Randomized incentives to increase participation in COVID testing in rural Kenya

Low participation in public health testing can lead to biased estimates of disease prevalence and inefficient allocation of public health resources. We evaluated the impact of monetary incentives and revisits on participation in door-to-door COVID testing in rural western Kenya. We conducted a cross-sectional study of all residents in 12 villages in rural western Kenya. We offered an incentive of KSh 200 (1.85 USD), KSh 350 (3.23 USD), or KSh 700 (6.47 USD), randomized at the household level, for each household member >3 months old who participated in COVID testing. Among 7,049 individuals, 5,659 individuals consented to testing. Overall, after revisits, 78.2% consented in the KSh 200 group, 80.6% in the KSh 350 group, and 81.9% in the KSh 700 group. Among individuals offered KSh 200, revisiting increased the consent rate from 68.1% to 78.2%, or 7.3 percentage points higher than the first visit consent rate in the KSh 700 group. Participation was very high (96.9%) among available individuals, but 17.2% of individuals were never available. Offering KSh 200 minimized the cost per consenting individual in our sample, even with revisits, compared to higher amounts. However, individuals in the KSh 700 group were slightly more likely to test positive, which suggests these individuals are missed at lower incentive amounts. Overall 0.3% (95% CI: 0.2, 0.5) tested positive for current infection by qPCR on nasal swabs, and 8.6% (7.7, 9.6) tested positive for SARS-CoV-2-specific antibodies by ELISA on dried blood spots. Accounting for nonresponse bias suggests this COVID population burden may be an underestimate. Our findings suggest incentives can increase participation in household door-to-door public health surveillance testing and may improve disease prevalence estimates by reducing nonresponse bias. However, high incentives may not be cost effective when low incentives motivate very high levels of participation. Significance StatementAchieving high levels of participation in testing can inform public health strategies for reducing the spread of infectious diseases. Cash incentives are one tool for increasing participation, but there is little evidence on the effectiveness of this approach in low-resource settings. We find that cash incentives increase participation in free COVID-19 testing in rural Kenya and may reduce nonresponse bias.