Yield of interview screening and chest X-ray abnormalities in a tuberculosis prevalence survey

Abstract

Background: Tuberculosis (TB) prevalence surveys generally rely on a combination of screening methods to identify suspects eligible for sputum culture.

Objective: To assess the yield of screening methods applied in a recent prevalence survey in Viet Nam and estimate the proportion of TB cases missed due to incomplete participation.

Methods: TB suspects were identified based on self-reported TB history or productive cough by interview and chest X-ray (CXR). We calculated the case yield of these two screening methods by dividing the number of cases detected per method by the total number of cases detected. As not all participants underwent the full screening procedure, we recalculated the maximum yield of the screening methods using multiple imputation methods.

Results: The yield from screening by interview and CXR were respectively 38% and 91%. Adjusting for missing data by multiple imputation, we estimated that we missed 9.9% (95%CI 6.8-14.2) of expected TB cases.

Conclusion: In prevalence surveys, screening by pre-structured interview is insufficient, and should be supplemented with CXR to achieve sufficient identification of TB cases. The effect of incomplete participation in the full screening procedure may be substantial and should be adjusted for in the analysis.