity of QMMs is crucial for accurate clinical and epidemiological applications.
• This study focuses on improving the statistical framework for evaluating QMM-derived pathogen density estimates.

Purpose of the Study:

• To develop and evaluate a statistical framework for assessing the reliability and diagnostic sensitivity of pathogen densities estimated using QMMs.
• To compare a traditional calibration method with a novel mixed model Bayesian approach for QMM data.
• To illustrate the method using Plasmodium falciparum gametocyte quantification via QT-NASBA.

Main Methods:

• Development of a statistical framework to quantify uncertainty in pathogen density estimates from QMMs.
• Comparison of a traditional calibration technique with a mixed model Bayesian approach.
• Application of the Bayesian approach to estimate Plasmodium falciparum gametocytaemia using quantitative nucleic acid sequence-based amplification (QT-NASBA).

Main Results:

• Traditional calibration methods yield unreliable and variable pathogen density estimates, inaccurately reflecting diagnostic sensitivity.
• The Bayesian mixed model approach accounts for inter-assay variability and improves the reliability and homogeneity of QMM estimates.
• This method provides a more accurate appraisal of both quantitative and diagnostic performance of QMMs.

Conclusions:

• Bayesian mixed model statistical calibration significantly outperforms traditional techniques for QMM-derived pathogen density estimates.
• This approach has the potential to substantially enhance the depth and quality of clinical and epidemiological inference across various pathogens.
• Improved statistical methods are essential for maximizing the utility of QMMs in public health and clinical diagnostics.