Feasibility of a Pilot Crowdsourced Syndromic and Virological Surveillance Platform for Respiratory Illness in South Africa, CoughWatchSA, 2022

Background Digital participatory surveillance (DPS) may provide information on non-medically attended influenza-like illnesses (ILI). Combining DPS with laboratory testing allows pathogen identification. We assessed the feasibility of DPS and home-based swabbing in South Africa. Methods We enrolled a cohort of individuals aged ≥18 years who completed weekly respiratory symptoms questionnaires from March–October 2022. We calculated the percent of weekly DPS-reports consistent with ILI and COVID-19 syndromes and compared with weekly private respiratory consultations. Symptomatic participants in 3 provinces were offered self-administered nasal swabbing for influenza and SARS-CoV-2 detection by polymerase chain reaction (PCR). We assessed six feasibility indicators: recruitment capability, representativeness, reliability, accuracy, acceptability, and usefulness. Results Recruitment capability: 26% (249/954) of participants who accessed the platform enrolled and 92% (81/88) of participants eligible for self-swabbing were enrolled. Acceptability: 84% (210/249) of participants remained enrolled until the end of follow-up, 15% (32/249) completed the acceptability questionnaire with 100% (32/32) indicating willingness to participate in future studies, whereas 16% (39/210) withdrew from the study. Representativeness: 50% (125/249) were aged 18-39 years and 50% were older than 40 years (124/249), 71% (178/249) were female and 79% (196/249) had a tertiary qualification. Reliability: median weekly participation was 23% (interquartile range (IQR): 19%-29%) and 38% (80/210) submitted >1 weekly symptoms questionnaire. Accuracy: 2% (31/1440) and 25% (359/1440) of weekly questionnaires met ILI and COVID-19 case definitions, respectively. Influenza and SARS-CoV-2 were detected in 7% (6/81) and 32% (26/81) of tested samples, respectively. There was low correlation with respiratory consultations (0.08, 95% CI -0.27–0.43) for ILI and (0.36, 95% CI 0.11–0.62) COVID19. Usefulness: symptoms were reported in 32% (459/1440) of reports and 11% (49/459) sought medical care. Conclusion The study was feasible, however, the low enrolment numbers limit power. Linkage to self-swabbing was successful and demonstrates the potential for pathogen confirmation.