A semiparametric two-component "compound" mixture model and its application to estimating malaria attributable...

Area of Science:

• Epidemiology
• Biostatistics
• Infectious Disease Modeling

Background:

• Malaria diagnosis is complex in endemic areas due to asymptomatic parasite carriers.
• Accurate estimation of clinical malaria cases is vital for public health interventions and resource allocation.
• Existing diagnostic methods face challenges in differentiating true malaria cases from asymptomatic parasitemia.

Purpose of the Study:

• To develop and evaluate a semiparametric likelihood approach for estimating the proportion of clinical malaria.
• To accurately quantify malaria cases in symptomatic individuals using parasite-level data.
• To provide a robust statistical framework for malaria intervention policy development.

Main Methods:

• Proposed a semiparametric likelihood approach utilizing parasite-level data from symptomatic individuals.
• Modeled the density ratio of parasite levels in clinical vs. non-clinical malaria cases using a logistic model.
• Employed empirical likelihood to effectively integrate both zero and nonzero parasite count data.

Main Results:

• The proposed maximum semiparametric likelihood estimate demonstrates superior efficiency compared to nonparametric methods.
• The method offers enhanced robustness over fully parametric approaches by not assuming a specific parametric model for nonzero data.
• Simulation studies confirmed the satisfactory performance and reliability of the developed statistical method.

Conclusions:

• The semiparametric likelihood approach provides a more accurate and robust estimation of clinical malaria proportions.
• This method improves upon existing techniques for analyzing parasite-level data in malaria diagnosis.
• The approach is applicable to real-world malaria surveys, as demonstrated by its use in a Tanzanian study.