Assessment of arbovirus vector infection rates using variable size pooling

Area of Science:

• Vector-borne disease surveillance
• Arbovirus ecology
• Entomology

Background:

• Pool testing is standard for arbovirus surveillance in mosquitoes.
• Minimum infection rate (MIR) assumes only one infected mosquito per pool, limiting accuracy with large pools or high infection rates.

Purpose of the Study:

• To compare maximum likelihood estimation (MLE) with MIR for estimating mosquito infection rates.
• To evaluate the impact of different pooling strategies on estimation accuracy.
• To identify optimal methods for West Nile virus (WNV) surveillance.

Main Methods:

• Applied MLE and MIR to West Nile virus (WNV) surveillance data from Chicago (2002).
• Conducted simulation studies with hypothetical mosquito samples (18 pools) using constant and variable pool sizes.
• Analyzed variable pool sizes (5, 10, 20, 30, 40, 50) against constant pool size (50).

Main Results:

• MLE revealed peak Culex mosquito infection rates of 60 per 1000, significantly higher than MIR estimates (<30 per 1000).
• Variable pool size strategies consistently outperformed constant pool size pooling in simulations.
• MLE with variable pool sizes demonstrated critical improvements for infection rate estimation.

Conclusions:

• MLE is superior to MIR for estimating mosquito infection rates across various field conditions.
• Variable pool size strategies enhance the accuracy of infection rate estimates.
• Combining MLE with variable pool sizes is essential for precise WNV surveillance during epidemic seasons.