Saraiva, Ana LeonorAfreixo, VeraGaio, VâniaMachado, Ausenda2026-03-042026-03-042025-07-10http://hdl.handle.net/10400.18/11135Introduction: The emergence of COVID-19 led to the rapid development of vaccines and diagnostic tests. To assess antibody responses in healthcare workers (HCWs), a 2021–2022 cohort study was conducted in three Portuguese hospitals. Antibody levels were measured at six time points: pre-vaccination, post-first dose, at 3, 6, 12 months after the second dose, and post-booster. Each hospital utilized a different assay: Abbott’s CMIA, Roche’s Elecsys® ECLIA, and Siemens’ ADVIA Centaur®, leading to challenges in data comparability. The study aimed to harmonize serological data across these hospitals to enable joint analysis and better understand the dynamics of humoral immunity among HCWs in Portugal. Methods: To ensure adequate conversion of antibody titers obtained from different laboratory methods, several harmonization strategies were applied and compared. These included the World Health Organization (WHO) international conversion and a main strategy based on quantile interpolation, with linear and spline methods tested, followed by Deming regression. Before interpolation, models were required to estimate the antibody titers due to the limitations of each method, such as linear regressions with polynomials and splines, followed by data linearization and normalization. The final choice between methods was based on a graphical assessment of which approach best preserved the original data patterns prior to harmonization. Results: The application of quantile interpolation followed by regression proved more effective than the WHO recommended approach for harmonizing data across hospitals. This method preserved the individual distribution patterns of each hospital, even after transformation. Importantly, it allowed the harmonized values to reflect the scale and magnitude observed in the reference method (Abbott’s CMIA). After harmonization, a marked increase was observed between the pre-vaccination time point and the post-vaccination time point, followed by a gradual decrease over the next three time points. Finally, a second, even more pronounced increase was observed after the booster dose. Conclusions: Quantile interpolation followed by regression enabled effective harmonization of serological data across the hospital centers, allowing for the comparison and joint analysis of anti-SARS-CoV-2 IgG antibody titers from different laboratory methods. This approach also preserved the expected temporal antibody response pattern. Although laboratory validation, by quantifying the same samples across different laboratories, was not performed, which would be necessary to fully confirm the adequacy of the method, the presented methodology shows strong applicability to other contexts, particularly in European studies that conduct pooled analyses of data from multiple countries and laboratories.engSARS-CoV-2HarmonizationHealthcare ProfessionalsEstado de Saúde e de DoençaDeterminantes de SaúdeInfecções Respiratóriasconference object