Abstract

The inception of lateral flow device (LFD) has existed since the 1940s and has proven to be a flexibly technology for rapid protein detection, ranging from pregnancy tests to more recently SARS-CoV-2 infection status. With the mass distribution of LFDs in response to the pandemic an overlooked issue came to fore, interpretation of results. Here we discuss inhouse studies conducted into result interpretation by the public, error rates and the suitability of self-defined results. To address the error rate in user-called results, in collaboration with a commercial partner, an automatic AI based calling app was incorporated and ‘trained’ for a variety of SARS-CoV2 antigen detection LFD. Test results called were compared between the app, a professional operator and members of the public, showing a significant difference in test outcomes. We propose a rationale for AI-based result interpretation of self-read tests for unbiased calling, increase fidelity of outcomes reported and with increased traceability.