Estimating prevalence when the true disease status is incompletely ascertained

Area of Science:

• Epidemiology
• Biostatistics
• Medical Diagnostics

Background:

• Estimating true disease prevalence is crucial for public health interventions.
• Screening tests are widely used but can lead to underestimation of disease due to incomplete diagnostic follow-up.
• Incomplete evaluation of individuals, particularly those testing negative, poses a significant challenge in accurately assessing screening test performance.

Purpose of the Study:

• To develop and discuss methods for adjusting case counts in screening studies with incomplete diagnostic evaluations.
• To address the underestimation of disease prevalence and test sensitivity caused by unreferred individuals and undetected cases among negatives.
• To propose an analytical framework for studies involving three or more screening tests.

Main Methods:

• Statistical adjustment for individuals not fully evaluated after positive screening tests.
• Methods to account for undetected cases among individuals testing negative on all screening tests.
• Procedures for estimating disease prevalence and test sensitivity using a random sample of negatives in multi-test screening scenarios.

Main Results:

• The study provides a framework for adjusting for biases introduced by incomplete diagnostic evaluations in screening studies.
• It offers methods to correct for missed cases among both positive and negative screening results.
• The proposed methods enhance the accuracy of disease prevalence and test sensitivity estimations.

Conclusions:

• Accurate estimation of disease prevalence and screening test performance requires accounting for incomplete diagnostic evaluations.
• The proposed adjustments are essential for valid inference in multi-test screening studies.
• The methodology facilitates more reliable public health assessments based on screening program data.