Sensitivity and specificity of OraQuick® HIV self-test compared to a 4th generation laboratory reference standard algorithm in urban and rural Zambia

Table 3 Agreement between user-conducted and user-read OraQuick® result and RDT Algorithm and Laboratory Reference Standard

User-read	Rapid diagnostic test (RDT) algorithm result*			Laboratory reference standard result*
User-read	Positive	Negative	Total**	Positive	Negative	Total***
Reactive	222	8	230	223	7	230
Non-reactive	14	2305	2,319	32	2287	2319
Sub-total	236	2313	2549	255	2294	2549
Agreement (%) Cohen's kappa	99.14 0.9481 (95% CI: 0.9265–0.9697)			Agreement (%) Cohen's kappa	98.50 0.9114 (95% CI: 0.8838- 0.9390)
Sensitivity (%)	94.1 (95% CI: 90.2–96.7)			Sensitivity (%)	87.5 (95% CI: 82.7–91.3)
Specificity (%)	99.7(95% CI: 99.3–99.9)			Specificity (%)	99.7 (95% CI: 99.4–99.9)

^*The performance of the RDT Algorithm compared to Laboratory Reference Standard was: sensitivity 93.4% (95% CI: 89.7–96.1) and specificity 100% (95% CI: 99.8–100) when compared with the laboratory gold standard
**Excludes 6 self-reported ART users, 2 users missing OraQuick® results, 5 clients missing RDT results, 9 OraQuick® results read as invalid by client, 5 clients with indeterminate RDT results (25 total; 2 clients missing both RDT and OraQuick® results)
***Excludes 6 self-reported ART users, 9 OraQuick® results read as invalid by client, 10 clients missing laboratory results, and 2 clients missing laboratory and OraQuick® results (25 total; 2 clients missing both laboratory and OraQuick® results)

ISSN: 1471-2334