Trend-based quality assurance of binary urinary antigen tests.

Abstract

Binary diagnostic assays, such as urinary antigen (uAg) tests, provide rapid results but lack quantitative output for conventional quality control. This study evaluated whether long-term monitoring of test positivity patterns can serve as a practical tool for quality assurance (QA) of pneumococcal and Legionella uAg assays. All pneumococcal and Legionella uAg tests performed in Uppsala County, Sweden (01/01/2007-31/12/2024), were retrospectively analyzed. Positivity trends were assessed over time, including seasonal variation and demographic subgroups. Outlier detection was performed using interquartile range (IQR) thresholds to identify potential analytical drift. Bayesian predictive values were additionally estimated as a theoretical comparison to illustrate the influence of prevalence on predictive performance. A total of 17,356 pneumococcal and 15,280 Legionella uAg tests were included. Pneumococcal positivity displayed significant seasonal variation, lowest in August (OR 0.49; 95% CI 0.35-0.68), whereas Legionella positivity varied mainly by year, with peaks in 2007, 2012, and 2021-2022. Pneumococcal positivity was highest in children and female patients, while Legionella showed no demographic trends. QA intervals derived from IQR thresholds captured expected long-term stability (7.4% for pneumococcus; 0.7% for Legionella), with outliers corresponding to known epidemiological events. Bayesian estimates highlighted the large discrepancy between incidence-based and test-based predictive values but were secondary to the trend-based framework. Rapid uAg tests pose challenges for QA. Positivity trends provide a feasible strategy for long-term, population-based monitoring of binary diagnostic tests and can complement conventional controls and detect analytical drift.