Background/Case Studies: Diverse recruitment strategies were used in SARS-CoV-2 serological surveillance studies, including testing residual samples from blood donations. The sociodemographic representativeness of blood donors has not been compared to other study populations.
Study
Design/Methods: We compared the sociodemographic representativeness of Canadian Blood Services’ SARS-CoV-2 serosurveillance study to four concurrent Canadian studies: three pre-existing longitudinal cohorts and a cross-sectional study using blood draws collected for outpatient testing. We calculated study specimen and population counts by age, sex, urban or rural residence, race/ethnicity, and neighborhood material deprivation quintile. For each study, we calculated a representation ratio for demographic subgroups by dividing the proportion of study specimens in a subgroup by the proportion of the population in the subgroup based on 2016 Canadian census data. We classified subgroups as underrepresented (representation ratio < 3/4), adequately represented (3/4 ≤ representation ratio ≤ 4/3), or overrepresented (representation ratio > 4/3). We considered subgroups statistically-significantly underrepresented if representation ratios were below ¾ in more than 95% of bootstrap resamples.
Results/Findings: Across all studies, 71% of racialized minority subgroups stratified by age and sex were underrepresented, and rural regions were consistently underrepresented compared to their urban counterparts (Figure A). Young racialized minorities and male rural residents had larger representation ratios in the blood donor study compared to others, indicating better representation. Of the two residual sample studies, blood donor samples were more representative of participants aged 18-46 across sex and urban strata, whereas outpatient laboratory samples were more representative of individuals aged 47 and older. Outpatient laboratory samples had adequate representation across quintiles of neighborhood material deprivation, but blood donor observations skewed toward less materially deprived areas. Conclusions: Some subgroups were underrepresented in all studies, particularly among rural residents and racialized minorities. Blood donors were more representative of some subgroups than other studies, such as younger racialized minorities, and may be an appropriate population for broader infectious disease surveillance depending on the pathogen and use case.
