ABSTRACT
Background: The COVID-19 antibody sub-study of the Canadian Longitudinal Study on Aging (CLSA) has been designed to examine the seroprevalence of the SARS-CoV-2 infection from 19,000 participants across Canada over three timeperiods. The largest set will be performed by in-home collection using a dried blood spot (DBS) device. Methods: We evaluated two self-collection devices, Mitra® (Neoteryx) and Velvet™ (Boston Microfluidics), considering cost, delivery, availability, ease of sample collection and return, sample volume, blood fractionation, normalization requirement, hemolysis, and analysis of additional analytes and ease of storage. We developed collection instructions, a quality of collection questionnaire, and a process to ship, return and store the devices. Results: Velvet™ was selected for several reasons. Hand eye coordination was less demanding for collecting drops of blood into a well compared to dropping blood on several microfiber tips for the Mitra®. Normalization for volume collected is needed for both, but hemoglobin interference is not an issue with Velvet™ as blood is fractionated into cell components and plasma. Storage of Velvet™ at -80°C was easier than Mitra® with a unique barcode on each device for scanning and were easily stackable in trays with its rectangular shape and smooth surface. Customized collection kits were made with devices, retractable lancets, gauze, Band-Aids, return pouch and in-structions. Partnership with FedEx and Canada Post provided timely delivery of collection kits and return of devices. An instruction video showing how to collect the sample was prepared and posted on the CLSA website. In the first collection we obtained 3339 devices of the 3803mailed out. Almost all participants completed the collection within 15 minutes, but some did not record the date and time when it was done. Some participants required collection from two fingers and almost all wiped their finger with an alcohol swab. Few participants found the instructions difficult to follow but most had to squeeze their finger to achieve good flow. About one-quarter of devices were underfilled. Conclusion: We were successful in our rapid implementation of DBS collection for the seroprevalence study, but improvements can be made in the participant instructions to achieve an easier and more complete collection using Velvet™.