Optimizing and validating a pediatric screening tool to more efficiently test and identify children living with HIV

Abstract

Objective: To develop and validate a screening tool to improve testing efficiency and increase case finding of children living with HIV.

Design: Cross-sectional study.

Methods: Between November 2020 and September 2021, children 18 months to 14 years presenting at outpatient departments in 30 health facilities in Zambia were administered a 14-question pediatric HIV screening tool and then tested for HIV. Data were analyzed using a randomly extracted 'validation' dataset and multivariable logistic regression to determine the highest performing and optimal number of screening questions. The final tool was then evaluated in the 'test' dataset. Sensitivity and specificity were calculated for both datasets. The final tool was then also implemented in 12 additional facilities to determine operational feasibility and uptake.

Results: A total of 9902 children were included in the final analysis. HIV prevalence was 1.3%. Six questions were significantly associated with HIV-positivity. The optimal screening cutoff score was to answer 'yes' to one or more of the six questions; using this cutoff sensitivity was 92.5% [95% confidence interval (CI) 85.7-96.7%] and specificity was 62.9% (95% CI 61.9-64%). In the test dataset, the same tool had a sensitivity of 84.6% (95% CI 65.1-95.6%) and specificity of 64.6% (95% CI 62.4-66.7%). Uptake was 89%.

Conclusion: The results of this study show sensitivity and acceptable specificity in a six-question validated HIV screening tool. Implementing this screening tool in settings where universal testing is not feasible should more efficiently accelerate identification of children living with HIV (CLHIV) and their timely initiation onto life-saving drugs.