Abstract
OBJECTIVES: To evaluate the diagnostic performance of urine HIV antibody rapid test kits in screening diverse populations and to analyse subjects' willingness regarding reagent types, purchase channels, acceptable prices, and self-testing. DESIGNS: Diagnostic accuracy studies PARTICIPANTS: A total of 2606 valid and eligible samples were collected in the study, including 202 samples from female sex workers (FSWs), 304 persons with injection drug use (IDU), 1000 pregnant women (PW), 100 subjects undergoing voluntary HIV counselling and testing (VCT) and 1000 students in higher education schools or colleges (STUs). Subjects should simultaneously meet the following inclusion criteria: (1) being at least 18 years old and in full civil capacity, (2) signing an informed consent form and (3) providing truthful identifying information to ensure that the subjects and their samples are unique. RESULTS: The sensitivity, specificity and area under the curve (AUC) of the urine HIV-1 antibody rapid test kits were 92.16%, 99.92% and 0.960 (95% CI: 0.952 to 0.968, p<0.001), respectively, among 2606 samples collected during on-site screenings. The kits showed good diagnostic performance in persons with IDU (AUC, 1.000; 95% CI, 1.000 to 1.000, p<0.001), PW (AUC, 0.999; 95% CI, 0.999 to 1.000, p<0.001) and FSWs (AUC, 1.000; 95% CI, 1.000 to 1.000, p<0.001). The AUC of the urine reagent kits in subjects undergoing VCT was 0.941 (95% CI: 0.876 to 0.978, p<0.001). The 'acceptable price' had the greatest influence on STUs (Pi=1.000) and PW (Pi=1.000), the 'purchase channel' had the greatest influence on subjects undergoing VCT (Pi=1.000) and persons with IDU (Pi=1.000) and the 'reagent types' had the greatest influence on FSWs (Pi=1.000). CONCLUSIONS: The rapid urine test kits showed good diagnostic validity in practical applications, despite a few cases involving misdiagnosis and underdiagnosis.