ABSTRACT
Background: Most programs use a screen and test strategy to identify SARS-CoV-2 infection, but this strategy does not identify individuals with asymptomatic infection. We determined the SARS-CoV-2 case detection rates in a test-all model compared to the standard screen-and-test model in Kenya and Cameroon. Method(s): A cluster-randomized trial was conducted in 20 health facilities between May-October 2022. In each country, 5 facilities were randomized to test all (testing offered regardless of screening outcome) or screen and test (testing offered if screened positive) arms. Additional staff were hired to support implementation of the two models in Kenya (K) and the test all model in Cameroon (C). Clients age>2 years attending HIV, TB and MNCH clinics were tested using SARS-CoV-2 rapid antigen tests. We estimated case detection rates (CDR) with facility level weighted averages and used a weighted t-test with robust standard errors for between arm comparison. Result(s): Overall, 80,828 attendee visits were reported in the test-all arm (63,492 C and 17,336 K) and 71,254 attendee visits were reported in the screenand- test arm (56,589 C and 14,665 K). In the test-all arm, 42,325 (52.4%) were screened for COVID-19 symptoms (46.7% C and 73.2% K) and 21,536 (26.6%) were tested (29.2% C and 17.4% in Kenya) with a positivity rate of 1.4% (2.0% C and 1.1% K). In the screen-and-test arm, 48,314 (67.8%) were screened (72.8% C and 48.6% K), and 3,629 (7.5%) were eligible for testing (8.2% C and 3.7% K) - of those, 2,139 (58.9%) were tested (57.1% C and 82.4% K) with a positivity rate of 4.1% (3.4% C and 10% K). The estimated CDR was 3.59 (95% CI:1.55-5.64) per 1,000 attendee visits in the test-all arm and 1.46 (95% CI:0.60-2.32) per 1,000 attendee visits in the screen-and-test arm. Compared to the screen-and-test arm, the test-all arm had significantly higher COVID-19 CDR in MNCH clinics (3.57 vs.1.29, p=0.034). There were no significant differences in COVID-19 CDR between the two arms in HIV (4.20 vs.1.98, p=0.174) and TB (10.33 vs. 5.03, p=0.283) clinics, though the number of SARS-CoV-2 infections was small. Conclusion(s): The test-all arm identified more SARS-CoV-2 cases than the routine screen-and-test model, despite overall low testing coverage. The test-all model should be considered in future epidemics to improve early detection of SARS-CoV-2 infection among vulnerable populations, but effective implementation requires additional human resources to manage the clinic volumes. COVID-19 Case Detection Rates Per 1,000 Attendees: Comparison of Screen-and- Test and Test-All Arms.